Skip Navigation

The David Bodian Seminar Series @ The Zanvyl Krieger Mind/Brain Institute




09/15/2003 4:00pm
James J. DiCarlo, MD, PhD
Asst. Professor of Neuroscience, McGovern Institute for Brain Research Dept. of Brain and Cognitive Sciences M.I.T.
Massachusetts Inst Technol (MIT)

Neuronal mechanisms in inferotemporal cortex underlying visual object recognition in primates

The long-term goal of our research is an understanding of the neuronal computations that support the brain’s remarkable ability to recognize visual objects. The key computational challenge of object recognition is the extraction of object identity irrespective of visual clutter, object position, size, pose and illumination. Our working hypothesis is that a series of brain processing stages gradually transform pixel-based images of the world into patterns of neuronal activity that emphasize object identity and discount object position, size, view, and illumination. Such neuronal representations would be well-suited to drive recognition behavior (report of object identity). Because several lines of evidence suggest that such high level neuronal representations are conveyed by neuronal responses in the inferotemporal cortex (IT), our studies are focused on understanding the IT neuronal representation, how it is produced, and its role in driving recognition behavior. In this talk, I will outline several of our recent studies of IT responses in monkeys engaged in visual object recognition tasks. I will review our studies in free viewing animals in which we asked if IT responses are altered by free viewing (relative to the standard laboratory fixation condition). I will then describe our recent observation that IT neuronal tolerance to changes in an object’s retinal position can be remarkably limited. Although several hypotheses `may explain this observation, one of the most intriguing is the possibility that visual experience plays a strong role in shaping IT position tolerance, and I will present results from our ongoing studies aimed at testing this hypothesis.

Kenneth Johnson, PhD

09/22/2003 4:00pm
Peter N. Steinmetz, MD, PhD
Assistant Professor, Brain Modeling Laboratory, Dept of Biomedical Engineering, University of Minnesota
University of Minnesota

Attentional shift abolishes response selectivity of single neurons in the human hippocampus but not in the amygdala

A substantial minority of neurons in the human medial temporal lobe (hippocampus, amygdala, entorhinal cortex, para-hippocampal gyrus) respond selectively to categories of pictures, such as faces, spatial layouts or cars. Given the role of the medial temporal lobe in processing visual stimuli for declarative memory, one would predict that these responses would be modulated by attention. We recorded from 153 neurons in the human medial temporal lobe while subjects were presented with visual stimuli of seven categories. Here we show that category selective visual responses in the hippocampus, which are present when a subject must attend to presented pictures, are abolished when attention is focused on a superimposed video game. By contrast, selective responses in the amygdala are largely independent of the task being performed.

Ernst Niebur, PhD

09/29/2003 4:00pm
Peter C. Holland, PhD
The Krieger-Eisenhower Professor, Dept of Psychological & Brain Sciences, Johns Hopkins University
Johns Hopkins University

Amygdalar-cortical interactions in attention for action and new learning of rats

Animals’ attention to stimuli is influenced by the relation of those stimuli to motivationally significant events, such as food. I will discuss a series of experiments with rats, in which we investigated amygdalar-cortical brain systems involved in the control of action by, and new learning about, visual and auditory stimuli paired with food. In general, more reliable predictors of food were more likely to control stimulus-directed action than unreliable predictors, but new learning was more likely to be directed to the unreliable cues. The learning-based enhancements of action and new learning were both dependent on the integrity of the central nucleus of the amygdala, but each was mediated by different amygdalar-cortical circuitry. Implications of these findings for concepts of top-down attentional control and decision will be considered.

Steven Hsiao, PhD

10/06/2003 4:00pm
Marco Atzori, Ph.D.
Research Assistant Professor, Blanchette Rockefeller Neuroscience Institute
Blanchette Rockefeller Neuroscience Institute, W.

"Acetylcholine dopamine interactions in the cortical release of glutamate"

The frequency of auditory hallucinations in schizophrenia suggests an involvement of the temporal cortex. In this brain area acetylcholine, released during attentive states, depresses glutamate release through muscarinic receptors. Using patch-clamp recording and applications of the muscarinic agonist oxotremorine in the rat temporal cortex we confirmed the muscarinic depression of the glutamatergic signal, and demonstrated that dopamine impairs the capability of acetylcholine to decrease the release of glutamate. A series of pharmacological experiments suggested that the effect is mediated by dopamine receptor D1R-D2R cooperativity inducing the activation of protein kinase A and the inhibition of phospholipase C. We speculate that a similar impairment in the control mechanisms of glutamate release could unleash the excitation associated with other cholinergic avenues in the temporal cortex of a schizophrenic brain.

Alfredo Kirkwood, Ph.D.

10/13/2003 4:00pm
Christos Constantinidis, PhD
Assistant Professor, Department of Neurobiology and Anatomy, Wake Forest University
Wake Forest University

"Neurophysiological studies of working memory in the primate prefrontal and parietal cortex"

The ability to hold information in memory over a time scale of seconds, known as working memory, is a critical component of higher mental functions such as language, planning and abstract thought. Neurophysiological studies in monkeys have shown that prefrontal cortical neurons continue to discharge after brief presentations of sensory stimuli, thus providing a neural correlate of working memory. In order to understand how the prefrontal cortex performs this operation it is important to examine the physiological properties of prefrontal cortical neurons, the patterns of connectivity between different types of neurons, as well as the dynamics of neuronal discharge. These characteristics are contrasted with the properties of neurons in the posterior parietal cortex and the unique role of each area in the maintenance of memory is discussed.

Steven S. Hsiao, PhD

10/20/2003 4:00pm
John Rinzel, PhD
Professor of Neural Science and Mathematics, NYU
New York University

"Coincidence detection and temporal processing in auditory brain stem"

Distinct biophysical properties including multiple voltage-dependent membrane conductances and well-timed transient inhibition contribute to the temporally precise processing characteristics of auditory neurons. We investigate the underlying mechanisms of coincidence detection through in vitro experiments (gerbil MSO) using dynamic clamp stimuli and with computational models of the Hodgkin-Huxley type. We focus particularly on what makes these neurons fire, i.e. on how they integrate subthreshold signals in the presence of a noisy synaptic (excitatory and inhibitory) background, as is typical in vivo. Consistent with previous reports, the partial blockade of low threshold potassium currents (IKLT) reduced coincidence detection (as well as reduced phase-locking and signal-to-noise ratio). By using spike triggered reverse correlation we showed that blockade of IKLT slowed the rise of Irevcor, indicating a less precise time window for integration. We also suggest, based on our experiments and modeling, that the sodium current (INa) is substantially inactivated at rest and thereby provides a second contributing factor to temporally sharpen integration of subthreshold inputs. These mechanisms of subthreshold negative feedback act to dynamically modulate the spike threshold., creating a brief temporal window for coincidence detection of small signals in noise.

Ernst Niebur, PhD

11/03/2003 4:00pm
Michael Leyton, PhD

Rutgers University

"Shape as Memory Storage"

In a series of books, Dr. Leyton has developed an approach to geometry which argues that shape is equivalent to memory storage. This has involved developing systems of inference rules for the extraction of process-history from shape. The theory is equally applicable for the extraction of history from complex natural shapes such as embryos, tumors, neurons, etc., as well as for the extraction of information from memory stores in computational processes. Since the publication of these inference rules, scientists have applied them in many areas including tumor and organ classification, blood-cell typing, neuronal growth models, embryological limb-formation, meteorology, chemical kinetics, computer vision, linguistics, forensic science, computer-aided design, dental radiography, magnetic resonance imagery, computer graphics, archaeology, human and robot motor control. Information on the speaker is available at

Ed Connor, PhD

11/14/2003 11:00am
Yves Trotter, PhD

Centre de Recherche Cerveau & Cognition, Toulouse

Please note the new time for this seminar


Rudiger von der Heydt, PhD

12/01/2003 4:00pm
Barrie J. Frost, PhD
Queen's University, Department of Psychology
Queen's University

"Neural mechanisms for detecting different types of visual motion"

Single neuron recording studies in pigeons and owl brains will be described to illustrate the subdivision of labour that occurs within the visual system. Neurons in the optic tectum are specially adapted to respond specifically to moving objects, or perhaps more correctly animate motion, while vetoing self produced visual flow fields. The receptive field mechanisms that underlie this specificity also permit figure/ground discrimination through coherent motion. Further along this visual pathway we expect to find neural structures specialized for detecting conspecific motion patterns involved in intra-species communication such as occurs in courtship and dominance hierarchy displays. The tectofugal pathway will be contrasted with the accessory optic system (AOS), a pathway highly conserved in all vertebrate species, that processes large visual flow fields (in vestibular coordinates) while ignoring local object motion. Subsequent binocular integration of these flow fields results in separate populations of neurons sensitive to roll, pitch and yaw rotations, and the three axes of translation. Both behavioural and physiological work on stereopsis and motion parallax processing in the owl wulst will also be presented. Finally, recent work will be described showing that a sub-population of neurons in the tectofugal pathway compute "time to collision" of approaching objects, but are not responsive to self motion toward stationary objects. The significance of Gibsonian conjectures to this research programme will be discussed.

Rudiger von der Heydt, PhD

02/02/2004 4:00pm
Todd W. Troyer, PhD, University of Maryland
Assistant Professor, Department of Psychology, University of Maryland
University of Maryland

Multiple Representations of Song During Vocal Learning

The template hypothesis for birdsong learning first posited by Konishi has served as an organizing framework for the field. However, this hypothesis says relatively little about the nature of the various representations of song and how they interact during learning. I will review some basic models for how these representations might interact during the sensorimotor phase of learning. I will also describe out attempts to use behavioral analysis to address these questions.

Ernst Niebur, PhD

03/01/2004 4:00pm
Mark Hallett, MD, NINDS/NIH


"Human Motor Learning"

Motor learning is defined as “a change in motor performance with practice”. Even considering only the aspect of a change in motor performance, there are at least three different phenomena: classical conditioning, adaptation learning, and skill learning. Classical conditioning is a change in motor output in response to specific sensory stimuli. Motor adaptation learning can be defined as a change in motor performance without a change in motor ability. Motor skill learning can be defined as a change in motor performance with a change in motor ability.

Takashi Yoshioka, PhD

03/15/2004 4:00pm
Alex Reyes, PhD, New York University


"Propagation of Signals in In Vitro Neural Networks"

The manner in which neural signals are represented in networks remains unclear. To examine experimentally how signals are processed, a multilayer feedforward network of neurons was reproduced in an in vitro slice preparation of rat cortex using an interative procedure. When constant and time-varying frequency inputs were delivered to the first layer of the network, the firing of neurons in successive layers became progressively more synchronous. Synchrony, which persisted under a wide range of physiological conditions, is crucial for the stable propagation of input rate signals.

Alfredo Kirkwood, Ph.D.

03/22/2004 4:00pm
David Freedman, Harvard University
Postdoctoral Fellow
Harvard University

Neuronal Mechanisms of Visual Categorization and Object Recognition

The ability to group stimuli into meaningful categories is a fundamental cognitive process. However, little is known about the neuronal mechanisms underlying the acquisition and representation of visual categories. We trained monkeys to classify computer-generated morphed stimuli into two categories, "cats" and "dogs", and conducted simultaneous neurophysiological recordings in the prefrontal cortex (PFC) and inferior temporal cortex (ITC) during performance of a categorization task. We found that the PFC and ITC play distinct roles in category-based behaviors: the ITC seemed more involved in the rapid analysis of visual shape, while the PFC showed stronger category signals

Ed Connor

03/29/2004 4:00pm
Yale E. Cohen, PhD, Dartmouth University
Assistant Professor
Dartmouth University

"Behavioral and neural correlates of semantic processing in rhesus macaques."

Human and non-human primates both produce vocalizations that convey information about objects and events (i.e. food, predators, and social interactions) in the environment. Though the human capacity far exceeds that of other primates, understanding the evolution of this communicative ability is hindered by a lack of data on the underlying neural mechanisms, especially in monkeys and apes. In this talk, we discuss how rhesus monkeys categorize vocalizations based on differences in the valence of their semantic meaning and not wholly on differences in their acoustic structure. Additionally, we show that information about semantic meaning is coded in the spike train of neurons in the ventrolateral prefrontal cortex (vPFC). Specifically, these neurons preferentially code transitions between SSVs that convey different semantic meanings. Thus, the vPFC may represent an important part of the neural circuitry for decoding the meaning of SSVs.

Xiaoqin Wang, PhD

04/05/2004 4:00pm
Jeffrey S. Taube, PhD, Dartmouth University

Dartmouth University

"Which way is that Aquarium? The Neurobiology for a Sense of Direction".

Animals require two types of fundamental information for accurate navigation: location and directional heading. Current theories hypothesize that animals maintain a neural representation, or cognitive map, of external space in the brain. Whereas cells in the rat hippocampus and parahippocampal regions encode information about location, a second type of spatial cell encodes information about the animal's directional heading in the horizontal plane in absolute space, independent of the animal's on-going behaviors. The preferred firing direction of HD cells can be controlled by both landmark cues and idiothetic cues, such as vestibular, proprioceptive, and motor efference copy. Head direction (HD) cells are found throughout the limbic system. Vestibular information is believed critical for the generation of the HD signal because labyrinthectomies abolish the direction-specific firing of these cells. Experiments have also shown that HD cells continue to discharge when the rat is locomoting in the vertical plane, provided the vertical locomotion begins with the rat's orientation corresponding to the cells' preferred firing direction. These cells are ideally suited to enable accurate navigation.

Kechen Zhang, PhD

04/12/2004 4:00pm
Anna Wang Roe, PhD., Vanderbilt University
Associate Professor of Neurobiology,
Vanderbilt University

"Functional organization of SI in the primate: optical imaging's point of view"

Using optical imaging methodology to examine cortical organization of SI in squirrel monkeys, we present new views of an old idea and new ideas about an old view. (1) Previous studies have suggested at least some segregation of modality-specific inputs to Area 3b in the primate. In support of this idea, we find evidence for interdigitated vibrotactile (pressure, flutter, and vibration)preference domains (~200 um in size) in Area 3b and 1. (2) The long-standing view that sensory topography in the somatosensory cortex reflects a 'body map' is well supported. However, by taking advantage of a tactile illusion termed the funneling illusion, we find evidence which suggests that the topographic map in Area 3b is a perceptual map rather than a physical body map.

Steven Hsiao, PhD

04/19/2004 4:00pm
Leonardo Cohen, PhD, NINDS/NIH

"Influence of interhemispheric interactions on motor function in health and disease"

In healthy volunteers, unilateral hand movements are associated with a transient decrease in excitability of the motor cortex that controls the opposite hand. Recent studies showed that corticospinal excitability targeting a resting hand muscle is differentially modulated according to the direction of the intended index finger movements in the opposite hand. These results indicate that inhibitory interactions operating in the process of generation of a skilled unilateral finger movement exhibit kinematic specificity and are more selective than previously thought, possibly to counteract the default production of mirror symmetrical motions. Careful analysis of these interactions may shed some light into abnormalities observed in motor disorders like stroke. Studies of interhemispheric interactions operating in the process of generation of a voluntary movement by the paretic hand in these patients has been done using IHI [1]. IHI was evaluated in both hands preceding the onset of unilateral voluntary index finger movements (paretic hand in patients, right hand in controls) in a simple reaction time (RT) paradigm. Closer to movement onset, IHI targeting the moving index finger turned into facilitation in healthy volunteers but remained deeply inhibited in patients. Deep IHI at movement onset in the patient group correlated with poor motor performance. IHI targeting the resting index finger was comparable in patients and controls. These results document an abnormally high interhemispheric inhibitory drive from M1intact hemisphere to M1lesioned hemisphere in the process of generation of a voluntary movement by the paretic hand in patients with chronic subcortical stroke that correlated with poor motor function. It is conceivable that this abnormality could negatively influenced recovery of motor function in the paretic hand in some patients with stroke, an interpretation consistent with models of interhemispheric competition in motor and sensory systems. Based on this information, possible approaches to ameliorate motor deficits after stroke will be discussed.

Amy Bastian, PhD

04/26/2004 4:00pm
Peter Strick, PhD
Professor, Co-Director CNBC
University of PIttsburgh

"Muscle" and "Movement" Representation in the Motor Cortex: New Anatomical and Physiological Correlates

There has been a long-standing controversy over whether "muscles" or "movements" are represented in the primary motor cortex (M1). We have used novel anatomical and physiological methods to gain a contemporary perspective on this issue. In anatomical experiments, we used retrograde transneuronal transport of rabies virus to map the location of cortico-motoneuronal (CM) cells that project to single hand muscles of macaques. In four animals, we injected rabies virus into either extensor digitorum communis (EDC, an extrinsic hand muscle with multiple insertions, n=2) or adductor pollicis (ADP, an intrinsic muscle with a single insertion, n=2). The survival period (4 - 5.5 days) was set to allow retrograde transport of virus from the injected muscle to the motoneurons that innervate it, and then retrograde transneuronal transport from infected motoneurons to the cortical neurons that innervate them. Standard immunohistochemical procedures were used to demonstrate the location of infected neurons, ETC.

Amy Bastian, PhD

05/03/2004 4:00pm
Donald Geman, PhD, Johns Hopkins University
Dept. of Applied Mathematics and Statistics and Center for Imaging Science Johns Hopkins University
Dept of Mathematical Sciences, JHU


It is unlikely that complex problems in machine perception, such as scene interpretation, will yield directly to improved methods of inductive and statistical learning. Some organizational framework is needed to confront the small amount of data relative to the large number of possible explanations, and to make sure that intensive computation is restricted to genuinely ambiguous regions. I will argue that this applies to natural vision as well, and propose a "twenty questions" approach to pattern recognition. The object of analysis is the computational process itself rather than probability distributions (Bayesian inference) or decision boundaries (statistical learning). Under mild assumptions, optimal strategies exhibit a steady progression from broad scope coupled with low power to high power coupled with dedication to specific explanations. Several theoretical results will be mentioned (joint work with Gilles Blanchard) as well as experiments in machine vision, especially object detection (joint work with Yali Amit and Francois Fleuret).

Ed Connor, PhD

05/17/2004 4:00pm
David Hinkle
A Neuroscience Thesis Seminar
Zanvyl Krieger Mind/Brain Institute

“The Neural Representation of Stereoscopic Disparity Tuning and Three-Dimensional Orientation Tuning in Area V4 of the Macaque Monkey Visual Cortex”


Ed Connor

05/24/2004 4:00pm
Earl Miller, PhD
Massachusetts Institute of Technology, Department of Brain and Cognitive Sciences
Dept Brain & Cog Sci, MIT

"The Prefrontal Cortex: Concepts, Rules, and Cognitive Control"

What controls your thoughts? How do you focus attention? How do you know how to act while dining in a restaurant? This is cognitive control, the ability to organize thought and action around goals. Results from our laboratory have shown that PFC neurons have properties commensurate with a role in "executive" brain function. They are involved in directing attention, in recalling stored memories, predicting reward value, and they integrate the diverse information needed for a given goal. Perhaps most importantly, they transmit acquired knowledge. Their activity reflects learned task contingencies, concepts and rules. In short, they seem to underlie our internal representations of the "rules of the game". This may provide the foundation for the complex behaviour of primates, in whom this structure is most elaborate.

Ed Connor

09/20/2004 4:00pm
Edward Hedgecock, PhD
Professor, JHU, Dept of Biology
Dept of Biology, Johns Hopkins Univ.

"The Game of Reason" - A Unified Theory of Knowledge and Cognition.

Is there a fundamental problem of cognition? Cognitive agents acquire and express knowledge through interactions with co-agents in their environment. Such interactions are organized as complex behavioral plans built from simpler percepts and actions. Modeling co-agents as finite automata, we outline a logical agent who explores and manipulates its environment by planning and playing programs for these automata. From this specific model of a cognitive agent as an inductive theorem-prover, we abstract a positive answer to the original question. To wit, any theory of reasoning, suitably formalized, may reduce to the basic problem of representation and reasoning about maps among finite sets. Finally, we speculate on some neuropsychological consequences of this model.

Steven Hsiao, PhD

09/27/2004 4:00pm (This seminar will not be broadcasted)
Jack Gallant, PhD

Univ of California, Berkeley

"Neuronal mechanisms of attention and working memory in area V4".

Natural vision is governed by two critical factors: natural scenes, which have characteristic statistical structure that is exploited by the visual system to facilitate perception; and natural eye movements, which constrain how the scenes are sampled and processed. My laboratory has simultaneously investigated how these factors influence processing in area V4, an important intermediate stage of visual processing. The task requires search for a natural image patch hidden in a large array of similar patches. Eye movements were permitted during search. Arrays were arranged so that whenever any patch was fixated, the receptive field of a recorded V4 neuron would be centered over a different patch. The overall activity of many V4 cells was modulated by the search target; in some cells modulation was additive, in others it was multiplicative. Fourier-domain turning profiles were also modulated by the search target, revealing top-down influences on spatial frequency and orientation tuning. Finally, the activity of many neurons predicted the direction of subsequent eye movements; the larger the response of the cell, the more likely it was that the next saccade would draw the fovea toward the receptive field. These observations suggest that area V4 acts as a parallel bank of spatially localized multidimensional filters. Their sensitivity and tuning is modulated by top-down influences to optimize performance. During natural vision, filters respond according to their match with the scene and their output, in combination with other areas, is used to drive subsequent eye movements.

Ernst Niebur, PhD

10/04/2004 4:00pm
Kwabena Boahen, PhD

University of Pennsylvania

"From Local Circuits to Cortical Maps: Modeling V1 in silicon".

Dr. Boahen will describe a neuromorphic chip that utilizes transistor heterogeneity, introduced by the fabrication process, to generate orientation maps similar to those imaged in vivo [1]. The cortex's input layer (L4) is modeled as a recurrent network in parallel with a push-pull stage. Similar to a previous model [2], locally connected excitatory and inhibitory cells in the recurrent network display hotspots of activity that give rise to orientation maps—in the absence of any patterned feedforward connections whatsoever. Unlike previous work, however, changing the sign of contrast does not change the orientation map. To anchor the map, cells in the recurrent network are driven by both ON and OFF channels, producing sign-invariant responses. And to recover contrast-sign, cells in the push-pull stage are inhibited by the complemetary channel, producing sign-selective responses. These two groups of cells are similar in their responses to complex and simple cells observed in V1. Thus, the model illustrates how heterogeneity, which is as ubiquitous in biology as it is in silicon, may be exploited to generate sign-invariant feature maps, and proposes a role for complex cells in this process. It makes the prediction that simple cells receive an orientation-selective signal from complex cells, rather than the other way around. [1] P Merolla and K Boahen, “A Recurrent Model of Orientation Maps with Simple and Complex Cells”, Advances in Neural Information Processing Systems 16, S Thrun, L Saul, and B Sholkopf Eds., MIT Press, 2004. [2] U A Ernst et al., Intracortical origin of visual maps. Nat Neurosci, 2001. 4(4): p. 431-6.

Ernst Niebur, PhD

10/18/2004 4:00pm
J. Yiannis Aloimonos, PhD
University of Maryland, College Park
U of MD College Park

"A road map to a silicon visual cortex."

Today, when we think computationally about the process of visual perception, we think mostly in a feedforward sense. We ascribe order to various processes, and some things happen before others. So, for example, 2D vision (image processing) happens before 3D vision. Neurobiologists have argued for quite some time that the operation of the cortex is characterized by substantial amounts of feedback. Computational theorists have assumed that the feedback is for the purpose of optimization and they forgot about it. At the same time, although Computer Vision blossoms as a discipline, we still cannot do very accurate things. Computer Vision sort of works, but not quite. In this talk I will explain why this is so and propose an architecture for vision which resembles a set of overlapping feedback loops. Vision is a chicken-egg problem. We usually think of Vision as a set of modules performing different operations, but, the talk will argue, these modules to work properly need information from several other modules. I will concentrate on the correspondence problem (matching images), which is at the heart of most processes. I will show that correspondence can be solved much more accurately than the state of the art when it is part of a loop that also solves for segmentation and 3D shape. I will show several experimental results for cases where current techniques fail and will present my view of the visual system as a whole, leading to a road map for a silicon visual cortex, a mechanism for integrating visual processes. I will finish with a few hypotheses about the structure and function of the cortex that arise on the basis of purely computational considerations. These hypotheses can be tested with appropriate neurobiological experiments. Yiannis Aloimonos studied Mathematics in Athens, Greece and Computer Science at the University of Rochester, NY (PhD, 1987). He is Professor of Vision and Geometry at the Department of Computer Science in the University of Maryland, College Park, and the director of the Computer Vision Laboratory in the Institute for Advanced Computer Studies. He has been working on active vision and the interplay of vision and action, visual motion analysis, multiple view vision and the integration of visual processes. Web: html

Ernst Niebur, Ph.D.

10/25/2004 11:00am
Ling Qin
University of Yamanshi, Japan

"Spectral-edge sensitivity of primary auditory cortex neurons in alert cats"

Although psychophysical studies have evealed involvement of spectral edges in auditory perception, little is known about neural processing. This study investigates how spectral edges are processed in neurons of alert cat primary-auditory-cortex (A1) with sustained response property. Stimuli are low-pass, high-pass, and band-pass tones with sharp spectral edges whose edge-frequencies were systematically shifted, constructing edge-frequency response functions. Pure- and two-tone stimuli served to delineate excitatory and inhibitory subfields of the frequency response field (FRF). Based on the response function characteristics, cells were divided into edge-sensitive and edge-insensitive cells: the edge sensitive cells had narrow tuning to the high-edge (type-II cells) or low-edge (type-III cells) frequencies, while the edge insensitive cells were driven by any static stimuli with energy on FRF (type-I) or only very narrowband stimuli with energy confined to FRF (type-IV cells). Edge-sensitive cells showed a close correlation between the best frequencies of the single-frequency (BFSF) and edge-frequency (BFEF) response functions and between their half-height bandwidths, suggesting that the edge-frequency identification is processed along the tonotopic axis in A1. BFSF shifted (mean 0.11 octaves) into the stimulus band from the BFEF (closely corresponding to pitch shift into the stimulus band from the edge frequency in human psychophysical data of edge-pitch), suggesting central mechanism of edge-pitch sensation. Type-I cells had non-significant inhibitory subfields of FRF; type-II cells had a significant inhibitory subfield on the higher frequency side; type-III cells, on the lower frequency side; and type-IV cells, on both sides, suggesting that the inhibitory mechanism characterizes the cell-type specific spectral-edge sensitivity.

Rudiger von der Heydt, Ph.D.

10/28/2004 4:00pm
Li Zhaoping, PhD
University College, London England
University College London

“Could V2 mechanisms alone account for its neural tuning to figure border ownership?”

A border between two image regions is normally owned by only one of the regions; determining which one is essential for surface perception and figure-ground segmentation. Border ownership is signaled by some V2 neurons, even though its value depends on information coming from well outside their classical receptive fields. I use a model of V2 to show that this visual area is able to generate the ownership signal by itself, without requiring top-down mechanisms or external explicit labels for figures, T junctions or corners. In the model, neurons have spatially local classical receptive fields, are tuned to orientation, and only receive information (from V1) about the location and orientation of borders. Border ownership signals that model physiological observations arise through finite-range, intra-areal interactions. The model licenses testable predictions.

Rudiger von der Heydt, PhD

11/01/2004 4:00pm
Anitha Pasupathy, PhD
Postdoctoral Fellow


To navigate our complex world, our brains have evolved a sophisticated ability to quickly learn arbitrary rules such as “stop at red”. Studies in monkeys using a laboratory test of this capacity – conditional association learning – have revealed that both frontal lobe structures, including the prefrontal cortex (PFC), and subcortical nuclei of the basal ganglia (BG) are involved in such learning. Both areas exhibit neural correlates of association learning, but whether or not their activity reflects unique mechanisms and/or functions is unclear; they have typically been studied separately with different tasks. To ascertain their relative roles, we studied the neural activity in the PFC and caudate nucleus (Cd) simultaneously, while monkeys learned the associations between two novel visual objects and two saccades. During such learning, neural activity in both areas underwent learning-related changes, albeit at different rates: the Cd showed rapid, almost bistable, changes compared to a slower trend in the PFC that was more in line with slow improvements in behavioural performance. Also, pre-saccadic activity appeared progressively earlier with learning in the striatum, but not the PFC. Neural activity in the frontal eye fields, another frontal lobe structure implicated in the control of eye-movements, underwent relatively little change during the course of learning. These results support the hypothesis that the BG “guides” associative learning in the frontal cortex and provide insights about the underlying mechanisms.

Ed Connor, PhD

11/15/2004 4:00pm
Gabriel Robles-De-La-Torre
Queen's University, Kingston, Canada
Queen's University, Kingston, Canada

“Haptic perception of shape: using real and illusory objects to uncover the role of force and geometry”

Haptic perception involves the active exploration of the environment through the sense of touch. When haptically perceiving the shape of an object, we commonly assume that our perception critically depends on the geometry of the object. For example, when sliding a fingertip along a surface, the spatial trajectory of the fingertip follows the geometry of the surface. It could be easily assumed that such kinesthetic/proprioceptive information is essential to perceive the shape of the surface. In this talk I will discuss how this is not always so. When haptically exploring an object, the contact forces we experience and the geometry of the object are, in general, correlated. It is necessary to experimentally decorrelate these cues to understand how they contribute to perception. This is accomplished by using a haptic interface to create paradoxical haptic objects, in which force and geometrical cues are selectively decorrelated. How do human subjects perceive these paradoxical objects? Regardless of object geometry, subjects use force cues (or correlates such as work) to judge the shape of paradoxical objects. These and other, related findings help illuminate the interaction between hand/limb movements and multiple sensory cues during perception. In particular, I will talk about the perceptual role of force, geometry and movement when actively or passively exploring an object. Finally, I will discuss why some paradoxical objects can be considered as examples of illusory haptic shapes.

Steven Hsiao, Ph.D.

11/29/2004 4:00pm
Emilio Salinas, PhD,Wake Forest School of Medicine

Wake Forest School of Medicine

"Context-dependent selection of visuomotor maps".

Behavior results from the integration of ongoing sensory signals and contextual information in various forms, such as past experience,expectations, current goals, etc. Thus, the response to a specific stimulus, say the ringing of a doorbell, varies depending on whether you are at home or in someone else's house. What is the neural basis of this flexibility? What mechanism is capable of selecting, in a context-dependent way, an adequate response to a given stimulus? One possibility is based on a nonlinear neural representation in which context information regulates the gain of stimulus-evoked responses. I will illustrate this mechanism using simulations of visuomotor tasks,and will discuss some of its properties. In the neural network models I'll present, any one of several possible stimulus-response maps or rules can be selected according to context. The underlying mechanism based on gain modulation is equivalent to switching on or off different subpopulations of neurons. Thus, the contextual cues can quickly turn on or off a sensory-motor map, effectively changing the functional connectivity between inputs and outputs in the network. The model predicts that sensory responses that are nonlinearly modulated by arbitrary context signals should be found in behavioral situations that involve choosing or switching between multiple sensory-motor maps.

Steven Hsiao, PhD

12/13/2004 4:00pm
Lauren Jones

University of Md Medical School

"Coding in the Whisker Somatosensory System"

The ability of rats to use their whiskers to perform fine tactile discrimination rivals that of humans using their fingertips. Rats perform these discriminations rapidly and accurately while palpating the environment with their whiskers. This suggests that whisker-derived inputs produce a robust and reliable code in the whisker-trigeminal system capable of capturing complex, high frequency information. The basis of any such code should be evident in first order neurons in the sensory system, since these will constrain all subsequent processing and coding strategies. Here we demonstrate that all classes of trigeminal ganglion neurons respond with highly reproducible temporal spike patterns to transient stimuli. We show that a single response train recorded from an individual neuron can reliably encode complex whisker deflection patterns, and that this encoding is improved by combining responses from cells with opposite directional preferences. Further, we test whether these stimuli can be reconstructed from neurons in primary (SI) or second (SII) somatosensory cortex.

Steven S. Hsiao, PhD

02/28/2005 4:00pm
Michael Kilgard, PhD

University of Texas Dallas

"Perceptual learning and cortical self-organization"

Sensory cortex is continually reorganized to meet changing behavioral needs. The general principles that allow behaviorally useful plasticity in large populations of neurons remain unclear. In this talk, I will describe a series of experiments that offer new insight into the neural basis of perceptual learning. Different forms of auditory input associated with identical modulatory stimulation can generate dramatic changes in A1 frequency map organization, temporal processing, or sequence selectivity. These experiments support the hypothesis that modulatory neurotransmitters inform cortical neurons which sensory events to learn. Once relevant stimuli have been identified, network-level plasticity rules control how these inputs alter cortical connectivity and dynamics. Understanding how sensory experience guides neural plasticity will be critical for the development of new therapies for neurological rehabilitation.

Alfredo Kirkwood, PhD

03/07/2005 4:00pm
Daeyeol Lee, PhD
University of Rochester
University of Rochester

“Cortical Mechanisms of Reinforcement Learning and Decision Making”

Flexible mapping from a sensory stimulus to a particular action is a hallmark of intelligent behaviors, and this in turns implies that decision making process by which one of often many alternative actions is chosen must be dynamically adjusted through experience. In order to investigate the cortical mechanisms responsible for evaluating the outcome of the animal's choice and optimizing the animal's decision making strategies, we employed a behavioral task modeled after a two-player zero-sum game, known as the matching pennies. By manipulating the exploitative nature of the computer opponent's strategy, we demonstrated that the animal's decision making strategy can be systematically influenced by the strategy of the opponent. In addition, single-unit recordings during this dynamic decision making task showed that many neurons in the lateral and medial prefrontal cortex displayed modulations in their activity related to the past history of the animal's choices and their outcomes. These results suggest that the primate prefrontal cortex plays a key role in optimizing the animal's behavioral strategy in a complex dynamic environment.

Veit Stuphorn, PhD

03/09/2005 3:00pm
Gilles Laurent, PhD - WED., 3/9/05
California Institute of Technology
California Institute of Technology

“Oscillations, circuit dynamics, invariance and cardinal neurons”

Since the seminal 1991 publication on odorant receptors by Buck and Axel, the last decade has seen major developments in our understanding of the molecular and developmental organizational principles of olfactory systems. I will focus here on a systems perspective of olfactory processing: How do neuronal populations operate in olfactory systems? What do they compute? And how? I will show that odor representations in the first olfactory relay station can be understood as population vectors that evolve over time, and that decoding mechanisms in the next station transform these distributed representations into sparse ones, carried by highly specialized cells (“cardinal neurons” or ‘grandmother cells”). I will emphasize the experimental evidence for such phenomena (from electrophysiological experiments in Drosophila, locust and zebrafish) and detail the known mechanisms underlying this transformation, such as oscillatory synchronization, local circuit organization and long-range connectivity.

Ernst Niebur, PhD

03/14/2005 4:00pm
George Ainslie, PhD
Coatsville VA Medical Center
Coatsville VA Medical Center

Does the Will Have a Seat?

Behavioral scientists still speak of the will as if it were an organ that sat, like Descartes’ pineal body, atop a pyramid of lower functions. Recent neuropathological and neurophysiological research has indeed suggested that parts of the prefrontal cortex are sites of will-like functions. However, these localizations have not yet revealed more about the functional properties of the will. I will start with behavioral evidence to argue that the self-control aspect of will depends on intertemporal bargaining in an internal marketplace, the properties of which are predicted by the hyperbolic shape of the discount function with which all organisms devalue delayed rewards. I will explore the implications of this non-organ model of will, and suggest how further research is possible despite this model’s internally recursive nature.

Veit Stuphorn, PhD

03/21/2005 4:00pm
Elizabeth M. Quinlan, PhD
University of Maryland, College Park
University of Maryland, College Park

"Regulation of Ocular Dominance Plasticity in Adult Post-Critical Period Visual Cortex"

Recent demonstrations that receptive field plasticity persists in adults suggest a revision in our understanding of the influence of experience on the mammalian cortex. Depriving one eye of visual experience, by monocular lid suture, induces a shift in ocular dominance (OD) of binocular neurons. During an early, postnatal critical period, monocular deprivation (MD) causes a rapid decrease in the strength of synapses serving the deprived eye, followed by a slower increase in the strength of synapses serving the non-deprived eye. OD shifts can also be induced in adult, post-classical critical period cortex, however longer periods of MD are required. In adults, deprivation engages only the slow component, increasing the strength of synapses serving the non-deprived input. To explore the possibility that ocular dominance plasticity in adults can be enhanced by sensory deprivation, we performed visual deprivation (dark exposure) on adult, post-critical period rodents. When MD follows DE we observe a decrease in the strength of synapses serving the deprived eye and acceleration of the increase in the strength of synapses serving the non-deprived eye. We hypothesized that dark-exposure enables MD plasticity be reducing the level of cortical inhibition, and demonstrate that DE reduces the level of cortical GABAARs. A change in the composition of synaptic NMDARs was also induced by dark exposure, resulting in receptors with a high level of the NR2b subunit relative to NR2a, reminiscent of NMDAR composition in the immature cortex. Our results demonstrate that cortical plasticity in the adult can be regulated by experience. We propose that a decrease in GABAergic inhibition is necessary to enable receptive field plasticity in adult cortex, and that the deprivation-induced regulation of NMDAR composition lowers the threshold for experience-dependent synaptic strengthening

Alfredo Kirkwood, PhD

03/28/2005 4:00pm
Bharathi Jagadeesh, PhD
University of Washington
University of Washington

Learning and "tuning" of neurons in inferior temporal cortex for realistic stimuli.

Neural selectivity in the primate inferotemporal (IT) cortex is thought to underlie the remarkable ability of primates to recognize and discriminate among the plethora of people, places and things that we see. The dimensions of selectivity in IT may be inherently psychological: neurons respond similarly to stimuli because they are deemed to be similar, along dimensions that may include the features of the image, but also include information that the viewer may have about the image as a result of previous experience. This hypothesis means that the "tuning" of IT neurons depends both on the physical characteristics of the image and the behavioral processing of the image. We have tested this hypothesis by examining both the relationship between IT neurons and both the physical characteristic of visual stimuli and the behavioral processing of them in demanding visual discrimination tasks.

Ed Connor, PhD

04/04/2005 4:00pm
Max Riesenhuber, PhD
Georgetown University
Georgetown University

"A simple model of human face recognition"

Understanding how objects are represented in cortex is one of the key and much debated questions in neuroscience. A large body of physiological evidence argues for a hierarchical shape-based organization of visual processing in the spirit of Hubel and Wiesel’s original simple-to-complex model. However, such a model appears to be incompatible with data from human behavior and fMRI that have been interpreted as requiring more complex schemes in which the recognition of some object classes, in particular faces, or the recognition of objects of expertise is based on special computational mechanisms. I will show how a simple shape-based hierarchical model can quantitatively (and parsimoniously) account for experimental data adduced to support a special status of faces, such as the Face Inversion Effect, the existence of a "face area" in fMRI, and data from prosopagnosic patients, without requiring any additional face-specific or expertise-specific neural mechanisms. I will present psychophysical results that argue against more complex theories of face representation, and I will finally present preliminary data from human behavior and fMRI that demonstrate the quantitative link between physiology, fMRI BOLD response in the "fusiform face area", and human face discrimination performance predicted by the model.

Ed Connor, PhD

04/11/2005 4:00pm
Judith Hirsch, PhD
University of Southern California

"Common patterns of organization for the synaptic inputs that build receptive fields in the visual thalamus and cortex".

Neurons in the lateral geniculate nucleus of the thalamus, like the retinal ganglion cells that supply them, are able to detect simple properties of visual signals such as spatial location. At the next stage of processing, layer 4 of the striate cortex, neurons become able to resolve sophisticated elements of the stimulus like orientation. To learn how neural circuits in the geniculostriate pathway resolve different features, we combine the techniques of whole-cell recording in vivo, intracellular labeling and quantitative receptive-field mapping. Our recent studies show that excitatory and inhibitory circuits in both the thalamus and cortical layer 4 share remarkably similar rules of organization. Thus, successive stages of the early visual pathway may use common strategies to analyze different aspects of sensory information

Rudiger von der Heydt, PhD

05/09/2005 4:00pm
Gregory Ball, PhD, Johns Hopkins University


“Seasonal Neuroplasticity in Songbirds”

The vocal control system of songbirds exhibits many properties that show a clear relationship between brain variation and behavioral variation. For example, the degree to which there is a sex differences in vocal production among various species is clearly related to the degree to which there is a sex difference in the volume of key song control nuclei. Such relationships are also apparent if you examine seasonal variation in brain and behavior. These seasonal changes in the vocal control system of songbirds are one of the most dramatic examples of naturally occurring adult neuroplasticity. In male European starlings for example, the volumes of telencephalic nuclei that control song such as HVC are substantially larger in volume in the spring than in the fall. Long daylengths promote testis growth and the concomitant increase in plasma testosterone that clearly can stimulate increases in HVC volume. Testosterone is most effects when birds are photosensitive (i.e. responsive to long days characteristic of spring) as opposed to photorefractory (non-responsive to long days characteristic of late summers). Long photoperiods can also stimulate HVC growth even in castrated birds. Thus testosterone interacts with other factors in the regulation of this seasonal neuroplasticity. Testosterone treatment increases HVC size in part by inducing the expression of brain-derived neurotrophic factor (BDNF). Interestingly, singing behavior itself promotes the release of BDNF in HVC independently of testosterone. We investigated the importance of such activity-dependent release of BDNF in testosterone-treated castrated birds by manipulating song rate with central lesions to the preoptic area (blocks the motivation to sing) and with peripheral lesions to the syrinx (the vocal production organ). In both cases preventing song production significantly decreases HVC volume, even in the presence of testosterone. Seasonal changes in HVC volume involves interrelations among variation in photoperiod, endogenous testosterone and behavioral responses to such variables.

Takashi Yoshioka, Ph.D.

05/16/2005 4:00pm
Arun Sripati, Graduate Student, JHU
Graduate Student, JHU
Johns Hopkins University

"A continuum mechanical model for mechanoreceptive afferent responses to indented stimuli."

Our understanding of vision begins with understanding the optics of the eye and the spectral absorption characteristics of rods and cones. A comparable understanding of touch requires that we understand the mechanics of the skin and mechanotransduction. To that end, we recorded the responses of slowly adapting (SA1) and rapidly adapting (RA) mechanoreceptive afferents to stimuli that varied widely in their spatial configurations. We then developed an analytical framework to characterize the chain of events that culminates in the neural response. Within this framework, we arrived at a model that reliably predicts the afferent responses from the local maximum tensile strain in the skin. The close agreement between model predictions and data supports the hypothesis that local maximum tensile strain drives mechanotransduction. This is work with Sliman Bensmaia and Kenneth Johnson at the Krieger Mind/Brain Institute.

Rudiger von der Heydt, Ph.D.

08/01/2005 4:00pm
Romana R. Dodla, PhD

Center for Neural Science, NYU

"Postinhibitory facilitation: A new inhibition induced excitability"

Inhibition has conventionally been viewed as a suppressor of spike probability of a neuron. But we know that inhibition was recognized as a spike facilitator about 50 years ago in the form of postinhibitory rebound. Because of a renewed interest on the role of inhibition both in auditory and other pathways, cortical circuitry as well as in network simulations, we would like to know if there is more to inhibition than there is to its name. I'm going to describe an inhibition-induced facilitatory mechanism that can (1) induce a spike in a single cell, (2) enhance the firing rate of a neuron that receives random inputs, (3) increase the spike response of a recurrent inhibitory neuron, and (4) help synchronize a network of mutually inhibiting neurons with a firing rate that is much bigger than both individual neuron's firing frequency and the background firing level. Presentation will provide detail of basic mechanisms responsible for this, what we term, postinhibitory facilitation.

Ernst Niebur, PhD

08/15/2005 4:00pm
Raphael Pinaud, PhD

The Vollum Institute

"Contributions of Inhibitory transmission to birdsong auditory processing"

The telencephalic caudomedial nidopallium (NCM) participates in auditory processing of conspecific vocalizations in songbirds. Almost 50% of NCM neurons are GABA-positive, suggesting a major role for inhibition in network dynamics, and a possible GABAergic influence on auditory response properties of NCM neurons. In this presentation I will discuss the anatomical-functional organization of the GABAergic system in the songbird NCM, focusing in three sets of studies: First we cloned a zebra finch homologue of the gene encoding the 65 kD isoform of glutamic-acid decarboxylase (GAD-65), a specific GABAergic marker, and conducted an expression analysis by in-situ hybridization to identify and to map the brain distribution of GABAergic cells. The results showed that the auditory telencephalic areas field L2, NCM and the caudomedial mesopallium (CMM) contain an unusually high number of GABAergic cells. We confirmed this finding with immunocytochemistry (ICC) using an anti-GABA antibody. Next, using double fluorescence in-situ hybridization and double-immunocytochemical labeling, we demonstrated that large numbers of GABAergic cells in NCM and CMM show inducible expression of the transcription regulator zenk in response to song auditory stimulation and are thus, song-responsive. Second, whole-cell patch clamp recordings from NCM neurons in a slice preparation were used to explore the relative contributions of inhibitory and excitatory inputs to spontaneous activity. Application of 50uM bicuculline methiodide (BMI), a competitive antagonist for GABA-A receptors, abolished IPSCs and sIPSCs, accompanied by dramatic increases in activity. We also found that both the frequency and amplitude of sIPSCs are regulated by excitatory input in NCM. These results show that tonic inhibition via GABA-A transmission plays a significant role in modulating excitability of NCM neurons. Third, multi-electrode extracellular recordings were combined with local injections of BMI in NCM of awake, restrained adult zebra finches during song playback to study the role of inhibition in the patterning of auditory responses. Simultaneous recordings were made from multiple NCM sites bilaterally, both before and after pressure injection of BMI into the right hemisphere at a dose just below threshold for local seizure-like activity. Before injection, song stimuli elicited responses in both hemispheres consisting of phasic bursts to most syllables and sustained firing between syllables that continued after the end of the stimulus. After BMI application, the right hemisphere showed changes in firing pattern: 1) the phasic responses increased dramatically, especially to the first song syllable, and 2) the sustained firing between syllables was largely abolished; however, the overall response magnitude was not changed. These results suggest that GABA-A transmission contributes not only to excitability but also to the temporal dynamics of auditory responses in NCM.

Alfredo Kirkwood, PhD

08/29/2005 4:00pm
Liam Paninski, PhD
University College London

“Statistical methods for understanding neural codes”

The neural coding problem --- deciding which stimuli will cause a given neuron to spike, and with what probability --- is a fundamental question in systems neuroscience. The high dimensionality of both stimuli and spike trains has spurred the development of a number of sophisticated statistical techniques for learning the neural code from finite experimental data. In particular, modeling approaches based on maximum likelihood have proven to be flexible and powerful. We present three such applications here. One common thread is that the models we have chosen for these data each have concave loglikelihood surfaces, permitting tractable fitting (by maximizing the loglikelihood) even in high dimensional parameter spaces, since no local maxima can exist for the optimizer to get ``stuck'' in. First we describe neural encoding models in which a linear stimulus filtering stage is followed by a noisy integrate-and-fire spike generation mechanism incorporating after-spike currents and spike-dependent conductance modulations. This model provides a biophysically more realistic alternative to models based on Poisson (memoryless) spike generation, and can effectively reproduce a variety of spiking behaviors. We use this model to analyze extracellular data from populations of retinal ganglion cells, simultaneously recorded during stimulation with dynamic light stimuli. Here the model provides insight into the biophysical factors underlying the reliability of these neurons' spiking responses, and provides a framework for analyzing the cross-correlations observed between these cells. (Joint work with E.J. Chichilnisky, J. Pillow, J. Shlens, E. Simoncelli, and V. Uzzell, at NYU and UCSD.) Next we describe how to use this model to ``decode'' the underlying subthreshold somatic voltage dynamics, given only the superthreshold spike train. We also point out some connections to spike-triggered averaging techniques. We close by discussing recent extensions to highly biophysically-detailed, conductance-based models, which have the potential to allow us to estimate the density of active channels in a cell's membrane and also to decode the synaptic input to the cell as a function of time. (With M. Ahrens and Q. Huys, GCNU.)

Joshua Vogelstein, Graduate Student

09/19/2005 4:00pm
Edward M. Callaway, PhD
Salk Institute for Biological Studies
Salk Institute for Biological Studies

"Parallel Pathways and Local Circuits in Visual Cortex"

We have studied primary visual cortex (V1) to better understand how neural circuits give rise to perception. We have found that cortical circuits are extremely precise, such that different neuron types with overlapping dendrites, and even neighboring neurons of the same type, are connected differently. This fine-scale and cell type-specific organization implies that studies of relationships between circuits and function should match this level of organization. We have exploited differences in the laminar projections within V1 of different types LGN neurons to correlate these cell types with their functional properties. We find that red-green and blue-yellow color opponent LGN neurons comprise parallel pathways that project to different layers of V1, and neurons with blue-on versus blue-off receptive fields project to distinct zones. To test hypotheses about contributions of specific cell types to neuronal responses and to perception, we have developed methods to allow reversible inactivation of selected cell types in primates. We find that expression of an insect neuropeptide receptor which couples to GIRK channels can be used to selectively, quickly, and reversibly eliminate the activity of LGN or cortical neurons in vivo. Future application of these methods will allow in vivo tests of the role of particular cell types within the functioning cortical network.

Joshua Vogelstein, Graduate Student

09/26/2005 4:00pm
Roger D. Traub, MD
SUNY Downstate Medical Center, NY
Health Science Center

"Network synchronization via recurrent synapses and via electrical coupling, in hippocampal and neocortical cortices: simulation and electrophysiological studies"

Very fast field oscillations (>70 Hz) occur in a number of contexts: superimposed upon physiological and epileptiform sharp waves, prior to focal seizure discharges, and in short bursts during persistent gamma (30-70 Hz) oscillations. Experimental data indicate that gap junctions are necessary for such very fast oscillations, and the requisite gap junctions most likely couple the proximal axons of principal neurons. Electrical coupling between principal neurons also facilitates the synchronization of neurons produced by recurrent synaptic excitation, when synaptic inhibition is weakened. A prediction of these ideas is that physically isolated axonal networks can generate very fast oscillations, and in vitro experiments are consistent. In neocortex in vitro, runs of synchronized bursts occur at frequencies of 10 Hz and above, the electrographic correlate of EEG polyspikes and tonic seizures; such runs are facilitated by an additional factor besides gap junctions: strong recurrent synaptic connections between layer 4 spiny stellate neurons.

Ernst Niebur, PhD

10/03/2005 4:00pm
Lena Ting, PhD
Biomedical Eng., Emory Univ.

"Dimensional Reduction of Spatial and Temporal Patterns of Muscle Activity for Postural Control"

Maintaining standing balance is an active process that requires that the nervous system coordinate the activity of multiple muscles across both space and time. Our work is focused on developing quantitative analyses and models that will allow us to interpret and predict the functional role of spatiotemporal patterns of muscle activation during normal and impaired postural control. Understanding the fundamental organizing principles underlying spatial and temporal patterns of muscle activation and their functional consequences is particularly critical to developing and evaluating strategies to restore movement. I will demonstrate how high-dimensional spatial and temporal patterns of muscle activation across multiple muscles can be specified by low-dimensional neural commands to muscle synergies for control of task-related variables. First, a non-negative factorization technique has allowed us to extract functional muscle synergies that can account for spatial coordination of muscle across various postural perturbation tasks. These muscle synergies may reflect a neural command signal that specifies endpoint force of a limb, and they can be used to robustly describe muscle coordination patterns and forces for balance control. Next, temporal features of muscle activation patterns can be predicted using a simple feedback loop on center of mass dynamics. The entire timecourse of the EMG activity as well as the motion of the center of mass can be predicted and simulated based on the acceleration, velocity, and position of the center of mass. Moreover, the changes in temporal muscle activation patterns and the resulting instability due to loss of group I afferents induced by pyridoxine intoxication can also be predicted. Taken together, these results suggest that just a few parameters can be used on characterize complex spatiotemporal patterns of muscle activation during postural responses. Complex patterns of muscle activation can be reduced in dimension spatially by muscle synergies and temporally by feedback loops that incorporate task-level, rather than local variables. Thus, it appears that the tendency of the nervous system to encode task-level variables is reflected in the limited dimension of the motor output. Our results will allow us to develop quantitative diagnostic tools for balance and movement disorders and to facilitate the design of effective interventional therapies, neural prostheses, and neural repair strategies for motor rehabilitation. Acknowledgements Whitaker RG-02-0747 and NIH HD46922

Amy Bastian, PhD

10/24/2005 4:00pm
John E. Lisman, PhD
Brandeis Univ., MA

Why does the brain oscillate? Role of theta/gamma coding

It has long been known that the brain shows multiple oscillatory rhythms, but their functional role remains unclear. In the hippocampus there is now general agreement that both theta (4-10Hz) and gamma (30-100Hz) oscillations occur. Indeed, the theta modulation of gamma amplitude indicates that the oscillations co-occur and inter-relate. The theta phase precession of hippocampal place cells was discovered by O’Keefe; this and other recent work establishes that the hippocampus uses a neural code based on theta phase. Because the firing of cells is also controlled by gamma (firing occurs preferentially on a given phase of gamma), the coding is best described as a discrete theta/gamma phase code. Theoretical work from several laboratories suggests that the phase precession can be interpreted as the cued recall from hippocampal memory of the sequence of upcoming places. Recent work suggests that theta/gamma coding may not be confined to the hippocampus, but is a more general coding scheme. Intracranial recording in humans has shown that theta can be recorded in widespread regions of cortex. Moreover, at many sites theta power is gated by working memory: power increases at the onset of a working memory task and stays elevated until the memory is no longer needed. Work from other laboratories shows evidence for theta/gamma oscillations in rat olfactory cortex. Most recently, single unit recording in monkey V4 (G. Rainer) shows that spiking during a working memory task occurs preferentially on certain phases of theta. Based on these lines of evidence it is suggested that theta/gamma oscillations are the clocking system for a general coding strategy for the brain. The cells that fire within a gamma cycle form a spatial code that represents a unit of information. The groups of cells that fire in sequential gamma cycles are an ordered set with different discrete theta phase. This order can be used to encode sequential expected places in the readout from long-term memory, sequentially presented items held in working memory, or sequential percepts within sensory structures. One role of gamma, which is due to the oscillatory nature of inhibition, is to synchronize cells, thereby promoting their detection by downstream networks. A second and overlooked role is to control which cells fire; the negative feedback of oscillatory inhibition promotes a winner-take-all function in cortical networks that appears to be critically important in many brain compu

Ernst Niebur, PhD

10/31/2005 4:00pm
Dana Ballard, PhD
University of Rochester
University of Rochester

Is the Cortex a Digital Computer?

Decades of single-cell recordings in mammalian cortex have revealed the correlation of increased firing rate with behavioral measures, suggestive of a rate-code. But is a rate code the fundamental signalling strategy of the cortex, or is it just a correlate of another quite different strategy? The two possible answers to this question both have ardent proponents. Defenders of the rate code as the basic signaling strategy point to the huge number of experimental studies that use rate code as a basic framework. In addition they argue that the correlations between neuros that are above the number predicted by chance are too small for any synchronous strategy to be a useful phenomena. Proponents of the idea of a more basic code based on spike timing point to the increasing number of data that show spike timing effects. There have been a wide range of models suggesting very different uses of timing, such as signal binding, fast computation, signaling the result of a computation and even consciousness itself. We argue that the cortex has adopted a signaling strategy that makes extensive use of synchrony for fast communication, but does it in a way that is consistent with the rate-code indications. The crux of the argument is that the cortex has to solve a handful of basic computational problems and that these can be expressed as constrants. When these constrants are intersected, one model that fits is that of a handful of simultaneously active digital feedback circuits. This is a very non-standard way of thinking about cortical computation that fits the data.

Ernst Niebur, PhD

11/04/2005 9:00am
Jeff Hawkins, Numenta
New time: 9:30 am
Redwood Neuroscience Institute

“Understanding the Neocortex as hierarchically organized temporal memory”

In his recent book titled "On Intelligence", Hawkins proposed that the neocortex can be understood as a hierarchical sequence memory. Since the book was written the theory has been mathematically formalized as a modified version of belief propagation. Prototype implementations can solve previously intractable vision inference problems but the theory relates to all perceptual tasks performed by cortex. In April 2005 Hawkins formed a new company, Numenta, to promote and develop this technology. Hawkins will describe the basics of the theory, show some of its biological derivations, demonstrate a working prototype, and discuss the potential impact of this theory on both technology and neuroscience.

Ernst Niebur

11/10/2005 4:00pm
Ichiro Fujita, PhD
Osaka University, Japan

Stereopsis and the ventral visual pathway

Horizontal binocular disparity, an important visual cue for the perception of depth and 3-D scene, was widely believed to be processed mainly in the occipitoparietal or dorsal visual pathway. Recent studies, however, show that a large population of neurons in cortical areas in the monkey occipitotemporal or ventral visual pathway, such as area V4 and the inferior temporal cortex (IT), are selective for binocular disparity. We addressed how binocular disparity signals are transformed along the ventral pathway, and whether neurons in V4 and IT differ from neurons in dorsal pathway areas such as MT and MST in terms of the processing of binocular disparity. We found that (1) when a random-dot stereogram (RDS) is contrast-reversed between the left and right eye images, the majority of disparity-selective V4 neurons attenuate their disparity selectivity, a decrease in sensitivity also reflected in depth perception; (2) a large portion of V4 neurons encode relative disparity between two surfaces in a visual stimulus; (3) trial-to-trial fluctuations of IT-neuron responses to a given stimulus correlate with the animal’s behavioral report of fine-depth judgment. These results indicate that activity in V4 and IT surpasses the local filter-like processing of V1, and correlates more with binocular depth perception than that in areas MT and MST. We suggest that V4 and IT contribute to fine-grade disparity discrimination. Two recent papers Tanabe, S., Umeda, K., Fujita,I. (2004) Rejection of false-matches for binocular correspondence in macaque visual cortical area V4. J. Neurosci. 24: 8170-8180. Uka, T., Tanabe, S., Watanabe, M., Fujita, I. (2005) Neural correlates of fine depth discrimination in monkey inferior temporal cortex. J. Neurosci., in press. Biosketch 1979 University of Tokyo (Biology, B.A.) 1984 University of Tokyo (Zoology, Ph.D) 1984-1987 National Institute for Physiological Sciences, Okazaki 1987-1989 California Institute of Technology (Mark Konishi lab) 1989-1994 RIKEN 1994 Professor, Osaka University Medical School 2002 Professor, Osaka University Graduate School of Frontier Biosciences

Ed Connor, PhD

11/17/2005 4:00pm
Alessandra Angelucci, MD., PhD

University of Utah, John A Moran Eye Center

"The contribution of feedforward, lateral and feedback connections to the classical receptive field center and extra-classical receptive field surround of primate V1 neurons"

The contribution of feedforward, lateral and feedback connections to the classical receptive field center and extra-classical receptive field surround of primate V1 neurons. A central question in visual neuroscience is to identify the circuits and mechanisms that generate the response properties of visual cortical cells. V1 cells respond best to oriented stimuli of optimal size within their receptive field (RF). This size tuning is contrast-dependent, i.e. a neuron’s optimal stimulus size is larger at low contrast. Stimuli involving the extra-classical surround suppress the RF center’s response. Surround suppression is fast and long range, extending well beyond a V1 neuron’s optimal stimulus size at low contrast (or low contrast summation RF, lsRF). To determine the contribution of feedforward, lateral and inter-areal feedback connections to V1 to the RF center and surround of V1 neurons, we have quantitatively compared the spatial extent of these connections with the size of V1 neurons’ RF center and surround. We find that FF afferents to V1 are coextensive with the size of V1 cell’s high contrast spatial summation field and can, thus, integrate signals within this RF region. V1 lateral connections are commensurate with the size of the lsRF and may, thus, underlie contrast-dependent changes in spatial summation, and modulatory effects from the “near” surround (i.e. the region of the surround close to the RF center). Contrary to common beliefs, the spatial and temporal properties of lateral connections cannot fully account for the dimensions and onset latency of surround suppression. Inter-areal feedback connections to V1, instead, are fast, and their spatial scale is commensurate with the full spatial range of center-surround interactions in V1. Feedback connections terminate in a patchy fashion in V1, and show modular and orientation specificity, and spatial anisotropy collinear with the orientation preference of their cells of origin. Thus, the spatial and functional organization of feedback connections is also consistent with their role in orientation-specific center-surround interactions. I will present a biologically-based recurrent network model, demonstrating how center-surround interactions in V1 neurons can arise from the integration of inputs from feedforward, lateral and feedback connections. I will show physiological data in support of the model’s predictions, revealing that modulation from the “far” surround is not always suppressive

Rudiger von der Heydt, PhD

11/18/2005 4:00pm
Hidehiko Komatsu, PhD

National Institute of Physiological Sciences, Japa

"Effect of task demand on the color selective neural responses in the inferior temporal cortex of the monkey"

Discrimination and categorization are two different aspects of our color vision. On the one hand, we can discriminate subtle difference in color; on the other hand, we categorize similar colors into a single group such as 'red'. To study the effects of these different aspects of color vision on the color coding in the inferior temporal cortex, we recorded neuronal responses from two Japanese monkeys trained a color categorization task and a color discrimination task. In both tasks, while the monkey was gazing at the fixation spot, a test stimulus was presented on the fovea. In color categorization task, fixation point and saccade point were shown after presentation of a test stimulus. Monkeys were trained to keep fixation when reddish color was presented and to make saccade, when greenish color was presented as test. In color discrimination task, two color stimuli (discrimination stimuli) were shown simultaneously after presentation of a test color. Monkeys were trained to make saccade toward the same color as the test. In both tasks, we used eleven test colors that were arranged at equal intervals from red to green. Monkeys were also trained a simple visual fixation task. Neurons were recorded from a sub-region in the anterior part of the inferior temporal cortex where neurons have strong color selectivity, but little shape selectivity. We found there are neurons that exhibit difference in the response magnitude between these two tasks. The difference could be explained by the change in the response gain; the effective color and the tuning width did not change. Categorical decision or motor command could not explain the response change. In several extreme cases, large gain change almost diminished the responses in one task. In some neurons, activity change occurred preceding the onset of the stimulus. These results suggest that color signals are gated by the top-down signal representing task demand in the inferior temporal cortex. I will also present some results of our recent experiment on the regional specialization of color signals in the inferior temporal cortex.

Ed Connor, PhD

11/21/2005 1:30pm
Stefan Mihalas,1:30-3PM

California Institute of Technology

"Modeling CaMKII Activation in Dendritic Spines"

We are interested in constructing a kinetic model of biochemical signaling in the glutamatergic spine in response to $Ca^{2+}$ influx produced by electrical activity. $Ca^{2+}$/calmodulin-dependent protein kinase II (CaMKII) is a principal target of $Ca^{2+}$ entering through NMDA receptors. It can be activated by $Ca^{2+}$ transients as short as 100ms and its activity lasts for minutes-hours, thus filing the gap between the time scale of the electrical activity and the time scale of LTP. We expanded a model for CaMKII activation to be accurate in low $[Ca^{2+}]$ and experimentally measured most of its parameters. To make predictions on CaMKII activation during physiologically relevant stimulations, we combined the model for CaMKII activation with a simple model for $Ca^{2+}$ dynamics in a spine. We plan to expand the biological realism of the simulations and to include more elements of the biochemical signaling.

Ernst Niebur

12/05/2005 4:00pm
Emery Brown, MD.,PhD

"Application of the State-Space Model Paradigm to Neural Data Analysis"

Neural systems encode representations of relevant biological signals in the firing patterns of their spike trains. Spike trains are point process time-series and their codes are both dynamic and stochastic. Even though the signal is often continuous, its representation in the nervous systems is as a high-dimensional point process time-series. Because neural spike trains are point processes, standard signal processing techniques for continuous-valued data will have limited utility in the analysis of neural systems. Accurate processing of neural signals requires the development of quantitative techniques to characterize correctly the point process nature of neural encoding. The advent in the last 10 years of the capability to record with multiple electrode arrays the simultaneous spiking activity of many neurons (>100) has made it possible to study information encoding by ensembles rather than by simply single neurons. Hence, an important question in neuroscience is developing algorithms to analyze dynamic, high-dimensional spike train (point process) measurements. The state-space modeling paradigm is a well-known engineering framework for studying systems that evolve through time. In this presentation, we will discuss the application of this paradigm in neural spike train data analysis. We use the Baye’s rule, Chapman-Kolmogorov equations to derive algorithms useful for neural spike train decoding, dynamic analysis of neural encoding (neural plasticity) and adaptive-decoding. We will illustrate the methods in three examples: Decoding position from the ensemble activity hippocampal pyramidal neurons and tracking the temporal evolution in hippocampal place receptive fields, and decoding motor cortex representations of movement velocity.

Joshua Vogelstein, Graduate Student

02/06/2006 4:00pm
Kirk Thompson, PhD
"Neural basis of spatial attention"

"Neural basis of spatial attention".

The premotor theory of attention proposes that developing oculomotor commands mediate covert spatial attention. Others have proposed the existence of a visual salience map somewhere in the brain that identifies behaviorally important locations to guide attention and eye movements. I will describe neurophysiological results in behaving monkeys showing that the frontal eye field contains a visual salience map for guiding eye movements and is involved in the allocation of spatial attention without eye movements.

Veit Stuphorn, PhD

02/13/2006 4:00pm
Eero Simoncelli, PhD

Howard Hughes Medical Institute



Rudiger von der Heydt, PhD

02/27/2006 4:00pm
Michael Goldberg, MD
Columbia University

"On the agnosticism of spikes: attention, intention, and salience in the monkey lateral intraparietal area"

Attention is the process whereby the brain filters out sensory information unimportant for behavior. Clinical studies show that the parietal lobe is important for the attentional processes. Neurons in the lateral intraparietal area (LIP) filter out visual stimuli that are behaviorally unimportant , for example stable objects in the environment, although they do respond to those same stimuli when they appear abruptly in the environment. Although LIP filters out behaviorally irrelevant visual stimuli, it does not filter out salient objects that are not the targets of a planned saccade. When a monkey plans a memory-guided saccade away from the receptive field of a neuron, the abrupt onset of a distractor in the receptive field evokes an enhanced response relative to the case when the monkey plans a saccade to the receptive field and the distractor subsequently appears in the receptive field. In these cases the distractor had no effect on the performance of the saccade. Attention, as measured by an improvement in contrast sensitivity at the attentionally advantaged site, lies at the goal of a memory-guided saccade during the delay period, but it can be briefly captured by the abrupt onset of a distractor. The activity of neurons in LIP correlates with the monkey’s attention both when it lies at the saccade goal and when it lies at the distractor site, and the time at which attention returns from the distractor to the saccade goal is predicted by the activity of neurons in LIP. Most studies of eye movements in awake, behaving monkeys demand that the animal make specific eye movements. We have developed a new paradigm in which the monkey performs a visual search for an upright or inverted T among 7, 11, or 15 cross distractors, and reports the orientation of the distractor with a hand movement. The search array is radially symmetric around a fixation point, but once the array appears the monkey is free to move its eyes. The monkey’s performance in this task resembles that of humans in similar tasks (Treisman and Gelade, 1980) : manual reaction time shows a set size effect for difficult searches (the crosses resemble the T’s) but not for easy searches (the T pops out). Saccades are made almost exclusively to objects in the array, and not to intermediate positions, but fewer than half of the initial saccades are made to the T. We recorded from neurons in the lateral intraparietal area (LIP) while the monkey performed the search. LIP neurons distinguish the saccade goal at an average of 86 ms after the appearance of the array. The time at which neurons distinguish saccade direction correlates with the monkey’s saccadic reaction time, suggesting that most of the jitter in reaction time for free eye movements comes from the discrimination process reflected in LIP. However, they also distinguish the T from a distractor on an average of 111 ms after the appearance of the array even when the monkey makes a saccade away from the target, suggesting that LIP has access to cognitive information about the target independent of the saccade choice. We suggest that LIP provides a salience map that can be used by multiple systems. The salience map is constructed from independent signals (visual, cognitive, saccadic) which are summed in a linear fashion. When a saccade is appropriate, the oculomotor system can use the peak of the salience map to drive a saccade. The visual system uses the same spikes to determine the locus of attention The source of the spikes, whether from a saccade plan or the visual system reporting the abrupt onset of a visual stimulus, is irrelevant to the use to which the recipient area puts the salience map.

Veit Stuphorn, PhD

03/06/2006 4:00pm
Patrick Cavanagh, PhD
Harvard University

"Coordinates of Attention"

Dr. Cavanagh will describe three methods to identify the coordinate frames for attentional processing: retinotopic, head-based, and position-independent. These coordinate properties identify plausible physiological sites for different components of attention. 1) Retinotopic coordinates: tracking multiple objects is limited independently in the left and right hemifields. This pattern is also seen with parietal patients who fail the tracking task only in the contralesional hemifield. Results so far show retinotopic, hemifield limits for location?based attentive processing, suggesting cortical sites in the dorsal stream. 2) Head-based coordinates: When the direction of gaze carries a retinal phosphene outside the visual field, it vanishes, despite the continued stimulation of the retina (delivered to a fixed location through the sclera). At some level, cortical activity is cropped at the edge of the visual field, reflecting the transformation from retinal to head-based coordinates. We assume that attentional selection is controlled from a head-based coordinate representation (again suggesting a dorsal site) that cannot select information lying outside its range. 3) Position independent: When an object moves across the retina, its features may have moved on before the local computations (based on retinotopic receptive fields) can complete their analysis. If percepts of moving stimuli remain well organized, then there must be position-independent computations that are accumulating partial results from each location. Our tests distinguish early, local computations that fail for moving stimuli, from later, object-based computations that are preserved. This mobile computation appears to require attention to the target and suggests non-retinotopic cortical sites in the ventral stream (human LOC, FFA, monkey IT). Our three methods link attentive processes to distinctive anatomical signatures: hemifield and visual field limited processing for early selection but position independent processing for mobile computation.

Rudiger von der Heydt, PhD

03/13/2006 4:00pm
Gregory DeAngelis, PhD
Washington University School of Medicine

"Perceiving and navigating through 3D space: neural mechanisms of depth and heading perception in the dorsal visual processing stream"

Research in my laboratory focuses on the neural circuits responsible for perceiving the location of objects in 3D space and for computing the direction of one's self-motion through 3D space. I will provide an overview of three ongoing projects in the laboratory. First, I will describe reversible inactivation experiments that probe the roles of area MT in two depth discrimination tasks. I will show that MT contributes to coarse depth discrimination but not to fine discrimination. In addition, I will show that training animals to perform the fine depth discrimination task profoundly alters the contribution that MT makes to the coarse depth task, suggesting remarkable training-induced plasticity. Second, I will describe a set of ongoing experiments that test whether MT neurons are involved in computing depth from motion parallax, the relative motion of objects that frequently results from self-motion of the observer. I will show that MT neurons combine retinal motion information with extra-retinal signals to compute depth sign from motion parallax. These data provide the first compelling evidence for a neural mechanism of depth perception that relies on motion parallax. Third, I will introduce a set of experiments that explore how visual motion (optic flow) is combined with vestibular signals to compute direction of heading. I will show that monkeys combine optic flow and vestibular signals near optimally in a heading discrimination task, and that a subset of neurons in area MSTd also integrate these two sensory inputs to achieve higher sensitivity under cue combination.

Ed Connor, PhD

03/27/2006 4:00pm
Christopher Moore, PhD

"New Principles in an Old Brain: Cortical Dynamics and Novel Feature Maps in Rodent Barrel Cortex"

The primary interest of our lab is perception, and specifically how rapid changes in neural organization ('dynamics') underlie rapid changes in perceptual capability. To this end, we conduct experiments across several preparations, including humans (psychophysics and fMRI) and monkeys (high-resolution 9.4T fMRI). A key model we employ is the rodent whisker (vibrissa) sensory system. The 'barrel' cortex, where one 'barrel' corresponds to each vibrissa on the face, provides a systematic representation to test hypotheses relating cortical organization, dynamics and sensory processing. Despite its isomorphic anatomy, barrel cortex shows robust sensory dynamics. Subthreshold receptive fields provide a broad substrate for these interactions, as a single barrel column receives input from ~25 different vibrissae. Further, the divergence of activation from one vibrissa to the cortex is frequency dependent and shows a more discrete spatial spread when vibrissae are stimulated at ~8Hz, the frequency at which animals move their vibrissae during natural exploration. These studies suggest that self-induced transitions in cortical activation patterns may shift the cortex between states optimal for the detection or discrimination of sensory stimuli. Further, we have discovered two novel feature maps overlaid on the somatotopic barrel map, a 'macro' columnar frequency representation and a 'micro' columnar direction map. The frequency map emerges due to vibrissa resonance properties. More posterior vibrissae are longer, and thus have lower fundamental resonance frequencies than more anterior, shorter vibrissae, creating a map of frequency across the face. Because whiskers in a dorsal-ventral arc have the same basic shapes and frequency tuning, common frequencies are preferentially transmitted to the corresponding arc of barrel columns, creating a multi-barrel 'isofrequency' column. This mapping from peripheral transduction to cortical maps parallels the cochlea-dependent frequency maps of the auditory system. The frequency map is complemented by a direction map organized on a sub-barrel column scale(<500 micron). Direction preference is linked to somatotopy, such that within a barrel column, neurons respond optimally to vibrissa motions directed toward the neighboring vibrissa that evokes the strongest surround input. This correlation of somatotopy and direction creates a 'pinwheel'-like outwardly radiating direction map. This map was robust in layer II/III but not in layer IV, indicating that it is an emergent property of the cortical circuit. These overlapping and systematic maps in rodent vibrissa cortex echo those seen in primary visual cortex of cats and monkeys. Given the absence of such detailed representations in the rodent visual system, our findings suggest that high-resolution maps are a common feature of high-acuity sensory systems, where selective pressures dictate the emergence of a more metabolically and computationally efficient cortical architecture.

Steven Hsiao, PhD

04/03/2006 4:00pm

"Light reception for non-image detection:role of rods, cones and the new photoreceptors (melanopsin-containing retinal ganglion cells"


Samer Hattar, PhD

04/10/2006 4:00pm
Paul W. Glimcher, PhD
Cntr for Neural Science, NYU

"The Psychophysics of Decision: Studying Inter-temporal Choice" Location: 234 Ames Hall

Traditional psychophysical theory rests on the measurement of ‘just noticeable differences’. These experiments quantify the accuracy of decisions like: ‘which of these two objects is heavier’. Weber noted that the accuracy of these decisions declines as a logarithmic function of stimulus intensity, in this example as a function of weight. A phenomenon we now refer to as the Weber Law. Fechner’s contribution to this process was his axiomatic assertion that the perceived intensity of a stimulus, not just perceptual accuracy, also varies as a logarithmic function of intensity. The resulting psychophysical theory thus discriminates choice functions (which is heavier) from perceptual states (how does it feel) using a simple mathematical set of assumptions. Perhaps surprisingly, nearly all neurobiological studies of perception that attempt to link neural and mental states have focused on correlating neural activity with the choice functions rather than with the underlying perceptual state. There have been almost no demonstrations that a specific neural signal can account for a specific mental state. One area in which this dichotomy is exceptional clear is in the study of inter-temporal choice. The subjective, or perceived, value of a reward drops hyperbolically as a function of the delay to that reward. Our goal in studying inter-temporal choice is thus two-fold. First, we hope to understand the neural mechanism that accounts for the hyperbolic decline in value. Second, we hope to use this example to demonstrate that mental states and neural states can be explicitly linked. To that end single-unit recording data from monkeys and functional imaging data from humans will be presented.

Veit Stuphorn, PhD

04/24/2006 4:00pm
Eero P. Simoncelli, PhD
Howard Hughes Medical

"Characterization of neural response with stochastic stimuli"

I'll describe some of our recent results on characterization of neural response using stochastic stimuli. The ingredients of the problem consist of a distribution over input stimuli, a model (or class of models) describing the input-output relationship, and a method of estimating the model parameters from the responses (spike trains) of a neuron. I'll start by describing the classical reverse-correlation solution, and then generalize to two new methods that are capable of characterizing a much broader and potentially more realistic set of models. I'll show example analyses of neurons in retina and primary visual cortex of macaques.

Rudiger von der Heydt, PhD

05/30/2006 11:00am
Dr. Bechir Jarraya

Functional recovery following continuous lentiviral dopamine production in a primate model of Parkinson's disease

Levodopa oral intake improves motor symptoms in Parkinson's disease; however intermittent administration results in involuntary movements known as dyskinesia. We conducted a preclinical study of continuous dopamine production in the striatum of MPTP primates using a tricistronic lentiviral vector encoding proteins for dopamine synthesis. Gene transfer significantly reduced akinesia, bradykinesia, posture abnormalities and tremor, in the absence of dyskinesia. This behavioral benefit was associated with an increase of dopamine levels in the transduced striatum. Metabolic activity and neuronal activity studies are ongoing. Exogenously administrated dopaminergic treatment similarly improved motor deficits but failed to prevent dyskinesia. These results suggest that lentiviral dopamine production might have a potential use in gene therapy for Parkinson's disease.

Ernst Niebur, PhD

05/30/2006 4:00pm
Dr. Stephanie Palfi

"Cortical stimulation in movement"

A concept in Parkinson's disease postulates that motor cortex may pattern abnormal rhythmic activities in the basal ganglia, underlying the genesis of observed motor symptoms. We conducted a preclinical study of electrical interference in the primary motor cortex using a chronic MPTP primate model in which dopamine depletion was progressive and regularly documented using 18F-DOPA positron tomography. High-frequency motor cortex stimulation significantly reduced akinesia and bradykinesia. This behavioral benefit was associated with an increased metabolic activity in the supplementary motor area as assessed with 18-F-deoxyglucose PET, a normalization of mean firing rate in the internal globus pallidus (GPi) and the subthalamic nucleus (STN), and a reduction of synchronized oscillatory neuronal activities in these two structures. Motor cortex stimulation is a simple and safe procedure to modulate subthalamo-pallido-cortical loop and alleviate parkinsonian symptoms without requiring deep brain stereotactic surgery. Publications related to the topic: Drouot X, Oshino S, Jarraya B, Besret L, Kishima H, Remy P, Dauguet J, Lefaucheur JP, Dolle F, Conde F, Bottlaender M, Peschanski M, Keravel Y, Hantraye P, Palfi S. Functional recovery in a primate model of Parkinson's disease following motor cortex stimulation. Neuron. 2004 Dec 2;44(5):769-78. Other publications related to non human primate models of basal ganglia disorders: Palfi S, Leventhal L, Chu Y, Ma SY, Emborg M, Bakay R, Deglon N, Hantraye P, Aebischer P, Kordower JH. Lentivirally delivered glial cell line-derived neurotrophic factor increases the number of striatal dopaminergic neurons in primate models of nigrostriatal degeneration. J Neurosci. 2002 Jun 15;22(12):4942-54. Kordower JH, Emborg ME, Bloch J, Ma SY, Chu Y, Leventhal L, McBride J, Chen EY, Palfi S, Roitberg BZ, Brown WD, Holden JE, Pyzalski R, Taylor MD,Carvey P, Ling Z, Trono D, Hantraye P, Deglon N, Aebischer P. Neurodegeneration prevented by lentiviral vector delivery of GDNF in primate models of Parkinson's disease. Science. 2000 Oct 27;290(5492):767-73. Palfi S, Conde F, Riche D, Brouillet E, Dautry C, Mittoux V, Chibois A, Peschanski M, Hantraye P. Fetal striatal allografts reverse cognitive deficits in a primate model of Huntington disease. Nat Med. 1998 Aug;4(8):963-6. Palfi S, Ferrante RJ, Brouillet E, Beal MF, Dolan R, Guyot MC, Peschanski M, Hantraye P. Chronic 3-nitropropionic acid treatment in baboons replicates the cognitive and motor deficits of Huntington's disease. J Neurosci. 1996 May 1;16(9):3019-25.

Ernst Niebur, PhD

06/12/2006 4:00pm
MBI Research Presentation 1
Computational Neuroscience
The Zanvyl Krieger Mind/Brain Institute

"Computational Neuroscience"


Ernst Niebur, PhD

09/25/2006 4:00pm
Corinna Darian-Smith

Stanford School of Medicine

"Neuronal Plasticity and behavioral recovery following spinal injury in monkeys"

We are interested in how sensorimotor pathways involved in primate hand function reorganize/ compensate following particular injuries that block normal sensory input from the hand. Recent investigations from our lab have addressed questions relating to neuronal and behavioral changes that occur in the adult macaque monkey following a dorsal rhizotomy. This work will be discussed from a behavioral, anatomical and electrophysiological perspective. I will also present recent work from our lab that shows that adult neurogenesis may also play a role in the recovery process.

Stewart Hendry

09/27/2006 4:00pm
Dora E. Angelaki, PhD
Self-motion perception: Multisensory integration in extrastriate visual cortex
Washington University in St. Louis/School of Med.

Self-motion perception: Multisensory integration in extrastriate visual cortex

Optic flow patterns generated during self-motion provide a strong cue for the perception of our own movement through space (heading). However, accurate judgments of heading often require integration of visual and nonvisual cues, including vestibular, kinesthetic, and eye movement signals. This sensory integration is complicated by the fact that signals from different modalities may originate in different coordinate frames (e.g., eye-centered or head-centered). To investigate the neural basis of self-motion perception, we record from neurons in area MSTd of macaque monkeys during optic flow (Visual condition), real motion (Vestibular condition) and congruent combinations of the two (Combined condition) using a novel virtual reality system that can move animals along arbitrary paths through a 3D virtual environment. To examine how visual and vestibular signals in MSTd contribute to heading perception, we train animals to perform a fine self-motion discrimination task. Heading directions are varied in small steps around straight forward, and monkeys report whether their heading is to the right or left of straight ahead. Psychophysical thresholds averaged 1-3º and, although the most sensitive MSTd neurons had thresholds close to behavior, the average neuron was much less sensitive than the monkey under both single-cue conditions. In the Combined condition, psychophysical thresholds were significantly lower than in the single-cue conditions and very close to predictions of optimal cue integration theories. According to whether the visual and vestibular heading preferences were well matched or nearly opposite, MSTd neurons could be divided into two distinct groups: ‘congruent’ and ‘opposite’ cells. We found that neuronal thresholds in the combined condition were strongly dependent on congruency of heading preferences, such that 'congruent' neurons showed substantially lower thresholds under cue combination, whereas ‘opposite’ cells showed elevated thresholds. The effect of cue combination on ‘congruent’ cells was very similar to that seen behaviorally, suggesting that this subset of MSTd neurons may contribute to sensory integration for heading perception. To assess functional coupling between MSTd responses and perceptual decisions, we compute 'choice probabilities' (CPs), which characterize the trial-to-trial covariation between neural responses and the animal's choices. CPs were strongest under the Vestibular condition, in which ~30% of MSTd neurons showed significant positive effects and the mean CP (0.56) was significantly greater than chance (0.5). In contrast, CPs in the Visual condition depended strongly on the congruency of visual and vestibular heading preferences. ‘Congruent' neurons were positively correlated with choices in the Visual condition (mean CP=0.59), whereas ‘opposite’ neurons tended to be negatively correlated (mean CP=0.45) with heading percepts. These results provide the first evidence that links cortical vestibular signals with heading perception. In addition, our Visual CP data suggest strongly that the visual responses of MSTd neurons are read out differently depending on the congruency of visual and vestibular tuning preferences. This result supports the notion that ‘congruent’ cells may have a privileged role in cue combination, and highlights the importance of studying multi-sensory integration for understanding spatial perception.

Ed Connor, PhD

10/02/2006 4:00pm
MBI Lab presentation

Johns Hopkins University/Krieger Mind/Brain Instit

"Neural Mechanisms of Visual Perception"


Rudiger vonderHeydt

10/09/2006 4:00pm
Minoru Kimura, PhD
Professor of Physiology
Kyoto Prefectural University of Medicine

"Neural basis of reward-based action selection in the basal ganglia and the centromedian nucleus of thalamus"

Estimation of the reward an action will yield is critical in decision-making. To elucidate the role of the basal ganglia in this process, we recorded striatal neurons of monkeys who chose between left- and right-handle-turns based on the estimated reward probabilities of the actions. During a delay period before the choices, activity of more than one-third of striatal projection neurons was selective to the values of one of the two actions. Fewer neurons were tuned to relative values or action choice. These results suggest representation of action values in the striatum, which can guide action selection through biasing decision for an action based on its reward value in the basal ganglia circuit. However, events do not always appear as expected, and sometimes we must switch from the biased higher-value action to lower-value, undesirable action. We found that over half of monkey centromedian (CM) thalamus neurons were selectively activated when a small-reward-action was required when a large-reward-option was anticipated. Thus, CM thalamus appears to participate in a mechanism complementary to decision- and action-bias.

Veit Stuphorn

10/23/2006 4:00pm
Gonzalo Marin Game
Facultad de Ciencias Biologia
Universidad de Chile

"A Spatial Attentional Mechanism in the Midbrain of Birds"

Dr. Marin will describe in pigeons the operation of a neural circuit in the midbrain that might help animals to select and attend to one visual stimulus from the myriad stimuli that are displayed in their visual environment. This mechanism is based on a modulating signal bearing upon the optic tectum, which is the main visual centre in the midbrain of most vertebrates. His laboratory has shown that whenever a visual stimulus appearing in the visual field activates neurons in a given location of the tectum, this location receives a strong cholinergic feedback from neurons of the nucleus isthmi pars parvocellularis, situated underneath the tectum. New results demonstrate that if a second stimulus appears in the visual field, the modulating feed-back in the first tectal location is diminished or interrupted, while it re-starts at the second tectal location. Our data also suggests that the upstream transmission of visual activity to higher visual areas is controlled by this modulation. In this sense, the modulating feedback that he describes seems to act as an internal spotlight that illuminates selected spots of an internal representation of visual space, which may enhance perception and drive orienting responses.

Alfredo Kirkwood

10/30/2006 4:00pm
Lee E. Miller, PhD
Associate Professor of Physiology
Northwestern University

Muscle-like commands in the motor cortex: The basis for natural brain machine interface?

In the early 1980’s, Georgopoulos discovered the dependence of motor cortical discharge on the direction of hand movement, and introduced the now-classic concept of the “preferred direction” organization of primary motor cortex (M1). The details of this model have been addressed by many subsequent experiments, and it was also the basis of one of the most successful brain machine interfaces for the control of movement. Recently, our group has introduced the alternate concept of a preferred direction in “muscle-space” (PDM). This model has grown out of research showing the strong correlation between M1 discharge and muscle activity. PDM is analogous in many ways to the classic kinematic preferred direction of hand movement. PDM expresses the similarity between the neuronal discharge, and each of N muscles. Its direction is quite varied across neurons, yet constant over time. Across different reaching tasks, this muscle-space preferred direction remains more stable than its hand-space analog. In a sense, PDM expresses the functional divergence from a single neuron to groups of controlled muscles. We have also studied the corresponding convergence from large numbers of neurons to single muscles. With simple linear filter methods, we have been able to use the activity of 10-40 simultaneously recorded neurons to make predictions of the activity of individual arm and hand muscles that account for 60-80% of the variance of the actual EMG signals. This compares quite favorably with the quality of the predictions of hand movement that have been made by other groups, despite the greater bandwidth of the EMG signal. These latest results have been the basis for a new brain machine interface project in my lab that uses predictions of EMG activity as control inputs to a set of stimulating electrodes implanted on several forearm muscles. The goal of that project is to restore grasp function in a monkey with a temporarily paralyzed hand. In addition to the potential clinical application of this BMI project, we hope they will reveal new information about the nature of the signals encoded in M1.

Veit Stuphorn

11/06/2006 4:00pm
Stefan Everling, PhD
Director, The Centre for Brain and Mind
University of Western Ontario

"Monkey fMRI and single unit recordings in PFC and ACC in an anti-saccade task"

A characteristic feature of primate behaviour is their ability to act flexibly in response to environmental events. This ability is a hallmark of executive control. The anti-saccade has emerged as an important task for investigating this executive control of action. In this eye movement task, subjects must suppress the automatic urge to look at a stimulus that is flashed in the peripheral visual field and must instead look away from the stimulus in the opposite direction. This lecture will review recent neurophysiological and neuroimaging studies demonstrating a crucial role of the prefrontal cortex and anterior cingulate cortex in the top-down inhibition of automatic saccades.

Veit Stuphorn

11/13/2006 4:00pm
Amy Bastian, PhD
Assistant Professor, Kennedy Krieger Institute
Kennedy Krieger Institute, Johns Hopkins Medical I

“Does our current understanding of motor learning actually help patients?”

Many interesting principles about normal human motor learning have been elucidated over the last decade. We now know more about: the time course, generalizability, interference effects, and augmentation of these processes. Lesion studies have shed light on the brain areas required for different aspects of motor learning. What remains unknown is whether patients can recover function using comparable motor learning or relearning strategies, and how/whether the principles defined above apply to optimize this process. Here, I will discuss whether information from learning studies in healthy humans provide insights that actually help patients to improve movement control.

Mind/Brain Institute

11/27/2006 4:00pm
Kevin D. Alloway, PhD
Professor of Neural and Behavioral Sciences, College of Medicine
Pennsylvania State University

Neuronal Convergence and the Neural Basis of Sensorimotor Integration

Spatial and temporal integration represent fundamental principles of neuronal processing throughout the brain, but few experimental studies have actually characterized integration at the systems level. We use the rodent SI barrel cortex as a model for characterizing the structural-functional relationships that govern the transmission of sensory information to target brain regions that integrate somatosensory inputs. As part of our comparative approach in characterizing sensorimotor integration in MI cortex, the neostriatum, and the pons, we use a dual anatomical tracing paradigm in conjunction with multiple neuronal recordings. These experiments, which are still in progress, suggest that corticopontine interactions are much stronger than corticostriatal or corticocortical interactions.

Ernst Niebur

12/04/2006 4:00pm
Allison Okamura, PhD
Human-Machine Interaction with Haptics
Johns Hopkins University/Mechical Eng. Dept

Human-Machine Interaction with Haptics

My research is devoted to developing the basic principles and tools needed to realize advanced robotic and human-machine systems capable of haptic interaction. In useful haptic systems, force and touch information cannot be obtained or displayed passively; haptics inherently requires physical contact and is greatly affected by system dynamics. My methodology to address these challenges is to use dynamical modeling and mathematical analysis to predict the behavior of haptic devices, develop new and creative solutions for the acquisition, display, and use of haptic information, and provide examples of practical application through working systems. This talk will describe a number of haptic systems developed in the Johns Hopkins University Haptics Laboratory, including teleoperated robots for surgery, virtual environments for surgical simulation, and novel haptic display devices.

Amy Bastian, PhD

12/11/2006 4:00pm
Ralph Mitchell Siegel, PhD

Rutgers University, Cntr for Molecular and Behavio

"There is no spoon": the representations of association cortex in monkey

Representation of visual and cognitive features in parietal association cortex in monkey are highly labile. What is the nature of these representations? How can projective cortices interpret such signals? Do they even need to? What is the outlines of a model that could generate such representations? Or is it all illusion? "There is no spoon"

Ed Connor, PhD

01/22/2007 4:00pm
March H. Schieber, MD, PhD
Professor, Dept of Neurobiology & Anatomy
University of Rochester

"An iconoclast in the motor cortex: Changing the circuits that control the fingers"

Traditionally, the primary motor cortex (M1) has been viewed as a somatotopic array of “upper motor neurons” which provide monosynaptic labeled lines to the spinal motoneuron pools of individual muscles. Our recordings of M1 neurons have shown, however, (a) that single M1 neurons often are active during the movement of multiple fingers; (b) that patterns of M1 neuron activity are diverse; and (c) that neuron activity in M1 during any finger movement is spatially distributed throughout the hand representation. These observations suggest that M1 neurons function as a diverse and distributed network in controlling the movement of any digit. To examine the operation of this diverse network during individuated finger movements, we used spike-triggered averaging to identify M1 neurons whose spikes were associated at short latency with the arrival of inputs in specific motoneuron pools. The activity patterns of such M1 neurons bore only limited similarity to the EMG activity patterns of their target muscles. Furthermore, summing the effects of all the M1 neurons with effects in the same muscle only partially reconstructed the muscle’s EMG pattern. These observations suggest that muscles (lower motoneurons) do not simply follow driving inputs from M1 (upper motor neurons). Rather, considerable integration of synaptic inputs, which may include not only corticospinal neurons but also rubrospinal and reticulospinal neurons as well as spinal interneurons, may be needed to shape the muscle activity that produces fine finger movements.

Steven Hsiao, PhD

01/29/2007 4:00pm
Peter Steinmetz, PhD
Separate image durations activate distinct neuronal populations in the human medial temporal lobe.
Arizona State University

Separate image durations activate distinct neuronal populations in the human medial temporal lobe.

Previous experiments have demonstrated that neurons in the human medial temporal lobe (MTL) may respond differentially to visual categories when human observers performed a face/non-face discrimination task in images presented for 1000 ms. This presentation duration, however, is much longer than needed to perceive an object and perform a simple discrimination task. Thus, it is unclear whether category-specific neurons, as well as image-specific neurons, would be observed for durations shorter than 1000 ms. To examine the effect of duration on category selective visual responses in the MTL, we varied presentation duration from 800, 500, to 300 ms. 15-20 pictures, selected from 5-6 image categories (animal, building, famous people, indoor scene, outdoor scene, and tools) were shown. Subjects were pharmacologically-resistant epilepsy patients who had undergone intracranial electrode implantation. Primary recording sites were in the medial temporal lobe (including the hippocampus, amygdala, and entorhinal cortex) and occasionally in the frontal lobe, as determined by clinical criteria. The patients performed a face/non-face discrimination for each of the images. We identified 42 well isolated neurons recorded from three patients. Neural responses were first tested for a category selective visual response at each of the 3 durations (response during 0-1000 ms from the onset of stimulus, 1-way ANOVA, p<0.05). Table 1 shows the number of neurons with category selective responses during each of the 3 durations and their combinations. The vast majority of neurons respond to only a single duration. A 2-way ANOVA (p<0.05) of the responses as a function of image duration and image category showed that 20 of 42 neurons had a primary effect of image duration. These findings suggest that separate populations of neurons within the human medial temporal lobe are activated by different durations of presentation and may have separate cognitive functions in memory or learning. Duration (ms) 300 500 800 300 & 500 300 & 800 500 & 800 300 & 500 & 800 5 5 3 0 0 1 1 Table 1. Number of neurons responded at specific durations or at a number of durations.

Ernst Niebur, PhD

02/05/2007 4:00pm
Steve M. Potter, PhD
Georgia Institute of Technology and Emory Universi

"Multielectrode Arrays for Embodied Cultured Networks"

I will describe recent advances using in vitro cultured cortical networks for the study of network plasticity and dynamics. We have developed a new paradigm for in vitro research, in which networks are re-embodied by using their outputs to control robots or animat simulations. We created the first closed-loop systems in which the networks’ activity controls their own multielectrode stimulation and chemical modulation. This technology is amenable to simultaneous multiphoton imaging, to make correlations between functional and morphological plasticity.

Matthew J. Roos, Graduate Student

02/12/2007 4:00pm
Timothy J. Ebner, M.D., PhD
“What cerebellar Purkinje cells encode about arm movements: Implications for theories of cerebellar function”
University of Minnesota

“What cerebellar Purkinje cells encode about arm movements: Implications for theories of cerebellar function”

The cerebellum is critical to the production of smooth, coordinated movements. Numerous hypotheses have been advanced to explain the cerebellum’s contribution to movement. Prominent among recent theories is that the cerebellum is the site of internal models of the motor apparatus. Internal models provide for representations of the input-output properties of the motor apparatus or their inverses. Inverse dynamics models, in response to the desired trajectory, generate the signals needed to control joint torques/forces. Forward models predict the state of system as a consequence of the current state of the arm and the motor commands. An advantage of internal models is the reduced need for sensory feedback to perform accurate and coordinated movements, overcoming the problem of long time delays and low gains in the feedback loops. This seminar will examine the psychophysical, imaging, electrophysiological, and lesion evidence that the cerebellum is the site of internal models and recent single unit recording studies in the monkey from the Ebner laboratory. The discharge properties of Purkinje cells during tracking under various external force fields are incompatible with Purkinje cells being the output of an inverse dynamics model of the arm. The robust kinematic signals carried by Purkinje neurons are potentially consistent with a forward internal model. Strategies being used to test the forward internal model hypothesis will be discussed.

Pramodsingh Thakur, Graduate Student

02/19/2007 4:00pm
Kirkwood Lab presentation

"Mechanisms of Cortical Modification"


Alfredo Kirkwood, PhD

02/26/2007 4:00pm
Dr. Gary Wilson

Moorpark College

"Insights on Primate Training"

Some common mistakes that trainers make when working with primates will be described along with methods for avoiding them. A problem-solving approach to behavior modification will be outlined which contributes to improved effectiveness while reducing distress in both the animal subject and the trainer.

Ernst Niebur, PhD

03/01/2007 1:00pm
Hideaki Shimazaki, PhD

Kyoto Univerisy, Japan

"A recipe for constructing a Peri-stimulus Time Histogram"

The peri-stimulus time histogram (PSTH) is a handy tool for capturing the instantaneous rate of neuronal spike occurrence. In most of the neurophysiological literature, the bin width that critically determines the goodness of the fit of the PSTH to the underlying spike rate has been selected by an individual author in an unsystematic manner. We propose an objective method for selecting the bin width of a PSTH from the spike data, so that the resulting PSTH best approximates the unknown underlying spike rate. The resolution of the PSTH increases, or the optimal bin width decreases, as the number of sampled data increases by repeating an experimental trial. It is notable that the optimal bin width may diverge if only a small number of experimental trials are available. In this case, any attempt to characterize the underlying spike rate by a histogram will lead to a spurious result. Given a paucity of data, we also provide a method that suggests how many more trials are needed until the set of data can be analyzed with the required resolution. Short summary: Shimazaki H. and Shinomoto S., Advances in Neural Information Processing Systems, Vol. 19 (2007)

Ernst Niebur, PhD

03/05/2007 4:00pm
Hugh Wilson, PhD
Perceptual Oscillations and Waves in Vision
York University

"Perceptual Oscillations and Waves in Vision"

When the two eyes are presented with mutually incompatible stimuli, such as orthogonal gratings, the cortex defaults to an oscillation, termed binocular rivalry, between the monocular inputs. Not only is rivalry a rich area for studying nonlinear dynamics in the cortex, but it also serves to elucidate a range of neuronal interactions. This talk will discuss the minimal physiological requirements for binocular rivalry and their relationship to both traveling rivalry waves and hierarchic levels of cortical processing. Additional topics will include the link between rivalry and stereopsis, rivalry hysteresis, and short term perceptual storage in rivalry.

Rudiger von der Heydt, PhD

03/12/2007 4:00pm
Dr. Tai Sing Lee

Carnegie Mellon University

"Statistical regularities in natural scenes and their implications in visual inference"

Classical approaches in computer vision rely on mathematical models of image formation and the strategy of "reverse optics". Many visual phenonmena in natural scenes, however, are too complex to describe with image formation models. Visual inference in a complex environment can benefit from a deeper understanding of the statistical relationships between natural scene structures and the images we see. In this talk, I will report some statistical patterns between 3D scenes and 2D images that we have discovered, and describe how these patterns can be used to improve computational algorithms for 3D inference. Finally, I will present new neurophysiological findings that illuminate how disparity-tuned neurons work together during 3D inference, and how the statistical patterns might be encoded by these neurons to facilitate multiple cue integration in the process. Our work illustrates the potential benefit of the interplay between computational and neurophysiological studies.

Rudiger von der Heydt, PhD

03/19/2007 4:00pm
Kim 'Avrama' Blackwell, V.M.D., Ph.D.
"The role of the striatum in reinforcement learning"
George Mason University

"The role of the striatum in reinforcement learning"

Operant conditioning is a form of associative learning in which rewarding an animal's response increases the likelihood of eliciting the response. The ability to use appropriately timed rewards to shape complex behaviour inspires scientists in psychology, neurophysiology, and modeling. Temporal difference models and experimental results agree that reward elicits dopamine release in the striatum and that striatal spiny projection neurons learn the association between the motor response and reward. Nonetheless, critical aspects of operant conditioning behavior have not been replicated. In particular, if synaptic plasticity underlies learning, then the temporal interval between dopamine and cortical inputs should be critical in producing plasticity of cortico-striatal synapses. I describe studies which test the temporal sensitivity of striatal synaptic plasticity, and investigate which molecules are critical for coincidence detection. We develop a computer model of the signaling pathways activated by dopamine and glutamate in the spiny projection neuron of the striatum to test whether plasticity is sensitive to temporal interval. Results show that kinases essential to plasticity are activated more in response to coincident inputs. Another aspect of reward learning models is that firing of the spiny projection neurons modulates motor actions. Though these neurons constitute 90% of the neuronal population, firing is critically shaped by both lateral and feedforward inhibition, the latter mediated by fast-spiking interneurons. I describe some experiments and modeling to investigate the synaptic inputs to spiny projection neurons and fast spiking interneurons, and characterize how the interneurons transform their inputs into inhibitory signals. Experiments reveal that synaptic inputs occur at a relatively low frequency of 10 - 40 Hz during the down-state, and increase to ~800 Hz during up-states for both neuronal types. The frequency and characteristics of synaptic inputs to both spiny projection neurons and fast spiking interneurons are similar, with approximately half of the inputs glutatmatergic, and the other half GABAergic. Modeling demonstrates that a transient potassium (KA) current allows FS interneurons to strike a balance between sensitivity to input and robustness to noise, thereby increasing its signal to noise ratio. Future studies to construct a network of striatal neurons should illuminate how the striatum performs action selection in reward learning.

Ernst Niebur, PhD

03/26/2007 4:00pm
Florian Engert, PhD
Assistant Professor of Molecular and Cellular Biology
Harvard University

"Visual processing in the developing zebrafish"


Joshua Vogelstein, Graduate Student

04/02/2007 4:00pm
John Bickle, PhD
Four Conditions on Sufficient Experimental Evidence to Establish a Molecular Basis for a Cognitive Phenomenon
University of Cincinnati

Four Conditions on Sufficient Experimental Evidence to Establish a Molecular Basis for a Cognitive Phenomenon

Although virtually unknown to philosophers and cognitive scientists, the neuroscientific field of ‘molecular and cellular cognition’ has been investigating and discovering the cellular and molecular bases of cognition for more than fifteen years. Learning and memory have so far provided the field’s greatest accomplishments, but experimental work is now underway on virtually all phenomena that comprise cognitive science. The key experimental practices that characterize this field are the use of genetically engineered mutant mammals in a variety of behavioral procedures widely accepted as measures of specific cognitive functions. Examples already well established in the experimental literature include manipulations of α-calcium-calmodulin kinase II (αCaMKII), cAMP-dependent protein kinase A (PKA), and cAMP response-element binding protein (CREB) in vivo, and behavioral tests of “declarative” (“explicit,” “hippocampus-dependent”) learning and memory such as the Morris water maze, contextual fear conditioning, social recognition memory, and the social transmission of food preference. Neurobiologist Alcino Silva was the first to propose four conditions on what constitutes sufficient experimental evidence to establish a molecular mechanism for a systems-level cognitive phenomenon. They are: negative alteration (decrease the probability of the hypothesized mechanism and reliably measure decreases in the probability of the cognitive phenomenon); positive alteration (increase the probability of the mechanism and reliably measure increases in the probability of the phenomenon); observation (establish that the mechanism precedes the phenomenon by the appropriate time interval); and integration (develop hypotheses about how to get from the mechanism to the phenomenon and integrate those hypotheses with other experimentally-established independent data about both the mechanism and the phenomenon). In collaboration, Silva and I have now refined those conditions based on a metascientific analysis of some paradigm cases from recent molecular and cellular cognition. In this talk I’ll explain each of these conditions in detail and give examples from the scientific literature (of the sort described in the paragraph above) to illustrate it. I’ll also elaborate the “ruthlessly reductionistic” core of these conditions and demonstrate how the account of reductionism-in-practice drawn from this analysis differs from the accounts of reduction that have been much discussed (and criticized) in recent philosophy of mind and cognitive science. My upshot will be that real reductionism as practiced in this branch of reductionistic neuroscience fares much better than the artificial reductionisms that have been considered in recent philosophy and cognitive science.

Steven Hsiao/Steven Gross

04/09/2007 4:00pm
David Leopold, PhD
What processes in the brain make a stimulus visible?
National Institute of Mental Health

What processes in the brain make a stimulus visible?

The role of the striate cortex (area V1) in shaping our perception of visual patterns and scenes is of fundamental importance in understanding how we see. Visual illusions where a salient stimulus can be temporarily rendered invisible have provided a paradigm by which neuroscientists can isolate neural processes directly involved in perception. This approach has been applied to study electrophysiological activity in monkeys, as well as fMRI activation patterns in humans. Interestingly, the different types of studies have provided divergent results regarding the role of V1 in perceptual processing. Human neuroimaging (fMRI) studies have repeatedly shown that during perceptual suppression there is a marked decrease in the BOLD response in V1. In contrast, single unit studies in monkeys have reported the near absence of such suppression in the firing of individual V1 neurons. Here I will present new data in which we directly compare neurophysiological and fMRI responses during perceptual suppression in monkey area V1. By monitoring the two signals in a variety of conditions within the very same monkeys, we found that the single-unit and BOLD activation specifically ceased to be correlated during periods of perceptual suppression. Slow changes in some bands of the local field potential reflected perception, but only over short time scales. I will discuss how the spatiotemporal pattern of neural events in V1 might contribute to shaping the BOLD response, and will argue that the fMRI signal provides a complementary, rather than incorrect, perspective on neural processing within a cortical area.

Rudiger von der Heydt

04/16/2007 4:00pm
Cynthia F. Moss, Ph.D.
University of Maryland

Spatial orientation by sonar: What the bat's voice tells the bat's brain

Echolocation in bats depends upon the dynamic interplay between auditory information processing and adaptive motor control. The bat produces ultrasonic vocalizations and uses information contained in the returning echoes to determine the direction and distance of objects in space. With this acoustic information, the bat builds a three-dimensional auditory representation of the world. The timing, frequency content, duration and intensity of echolocation signals used to ensonify the environment directly impact the information available to the bat's acoustic imaging system. In turn, the bat's auditory representation of the environment guides its actions--ear movements, head aim, flight path, and the features of subsequent sonar vocalizations. My talk will emphasize the importance of the bat's vocal-motor system to spatial orientation by sonar, and I will present this view in the context of three problems that the echolocating bat must solve: (i) auditory scene analysis; (ii) sensorimotor transformations; (iii) spatial memory and navigation. I will summarize our research findings from behavioral studies of echolocating bats engaged in natural tasks and from neurophysiological studies of the bat superior colliculus and hippocampus, brain structures implicated in sensorimotor integration, orientation and spatial memory.

Ernst Niebur, Ph.D.

04/23/2007 4:00pm
COL Geoffrey Ling, MD., PhD
Arlington, Virginia
DARPA Defense Sciences Office

"DARPA'S Approach to Advancing Science"

The pursuit of scientific knowledge and the application of that knowledge to improving the human condition are both worthy endeavors to commit one’s life to. As one embarks on such a career, it is time to determine how best to begin. The DARPA process is one that can be considered. DARPA, the Defense Advanced Research Projects Agency, is an federal organization whose mission is to ensure that the US maintains its technological supremacy. Colloquially at DARPA, it means “do show that the impossible is possible.” This begins by first developing a challenge problem, i.e., to know what one wants ultimately to accomplish. This is analogous to a hypothesis except that in addition to knowledge, some product will be expected. For many, it will become known as “the deliverable.” The next step is to understand what intermediate goals will clearly demonstrate that the research is proceeding as expected. As research is exploring the unknown, the unexpected outcome should be expected. Further, what the implications are relative to the particular work. Referred to as “go/no go” milestones, it is valuable to both the investigator and the program manager. Finally, an aggressive but realistic timeline is needed. Once this foundation is laid, the proper performance team needs to be assembled. Unlike traditional collaborations which are loose associations of individuals, this is a committed cohesive group led by a single person. The “chain of command with a responsible leader” is paramount for success. Naturally, members of the team need to be expert, dedicated and committed to the effort. What needs to be understood at the outset is that the DARPA process culminates with achievement. Too often, in the pursuit of funding, the process becomes more important that the goal. Active DARPA programs will be used as examples.

Matthew Roos, Graduate Student

04/30/2007 4:00pm
Jorge Golowasch, PhD
New Jersey Institute of Technology

"Activity in gap-junctionally coupled networks depends on dendrite diameter"

Electrical signal transmission between passive and electrically compact neurons coupled via gap junctions is well understood. However, signal transmission via gap junctions between realistically shaped neurons, including dendrites and axons, is complex and depends nonlinearly on the diameter of the coupled processes. Specifically, I will show that there is a unique optimal diameter, depending on membrane and gap junction properties, for which signal transfer between two electrically coupled processes is maximal (or attenuation is minimal). These observations predict that in dendritically gap junction coupled neurons that have an excitable axon, action potential generation may also depend on dendrite diameter. Using a modeling approach, neurons with a passive soma, passive dendrites and an action potential-generating axon were connected by dendro-dendritic gap junctions to form networks of various architectures. We found that action potentials can propagate through the network and can generate sustained rhythmic network activity. The frequencies of the rhythmic activity generated by these networks fall within observed spinal cord (0.3 – 25 Hz) and brain rhythm oscillations (i.e. theta rhythms, 4-12 Hz and beta/gamma rhythm, 15-70 Hz). These gap junction-dependent rhythms may explain observed oscillations in the developing mammalian nervous system at times when chemical synapses are not yet present.

Alfredo Kirkwood, PhD

05/07/2007 4:00pm
John Evenden, PhD
"Determinants of impulsivity: Evidence from animal studies"
AstraZeneca Pharmaceuticals LP

"Determinants of impulsivity: Evidence from animal studies"

Impulsive behaviour is a common feature of many psychiatric and neurological disorders. However until recently it was relatively little studied, and, given its importance in societal impact and treatment compliance, remains under-researched. As behaviour rather than a cognitive or emotional process, it lends itself ideally to translational research. Today’s talk reviews the presenter’s analysis of key factors determining whether an individual displays impulsive behaviour or self-control based upon objective tests carried out in rodents using challenge with psychoactive drugs as a tool to identify key determinants. Impulsive behaviour is multifactorial, and evidence is presented that these different aspects of impulsive behaviour may be mediated by different biological mechanisms, e.g. distinct effects of stimulation of 5-HT1A or 5-HT2 receptors. Impulsivity as a personality trait may be distinguished from an increase in impulsive behaviour in response to a change in the state of an organism. The effects of a drugs on impulsivity can be influenced by the presence of external conditioned stimuli, or by the genetic background of the subjects. Finally, the motivational drive for self-control may influence the response to the drug challenge. Clinical research and on impulsivity development of treatments for impulsivity are hampered by a lack of agreed diagnostic criteria and absence of validated assessment instruments. Adaptation of procedures for testing non-human subjects may stimulate translational research as well as helping define objective endpoints to measure the effects of potential pharmacological and behavioural treatments on impulsive behaviour.

Veit Stuphorn

05/21/2007 4:00pm
Arash Yazdanbakhsh, PhD
Harvard Medical School

"Surface and contour representation in primary visual cortex"

Although the perceptual experience of a uniform surface, a sharp edge, or a simple line may seem straightforward and effortless, the study of their neural correlates has been facing a long term controversy. For an isomorphic representation of uniform surfaces, a filling-in mechanism by which the edge signal diffuses within a closed surface has been adopted by certain models. Such models inspire relevant single cell studies to investigate the possibility of such diffusion mechanism. Here, by using different versions of spike-triggered reverse correlation method, I show how the surface and edge representation can be emerged from the basic spatio-temporal response of the neurons in the primary visual cortex. The advantage of the reverse correlation method is the possibility of having a fine-spaced sampling over the stimulus region accompanied with a precise registration of the temporal dynamics of spikes after each stimulus onset. This ideally can give us a direct evidence of the spatio-temporal dynamics of edge and surface representation to test different models and hypothesis, in particular the ones which suggest filling-in. We implemented the same method to investigate the role of the direction of contrast in contour processing. In this respect, both receptive field dynamics and the interactions within the receptive field were obtained by the same method. A careful selection of stimulus size and the presentation onset and offset timing can cope with different concerns regarding the confounding second-order interactions between the consecutive frames. The results of the above experiments suggest that one can significantly reduce many aspects of the neural response to edges, surfaces, and other perceptually inspired stimuli by studying the spatio-temporal dynamics of the neural response, which can be readily estimated with a high resolution in space and time by implementing spike-triggered reveres correlation method.

Rudiger von der Heydt

07/09/2007 4:00PM
James Knierim, PH.D.
Professor Neurobiology&Anatomy
University of Texas

Memory Formation in the Hippocampal Formation: A Multi-region Analysis

One of the key questions of cognitive neuroscience is how the brain constructs high-order representations of experience and how those representations are stored and recalled as conscious memories. Place cells of the hippocampus are an outstanding model system for deciphering the neural network mechanisms by which the brain constructs these cognitive representations from multimodal input. Converging anatomical, behavioral, and physiological evidence suggests that there are two parallel processing streams that convey different types of information into the hippocampus. One stream conveys spatial information to the hippocampus via the medial entorhinal cortex (MEC) “grid cell” system. The second stream conveys nonspatial information via the lateral entorhinal cortex (LEC). The hippocampus is thought to combine these streams and create conjunctive representations of the spatial and nonspatial information (e.g., object in place, event in context) that are necessary for hippocampus-dependent contextual learning and episodic memory.

Ed Connor

09/17/2007 4:00PM
Nitish Thakor, Ph.D.
Professor Biomedical Engineering
Johns Hopkins University

"Decoding Multi-finger movement for Neuroprosthetic Control"


Ernst Niebur

10/05/2007 1:00PM
Ned Block, Ph.D.
Professor of Philosophy and Psychology
New York University

"Consciousness, Cognitive Accessibility, and the Mesh between Psychology and Neuroscience"

Abstract: How can we disentangle the neural basis of phenomenal consciousness from the neural machinery of the cognitive access that underlies reports of phenomenal consciousness? We see the problem in stark form if we ask how we can tell whether representations inside a Fodorian module are phenomenally conscious. The methodology would seem straightforward: Find the neural natural kinds that are the basis of phenomenal consciousness in clear cases – when subjects are completely confident and we have no reason to doubt their authority – and look to see whether those neural natural kinds exist within Fodorian modules. But a puzzle arises: Do we include the machinery underlying reportability within the neural natural kinds of the clear cases? If the answer is "yes," then there can be no phenomenally conscious representations in Fodorian modules. But how can we know if the answer is "yes"? The suggested methodology requires an answer to the question it was supposed to answer! This paper argues for an abstract solution to the problem and exhibits a source of empirical data that is relevant, data that show that in a certain sense phenomenal consciousness overflows cognitive accessibility. I argue that we can find a neural realizer of this overflow if we assume that the neural basis of phenomenal consciousness does not include the neural basis of cognitive accessibility and that this assumption is justified (other things being equal) by the explanations it allows.

Steve Hsiao

10/12/2007 1:00PM
Eberhard Fetz, Ph.D.
Professor (Department of Physiology and Biophysics)
University of Washington

"Volitional control of cortical activity and recurrent brain-computer interfaces"

A variety of brain-computer interfaces [BCI] have been developed to transform neural activity into signals that control a computer cursor or other external devices. Effective BCI control depends significantly on the ability of the subject to modify neural activity appropriately. Volitional modulation of neural activity is evident in many conventional experimental paradigms, and the degree of neural control has been directly tested in biofeedback experiments and BCI applications. While the usual BCI paradigm involves brain control of external devices, a recurrent BCI [R-BCI] generates output that is fed back into the nervous system or muscles. We are investigating an implantable R-BCI consisting of autonomously operating electronic circuitry, including a computer chip that interacts continuously with the brain of a monkey. The so-called "Neurochip" can document the activity of motor cortex cells and arm muscles during free behavior and sleep, storing this activity for subsequent download via an infrared port. In a recurrent mode, the Neurochip can convert cell activity to electrical stimuli delivered back to the spinal cord or muscles, implementing neurally controlled functional electrical stimulation. The R-BCI has also converted motor cortex cell activity into stimuli delivered at an adjacent cortical site. Continuous operation of such spike-triggered stimulation for a day or two generated long-lasting changes in connections between the synchronized sites. The R-BCI paradigm has numerous potential applications, depending on the input signals, the computed transform and the output targets.

Steven Hsiao

10/22/2007 4:00PM
Philip N. Sabes, Ph.D.
Professor University of California SF
University of California San Francisco

"Sensory integration and adaptation: behavior, optimality, and neural models"

When planning and executing movement, the brain takes into account information from multiple ("redundant") sensory streams. For example, both vision and proprioception are used to localized the hand before a reaching movement. This ability relies on two experimentally observed processes. First, these sensory signals must be integrated, e.g. a "consensus" position estimate is computed. Second, when sensory signals conflict, sensory adaptation occurs to bring the signals back into alignment. I will show that these two processes are linked in a statistically optimal manner. I will present human behavioral experiments illustrating how the brain solves these problems and an "optimal estimator" algorithm that can account for the experimental results. I will then show that this algorithm can be implemented by a fairly simple a network model, which we propose as a model of sensorimotor areas in human parietal cortex.

Amy Bastian

10/29/2007 4:00PM
Andreas Andreou, Ph.D.
Professor Electrical & Computer Engineering
Johns hopkins University



Rudiger von der Heydt

11/12/2007 4:00PM
Takehiro Matsumora
Ph.D. Candidate
Nat'l Institute Physiological Sciences Japan

"Quantitative relationships between the activities of color selective neurons in area TE of the monkey and color discrimination behavior"

On the basis of lesion or electrophysiological studies, it is suggested that area TE (anterior part of IT), which is the highest area in the ventral stream of visual information processing, plays an important role for information processing of color. To study whether there exists close link between color selective responses of TE neurons and perception, we examined the correlation between monkey's color judgment and responses of color selective TE neurons. We trained a fine color discrimination task in two monkeys and recorded single neuron activities from the color-sensitive sub-region in area TE. In this task, a sample color stimulus was presented at the center of the display, and the monkey had to make either the rightward or leftward saccade depending on the similarity of the sample color to the two target colors. Two target colors were determined based on the color selectivity of each neuron such that one target color (preferred target color) evoked stronger response than the other color (non-preferred target color). Sample color in each trial was chosen from seven colors equally spaced between the target colors on the CIE-xy chromaticity diagram. First, we quantitatively compared the color discrimination abilities of neurons and the monkeys. To calculate the discrimination ability of the neuron, we computed 'neurometric function' on the basis of a competition model. This model assumes that the ideal observer choose the preferred-target color if the spike count of the recorded neuron ('neuron') is larger than that of a hypothetical neuron with the opposite tuning ('anti-neuron') that is selective to the non-preferred target color. We applied ROC analysis to compute the probability that an ideal observer choose the preferred target color for each sample color using spike counts of the 'neuron' and 'anti-neuron' to that color. Then we constructed 'neurometric function' that indicates the performance of a given 'neuron'/'anti-neuron' pair to the entire set of sample colors. We compared neural threshold based on the neurometric function and behavioral threshold based on the psychometric function as the color separation which yielded 80% correct color judgments. We found that although some neurons outperformed the monkey, neural threshold was on average slightly higher than the behavioral threshold. Variation of the neural threshold across the color space corresponded well with that of behavioral threshold. We then calculated the 'Choice Probability (CP)' that is a quantitative measure of correlation between trial-to-trial fluctuation of neural responses and the monkey's color judgment. We found that CPs are on average significantly larger than 0.5 and this is consistent with the expectation that the activities of these TE neurons positively correlate with the monkey's color judgment. Finally, we found that there is little correlation between the neural sensitivity for color discrimination and CP. This suggests that a population of color selective TE neurons having various color sensitivity rather than a specific subset of neurons with particularly high color sensitivity contribute to the color discrimination performance.

Ed Connor

11/19/2007 4:00PM
Cornelia Fermuller, Ph.D.
Associate Research Scientist
University of Maryland at College Park

“Computational insights into visual motion processing”

Visual navigation encompasses a large set of capabilities ranging from low level ones such as kinetic stabilization, image motion and 3D motion estimation, through middle level ones such as obstacle avoidance and motion segmentation, to high level ones such as homing (finding your home), reading a map or building visual memories of a location or movement through space. In this talk I will present a computational architecture for implementing navigational competences, most of which are also realized in the "motion pathway" of the primates. In particular, I will discuss the recovery of image motion and 3D motion (or egomotion). An insight gained from the analysis of visual motion led me to understand a principle which explains (predicts) a large number of illusions, and I will discuss some of them. I will finish the talk with a discussion of the motion segmentation problem and my thoughts on its interaction with the overall segmentation process.

Rudiger von der Heydt

11/26/2007 4:00PM
Doris Y. Tsao, Ph.D.
Head of Independent Research Group
University of Bremen

A specialized system for processing faces in the macaque temporal lobe

FMRI has revealed cortical regions in the temporal lobes of both human and macaque brains that show stronger activation to faces than to other visual objects. Single-unit recordings targeted to face-selective regions make it possible to study form processing machinery within these regions in unprecedented detail. In my talk, I will describe results of combined fMRI and electrophysiology experiments in the macaque addressing the following questions: 1) Are each of the face patches specialized for processing faces? 2) How do the different face patches differ functionally? 3) What is the detailed mechanism used by each face patch to encode faces? 4) What is the anatomical connectivity of the face patch system? The answers to these questions make clear that the macaque brain has evolved a highly specialized system for processing faces, made up of tightly connected but functionally distinct nodes.

Katherine Bowman

12/03/2007 4:00PM
David Linden, Ph.D.
Professor of Neuroscience
Johns Hopkins University

"Axonal motility in the intact, adult cerebellum"

We performed two-photon in vivo imaging of cerebellar climbing fibers (CFs; the terminal arbor of olivocerebellar axons) in adult mice. CF ascending branches innervate Purkinje cells while CF transverse branches show a near complete failure to form conventional synapses. Time-lapse imaging over hours or days revealed that ascending branches were very stable. However, transverse branches were highly dynamic, exhibiting rapid elongation and retraction and varicosity turnover. Thus, different branches of the same axon, with different innervation patterns, display branch type-specific motility in the adult cerebellum. Furthermore, dynamic changes in transverse branch length were almost completely suppressed by pharmacological stimulation of olivary firing.

Steve Hisao

12/10/2007 4:00PM
Rick Born, PhD
Professor of Neurobiology
Harvard University

”Integrating Motion and Depth via Parallel Pathways”

Abstract: Visual information processing is both parallel and hierarchical, with each visual area richly interconnected with other visual areas. One example of the parallel architecture of the primate visual system is the existence of two major pathways providing input to the middle temporal visual area (MT): a direct projection from striate cortex (V1), and a set of indirect projections also originating in V1 but then relaying through V2 and V3. Here we reversibly inactivated the indirect pathways while recording from MT neurons and measuring eye movements in alert monkeys, allowing us to assess whether the two different input pathways are redundant or whether they carry different kinds of information. We found that this inactivation caused a disproportionate degradation of binocular disparity tuning in MT neurons relative to direction tuning, suggesting that the indirect pathways play an important role in the recovery of depth in three-dimensional scenes. Single neurons integrate this depth information with information on direction of motion, likely provided by the direct pathway from V1, which might endow them with the ability to segregate motion signals arising from different depth planes.

Sliman Bensmaia

01/07/2008 4:00PM
Tara Thiagarajan, Ph.D.
IRTA Fellow Critical Brain Dynamics
National Institute of Health

“Coherence Potentials: A network level action potential-like phenomenon in the cortex”

Coherence Potentials: a network level action potential-like phenomenon in the cortex The idea that percepts are represented as distributed activity in the cortex dates back over fifty years. However, little is known about how participating neurons are ‘bound’ into an assembly and what aspect of the activity in the assembly is meaningful. In theory, if one could identify a clear network level representation of functional associations, these could potentially reveal more direct relationships with cognitive outcome than those found with the spiking activity of individual neurons. Functional associations between sites in the cortex have been identified in the local field signal in the form of phase relationships or increased coherence between sites during particular behaviors, suggesting that relationships in the temporal structure may be important. However, thus far, analysis has generally been restricted to particular frequency bands and discards amplitude information on the presumption that the spread of activity from a local field is linearly dependent on the individual components of the local field activity. This lies in contrast to the spread of charge within an individual neuron where large amplitude currents in local fields of membrane represent a special regime (i.e. the action potential or spike) that allows for a non-linear enhancement of the ability to transfer charge in a loss less manner. Using multielectrode array recordings of ongoing local field activity from awake macaque monkey we demonstrate brief periods of near-perfect coherence across all frequency bands at other sites (i.e. an identical temporal structure in the waveform) that is phase shifted by up to 50 ms. This increase in the spatial spread of coherence occurs sigmoidally as a function of amplitude, is sequential in nature and depends crucially on fast excitatory and inhibitory transmission. Thus, like the action potential, these ‘Coherence Potentials’ represent a special regime of loss-less propagation of temporal information, occurring only when a sufficiently high degree of synchronization of neuronal activity has been achieved in a local field. ‘Coherence Potentials’ can last from a few tens of milliseconds to several hundred milliseconds and are diverse in their temporal structure and spatial configurations, allowing concurrent associations to be distinguished from one another. The metastable associations formed by these coherence potentials thus suggest key elements appropriate for the representation of percepts.

Steve Hsiao

01/14/2008 4:00PM
Asaf Keller, Ph.D.
Professor of anatomy and Neurobiology
University of MD School of Medicine

“Gateways to tactile perception: Parallel processing of pain and somatosensation”

Vibrissal information is relayed to the barrel cortex through at least 2 parallel pathways: a lemniscal pathway involving the ventroposterior medial thalamic nucleus (VPM), and aparalemniscal pathway involving the posteromedial nucleus (POm). I will review the roleof the lemniscal system, focusing on the mechanisms by which VPM shapes theresponse properties of neurons in cortical barrels. I will argue that although analyses ofthese properties (e.g. receptive field structure and angular preference) have illuminatedthe process of input transformation in sensory pathways, they may have only limitedethological role. I will show that this lemniscal pathway is critical for temporal coding ofsomatosensory inputs. In the paralemniscal pathway, and in POm in particular, neurons respond poorly and unreliably to physiologically relevant stimuli. We tested, and disproved, 2 prevailing hypotheses on POm’s role: as a temporal-to-rate code converter,and as an encoder of vibrissae movements. I will show that the GABAergic nucleuszona incerta (ZI) regulates POm activity is a state-dependent manner. This regulation is mediated by the cholinergic activating system, which enhances POm activity duringstates of arousal and vigilance. However, even in these states, POm neurons fail to reliably encode sensory inputs. I will show that POm is critically involved in codingnoxious stimuli. Specifically, I will argue that the phenomenon of central pain may be theresult of suppressed inhibitory regulation of POm activity.

Alfredo Kirkwood

01/28/2008 4:00PM
Patrick Kanold, Ph.D.
Assistant Professor Biology
University of MD College Park

"Early circuits that regulate cortical development and plasticity"

"Early circuits that regulate cortical development and plasticity" Neuronal circuits are shaped by sensory experience during critical periods of development. After the critical period only limited remodeling is possible. Thus, when peripheral sensory processing is impaired, cortical circuits can become irreversibly miswired. By understanding the circuits and processes that control the critical period we may be able to develop treatments and/or devices that can induce plasticity mechanisms in older individuals, even after the critical period. The young brain is structurally different from the adult brain and contains additional circuits that are formed by subplate neurons. These neurons are among the earliest born cortical neurons, reside in the white matter and disappear during development. After the critical period ends – when subplate neurons are no longer present - only limited plasticity is present. Thus we hypothesize that these neurons participate in types of synaptic plasticity that occur only during the critical period. We investigated this hypothesis with a combination of experiments and computational models, which together demonstrate that subplate neurons are required for the functional maturation of the cortical columnar organization, the development of intracortical inhibitory circuits and the outcome of plasticity during the critical period. These results suggest that subplate neurons act like a "teacher" helping thalamic neurons to make strong and precise connections to their cortical target neurons. By controlling the balance of excitation and inhibition, subplate neurons influence the correlations between thalamic and cortical activity and thus the amount of cortical activity driven by sensory inputs. Therefore, this work provides a framework demonstrating that plasticity during the critical period is the product of a complex and dynamically changing circuit in which subplate neurons play a key role. Characterizing subplate neurons function and connectivity is therefore critical to understanding cortical development and is the current focus of our research. This knowledge offers, for example in the case of developmental disorders, the possibility of providing the functional effect of subplate neurons to induce cortical plasticity after the critical period.

Steve Hsiao

02/18/2008 4:00PM
Dietmar Plenz, Ph.D.
Chief of Neural Network Physiology
National Institute of Mental Health

"Neuronal Avalanches and the Organization of Ongoing Cortical Synchrony in the Awake Macaque Monkey"

Ongoing cortical activity in the absence of any obvious sensory stimulus or motor output carries coherent structure that reflects and influences stimulus responses. This dynamic incorporation of inputs raises the question of whether there are aspects in this coherent structure that are invariant across cortical areas and individuals. Using microelectrode array recordings from premotor and motor cortex of awake macaque monkeys, we analyze the spatiotemporal occurrence of ongoing synchronized activity by tracking negative deflections in the LFP (nLFPs) that are found to correlate with neuronal spiking. We demonstrate that nLFPs form spatiotemporal clusters with an identical organization in both monkeys and cortical areas. This organization, termed neuronal avalanches, is invariant to spatial and temporal scale, as previously reported in vitro, as well as to the threshold of nLFP detection. This new dimension, i.e. threshold invariance, describes a fractal organization of nLFP amplitudes such that small nLFPs are embedded into clusters of larger nLFPs without destroying the spatial and temporal scale invariance of the dynamics. These findings suggest an organization of cortical synchronization that is scale-invariant in its three fundamental dimensions - time, space, and amplitude - and imply a functional blueprint that may define the constraints on the internal representation of the external world.

Ernst Niebur

03/03/2008 4:00PM
Tomoyuki Furuyashiki , MD, PhD
Psychological and Brain Sciences
Johns Hopkins University

Action monitoring and outcome expectancy in rodent orbitofrontal cortex

The prospective information of rewarding outcomes is crucial to plan appropriate goal-directed behavior. Once the behavior is initiated, the behavioral action being executed is monitored to update outcome expectancy so that the original plan of actions can be modified or even terminated. Orbitofrontal cortex (OFC), a subregion of the prefrontal cortex, encodes rewards and the cues that predict them, providing a neural substrate for outcome expectancy in goal-directed behavior. Evidence from recent single-unit recordings mainly from rodents have revealed that OFC additionally encodes the behavioral response in obtaining the reward. In this seminar, I will present our recent findings about distinct properties of response and outcome encoding in rodent OFC in relation to behavioral events in goal-directed behavior. Response selective encoding appears concurrently with outcome expectant encoding in separate, but overlapping, neuronal populations. A novel view of an OFC function based on action monitoring will also be discussed.

Takashi Yoshioka

03/10/2008 4:00PM
Katalin M. Gothard, Ph.D.
Assistant Professor Neurobiology
University of Arizona

"Emotion and Attention in the Primate Amygdala"

The primate amygdala plays an important role in evaluating the emotional significance of all stimuli encountered by the organism. In primates, that live in complex and hierarchical societies, this process of evaluation is geared toward the evaluation of social stimuli, such as facial expressions. In this talk, I will show that single neurons the amygdala respond selectively to monkey faces. The majority of these neurons encode unique combinations of identity and expression suggesting that in the amygdala identity and facial expressions are merged into a single representation. The amygdala is also involved in orchestrating overt behavioral and autonomic responses to images with emotional value. I will show that neural activity in the nuclei that are connected to autonomic effectors is correlated with the skin conductance response and other measures of autonomic arousal. I will report on the division of labor among the basolateral and centro-median group of amygdaloid nuclei. The basolateral group, which has cortical connections and cortical-type cytoachitecture transmits information about facial expressions, whereas the centro-medial group, which is connected to autonomic effectors, responds to direction of gaze. The relationship between identifying facial expressions and detecting direction of gaze will be illustrated by scanpaths recorded while monkeys examine images of conspecifics faces. The process of hierarchical scanning of facial features appears to be guided, at least in part, by the amygdala. These data will be interpreted in the context of the connections between emotion and attention and the ethology of the species.

Veit Stuphorn

03/24/2008 4:00PM
Giulio Tononi, Ph.D.
Professor of Psychiatry
University of Wisconsin

“Sleep Function and Synaptic Homeostasis”

I will review a novel hypothesis about the functions of slow wave sleep—the synaptic homeostasis hypothesis. According to the hypothesis, plastic processes occurring during wakefulness result in a net increase in synaptic strength in many brain circuits. The role of sleep is to downscale synaptic strength to a baseline level that is energetically sustainable, makes efficient use of gray matter space, and is beneficial for learning and memory. Thus, sleep is the price we have to pay for plasticity, and its goal is the homeostatic regulation of the total synaptic weight impinging on neurons. The hypothesis accounts for a large number of experimental facts, makes several specific predictions, and has implications for both sleep and mood disorders.

Alfredo Kirkwood

03/31/2008 4:00PM
Hey-Kyoung Lee, Ph.D.
Assistant Porfessor
University of Maryland

“Visual experience-induced regulation of excitatory synapses in primary sensory cortices”

It is well documented that visual experience can alter the functional connectivity of the visual cortex. In addition, a loss of vision can produce cross-modal changes in the remaining sensory modalities. The latter is thought to allow sensory compensation in blind individuals. My group has recently uncovered that depriving vision in rodents not only alters excitatory synaptic transmission in the primary visual cortex, but produce opposite changes in primary somatosensory and auditory cortices. I will present data demonstrating that synaptic changes in visual cortex, as well as cross-modal changes in other sensory cortices, follow the rules of a homeostatic synaptic plasticity mechanism. In addition, I will discuss potential cellular mechanisms underlying these phenomena.

Steven Hsiao

04/28/2008 4:00PM
Harel Shouval, Ph.D.
Assistant Professor
University of Texas Medical School at Houston

"Learning Reward Timing in cortex: A theoretical study"

The ability to represent time is an essential component of cognition but its neural basis remains unknown. It is commonly believed that the underlying cellular mechanisms reside in high order cortical regions but recent studies show sustained neural activity in primary sensory cortices that are reward dependent and can represent the learned time of expected reward. Although extensively studied in both behavioral and electrophysiological studies, a general theoretical framework capable of describing the elementary neural mechanisms used by the brain to adaptively learn temporal representations is lacking. Here we present two different models by which the brain can learn temporal representations through a simple theoretical framework predicated on reward dependent expression of synaptic plasticity. One model asserts that temporal representations are stored within the lateral neuronal connection matrix, while the other model assumes that temporal representation is accomplished by single neuron conductances. We show that in the network model the representation can be accomplished by setting the connectivity matrix to have an appropriate form. For the single neuron case we postulate a simple experimentally motivated non-linear dynamical model and analyze its dynamics to show how they depend on the system parameters. We demonstrate that in both cases a simple model of reward dependent expression of neuronal plasticity is sufficient to learn the appropriate temporal representations. We implement our models computationally to explain reward-time learning in the primary visual cortex (V1) and suggest experimentally verifiable predictions. We also show how a population of such neurons can estimate the reward time given the spiking activity of the network, and how the error of this estimate scales with the reward time.

Marshall Shuler

05/12/2008 4:00PM
Charles E. Schroeder, Ph.D.
Professor of Psychiatry
Nathan S. Kline Insitute for Psychiatric Research

"Neuronal Oscillations as Instruments of Sensory Selection"

It has been known for over 75 years that, independent of frequency, neuroelectric oscillations reflect rhythmic shifting of neuronal ensembles between high and low excitability states. Recent findings indicate that neuronal oscillations have a highly structured hierarchical organization across frequency bands. These findings also strongly reinforce the conclusion that neuronal oscillations both enable and constrain the brain’s processing of sensory inputs, as well as its generation of motor outputs. I will discuss recent findings concerning the way the brain uses oscillations to process input. I will also discuss a conceptual framework that can help to integrate the evidence on the neuronal oscillation as a mechanism of brain operation, with prior findings in traditional vision research paradigms (e.g., vigilance). The discussion will range across several sensory modalities, and will end with a focus on selective attention’s manipulation of oscillatory phase as a mechanism of sensory selection in Primate V1.

Jeff Yau

09/08/2008 4:00pm
Mario Livio, PhD
Professor Physics & Astronomy
Johns Hopkins University


THE GOLDEN RATIO What do sunflowers, Salvador Dali's painting "Sacrament of the Last Supper", and certain types of crystals have in common? These very disparate elements share a certain number commonly known as the GOLDEN RATIO, expressed by the Greek letter PHI. In a journey through mathematics and astrophysics, botany and zoology, art and architecture, taking in fractals and psychology on the way, this extraordinary number, that has captured the imagination for millennia, will be explored. This will also be a story of obsession-of phi-fixated individuals who have devoted their lives to discovering the properties of this number. This tale begins with the ancient Egyptian and Greek mathematicians, continues with the scientists and artists of the Renaissance, and takes us right to such scientists and masters of the modern world as Penrose, Debussy, and Le Corbusier. Even more importantly, the intriguing question of the unreasonable effectiveness of mathematics in explaining the world will be explored.

Steven Hsiao, PhD

09/15/2008 4:00pm
James Stinear, PhD DC
Director Neuralplasticity Laboratory
Rehabilitation Institute of Chicago

"Manipulating intra-and inter-hemispheric motor cortical excitability: a candidate priming mechanism for walking training post-stroke"

An aim of the research conducted in our Laboratory is to develop stimulation and movement-based priming mechanisms as adjuvants to post-stroke walking therapies. Priming the human motor cortex using either transcranial brain stimulation or repetitive bimanual movement enhances the effects of upper limb therapy. Up-regulating the ipsilesional motor cortex and/or down-regulating the contralesional motor cortex induces similar improvements in paretic hand function. Over the last four years we have investigated priming mechanisms for the lower limb motor system. Our research program seeks answers to the following questions: 1. Can non-invasive transcranial stimulation techniques modulate lower limb motor cortex excitability assessed during walking? 2. Given the close proximity of lower limb primary motor cortices, and the low spatial resolution of transcranial brain stimulation techniques, is it possible to selectively modulate one lower limb cortex? 3. If so, does hemisphere-specific lower limb cortical modulation induce opposite sign cortical modulation in the other hemisphere, as demonstrated for the upper limb cortex? 4. Does priming make the lower limb motor system more susceptible to walking rehabilitation therapies? 5. Following stroke, does the contralesional cortex participate in paretic leg control during walking? 6. If so, is contralesional cortical activity beneficial or maladaptive? 7. How do the answers to the above help us choose priming agents that are appropriate for walking rehabilitation? 8. Finally, can the lessons learned from our adult stroke research be applied to perinatal stroke? The presentation will describe how we answered Questions 1 through 3. Study outcomes and the technical challenges we met along the way will be described. The motivation for asking Questions 4 through 8, and the challenges we foresee in answering them will also be discussed.

Amy Bastian, PhD

09/22/2008 4:00pm
Jason Maclean, PhD
Assistant Professor of Neurobiology
University of Chicago

UP States Render Neocortical Circuits Less Sensitive to Thalamic Inputs

Patterned neuronal activations, produced by network connectivity, have been postulated for decades to be the potential neuronal scheme for informational representation. Critical to this model, multineuronal patterned activations in cerebral cortex arise during UP states—prolonged neuronal depolarizations— occurring simultaneously across many cells. UP states are found in the cerebral cortex both in vitro and in vivo and are particularly prominent during slow wave sleep (SWS). Recently it has been shown that UP states repeatably involve a particular population of neurons activating in a particular sequence. These network events are statistically indistinguishable in the numbers, identities and sequences of cells activated from UP state to UP state activation suggesting that these sequences are representative of highly favored network states and are likely enabled by structured cortical network connectivity. While cortical UP states can be robustly triggered by thalamic input, our data suggest that spontaneous cortical UP states actually originate in the cortex and not the thalamus. While previous studies have probed the responsivity of single neurons during UP states the responsivity of an active network of hundreds of neurons at once has not been evaluated. This is a critical consideration when attempting to evaluate the potential function of these patterns of activation—are the patterns dynamic facilitating online dynamical data processing?—or are the patterned activations carrying out a fixed function as would be implied by the hypothesis that they are integral to the process of memory consolidation? Experiments at the level of the network are essential for answering this question, because studies in which single or even a few cells are monitored at once fundamentally miss the emergent network-level properties of groups of cells. Here we examine how stable network states, characterized by stereotyped multineuronal spatiotemporal dynamics behave in response to new inputs arriving as they are ongoing. We find that although thalamic input during the DOWN state is capable of triggering multicellular patterns of activity in layer 4 the same input during UP states does not perturb these patterns. Correspondingly, the large majority of neurons in layer 4 are insensitive to thalamic input if they are in an UP state. The insensitivity occurs regardless of the timing of the input and the electrophysiological and morphological class of neuron. Finally, we analyze the mechanisms responsible for this cortical insensitivity, finding that both thalamocortical and corticocortical EPSPs are greatly attenuated in excitatory neurons during UP states and that this attenuation is likely the result of a major decrease in input resistance that always accompanies the UP state. Our data indicate that DOWN states appear to make the cortex especially sensitive to thalamic inputs while UP states render the neocortex insensitive to impinging thalamic input—perhaps playing a role in repetitively burning in sequences of multineuronal firing corresponding to activity occurring during waking behavior.

Sliman Bensmaia, PhD

09/29/2008 4:00pm
Stefan Treue, PhD
Director German Primate Center
Center for Systems Neuroscience

"Attentional modulation of visual motion processing: of space, features and objects."

Attentional modulation is a powerful influence on the processing of visual information in primate cortex. I will present experimental findings focusing on the influence of attention on area MT in macaque visual cortex. Here electrophysiological recordings have demonstrated the influence not only of spatial attention but also the neural correlates of attention to stimulus features and of object-based attention. The attentional modulation appears to have a multiplicative influence on neural responses but it is still able to create non- multiplicative changes in receptive field profiles and population responses.

Ernst Niebur, PhD

10/03/2008 4:00pm
Rodney Douglas, PhD
Institute Neuroinformatics Zurich

"The challenge of the neocortex for information technology"

Brains solve in real-time a variety of difficult computational problems that cannot yet be solved by computers using very much faster hardware. Their superior performance is likely due to a fundamentally different style of computational that nature has evolved. The differences lie in the basic processing components, their system architecture, methods of information encoding and transformation, and processes of self-construction, -configuration and -calibration. More than any other part of the vertebrate brain, it is the neocortex that is crucial in the ability of animals to behave intelligently, and so the analysis of its organization and operation is likely to reveal the principles of computation that nature has gained over eons of evolution. In this lecture I will briefly describe four lines of research at INI that pursue these principles: quantitative studies of the adult neuronal circuits; simulation of their developmental self-construction; exploration of an interesting primitive processing operation that is likely embedded in these circuits; and the implementation of this operation in neuromorphic electronic circuits.

Ernst Niebur, PhD

10/06/2008 4:00pm
Jerome R. Busemeyer, PhD
Prof. Psychological& Brain Sciences
Indiana University, Bloomington

“Integrating sophisticated choice models with basic processes to more fully account for complex choice behavior”

Two very different paradigms have been used to experimentally study decision making: in the descriptive paradigm, all the information concerning available options and their possible outcomes is described to the decision maker; in the experiential paradigm, everything about the outcomes produced by each option is learned from trial by trial experience. Different phenomena have been discovered within each paradigm, and consequently, different theories have evolved to account for these disparate findings. The purpose of this presentation is to accomplish four goals: (1) demonstrate the shortcomings of the most commonly used class of choice models, (2) combine basic learning with more sophisticated decision making models so that the combination is capable of addressing the phenomena of both paradigms, (3) evaluate the ability of this integrated model to successfully explain both sets of empirical findings, and (4) present neuroimaging (fMRI) findings that explore possible differences in neural substrates recruited to perform these different types of decision tasks

Veit Stuphorn, PhD

10/20/2008 4:00pm
Joshua W. Brown, PhD
Asst. Prof Psychological&Brain Sciences
Indiana University

“Risk Prediction and Cognitive Control of Decision-Making”

Cognitive control involves a network of brain regions that monitor performance and increase goal-directed control of behavior. The anterior cingulate cortex (ACC) and dorsolateral prefrontal cortex (DLPFC) have been implicated as central players in this network. The error likelihood hypothesis (Brown and Braver, 2005) accounts for performance monitoring effects and suggests that anterior cingulate cortex (ACC) and surrounding areas will become active in proportion to the perceived likelihood of an error. The error likelihood hypothesis was originally derived from a computational model prediction. The same computational model further predicted that ACC will be sensitive not only to error likelihood, but also to the magnitude of the consequences should an error occur. The product of error likelihood and error consequence magnitude collectively defines the “expected risk” of a given behavior in a manner analogous to subjective expected utility theory. Subsequent fMRI results from an incentive change signal task have supported this prediction and shown weaker effects in risk-seeking individuals (Brown & Braver, 2007; 2008), suggesting that ACC in part drives risk avoidance. Further studies suggest that risk prediction effects are reduced in substance abusers and individuals with schizophrenia, and they can be enhanced by effective persuasive messages against risky behavior. These results motivate a new computational neural model of learned response-outcome prediction in ACC, in which the need for control can be anticipated in such a way as to proactively bias decision-making against disadvantageous options.

Veit Stuphorn, PhD

10/27/2008 4:00pm
Jack Gallant
University of California Berkeley

"See what you think: Bayesian reconstruction of perceptual experiences from human brain activity"

A recent paper from our laboratory demonstrated that fMRI measurements of hemodynamic brain activity contain far more information than was believed previously. In fact those data suggested that fMRI might provide enough information to permit true noninvasive reconstruction of perceptual experiences: a direct window into visual perception. This talk will focus on a new Bayesian decoder that can reconstruct natural images seen by an observer from measured brain activity. The decoder combines three elements: a structural encoding model that characterizes signals from early visual areas; a semantic encoding model that characterizes signals from higher visual areas; and appropriate priors that incorporate statistical information about the structure and semantics of natural scenes. By combining all these elements the decoder produces reconstructions that accurately reflect the distribution, structure and semantic category of the objects contained in the original image. These results help clarify why the brain contains many distinct representations of the visual world, and highlight the important role of prior knowledge in visual perception. This decoding framework might form the basis of practical new brain-reading technologies.

Ed Connor, Ph.D.

11/03/2008 4:00pm
John Reynolds, PhD
Associate Professor Neurobiology
Salk Institute for Bilological Studies

“Mapping the microcircuitry of attention: attentional modulation varies across cell classes in visual area V4”

Cortical neurons differ from one another in important ways, including their neurochemical properties, patterns of connectivity, laminar distribution, gene expression patterns and developmental origin. Previous studies of attention have not sought to distinguish among different classes of neurons. We therefore know almost nothing about the complex circuitry that transforms attentional feedback signals into improved visual processing. Studies in the slice and in anesthetized animals find that parvalbumin expressing GABA-ergic interneurons with the morphologies of basket and chandelier cells have short duration action potentials, whereas most excitatory cell classes have longer duration action potentials, a difference that is due to expression of different classes of sodium and potassium channels. We thus examined differences in attentional modulation across visual area V4 neurons classified on the basis of action potential width. The distribution of action potential widths in area V4 is clearly bimodal. We find substantial differences in the basic response properties of these two classes of neurons, including their baseline firing rates, the strength of their stimulus-evoked responses, as well as qualitative differences in the types of variability of the neuronal response across classes. We also find qualitative differences in how the two neuronal classes are modulated by attention, including differences in how attention modulates firing rate and differences in the attention-dependent reduction in response variability among the two classes of neurons. Narrow spiking neurons show a marked low frequency fluctuation in firing rate that is diminished by attention. Many broad spiking neurons show burstiness in their responses that is diminished by attention. The discovery of differences in attentional modulation of firing rate and neuronal noise represents a key step forward in developing circuit-level models of attention and visual processing.

Rudiger von der Heydt, PhD

11/12/2008 3:00pm
Mathew Diamond
Int'l for School Advanced Studies

“Whisker-mediated texture classification: The transformation carried out by the sensory system”

Categorization is one of the fundamental transformations carried out by sensory systems. This talk will be about how the tactile sensory system in rats represents stimuli, beginning at the whisker and progressing to hippocampus. We focus on a texture discrimination task. At the level of the sensory receptors and at early stages of the sensory pathway, neurons encode mainly the elemental physical features of stimuli, while at later stages they encode more complex properties relating to the convergence or intersection of elemental physical features. In parallel with this transformation, the sensory system also generates abstract representations, where neuronal activity encodes the identity and category of the stimulus. Explicit representations of identity and category are necessary both to guide behavior (because actions usually must be taken based on the identity of encountered objects rather than merely on their physical characteristics) and to store memories about the animal's own behavior.

Takashi Yoshioka, Ph.D.

11/21/2008 4:00pm
Jenny Read, PhD
Newcastle University

“The Neuronal Basis of Stereo Vision”

Stereo vision, or stereopsis, has several properties which make it an ideal model system for studying perception. Its underlying neuronal mechanisms are understood in more detail than those of perhaps any other perceptual ability, but we are still far from our goal of a computer algorithm reproducing human stereo vision. In this talk, I will aim to give an overview of key results concerning the neuronal substrate of stereo vision, concentrating on primary visual cortex and the stereo energy model of disparity tuning. I will begin with a brief introduction to stereo geometry and the underlying anatomy, and discuss in some detail what it means for a binocular neuron to qualify as a disparity sensor. I will discuss the stereo correspondence problem, posed most starkly in random-dot patterns, and the use of anti-correlated stimuli to probe whether global stereo correspondence has been achieved. I will argue that primary visual cortex shows several specialisations which make it suitable to perform the initial encoding of disparity, but that the neuronal correlate of stereo perception must take place in higher, extrastriate areas.

Rudiger von der Heydt, Ph.D.

11/24/2008 4:00pm
Teng Leng Ooi, PhD
Prof. Optometric Medicine
Pennsylvania College of Optometry

“Binocular surface perception affected by boundary contours and surface features”

The optical images of most objects have surrounding boundary contours and interior surface features (e.g., texture and color information). To understand the mechanisms of surface representation, a major component of mid-level visual processing, we need to understand how the visual system relies on the boundary contours and surface features to represent surfaces. With this major goal in mind, we have conducted a series of psychophysical experiments. In this presentation, I will first focus on how boundary contours and their projection geometry with respect to the two eyes affect global surface formation through filling-in and integration, and binocular rivalry perception. I will also present our findings on the role of color contrast and luminance contrast in surface representation, in particular, in surface interpolation.

Rudiger von der Heydt, PhD

12/08/2008 4:00pm
Daniel Margoliash, PhD
Prof. Organismal Biology&Anatomy
University of Chicago

“Early Events in Song Learning”

Bird song learning is an intensively studied model system for skill learning that requires evaluation of auditory feedback via an acquired sensory model ("template"). I will describe recent results demonstrating that early tutor song exposure specifies a small percentage of cells to be tutor-song selective. Both the template and auditory feedback affect discharge of premotor cells during sleep, giving insight into the interaction between sensory and sensorimotor plasticity that drives song learning.

Sliman Bensmaia, PhD

01/26/2009 4:00pm
Nicholas Hatsopoulos, PhD
Asst Prof Organismal Biology&Anatomy
University of Chicago

“Exploiting vision and proprioception to build and augment cortically-controlled brain machine interfaces”

A fundamental challenge in developing brain machine interfaces (BMIs) is building a mapping between the spatio-temporal patterns of cortical activity and movement because severely disabled patients cannot move. We demonstrate how visual observation of action can automatically trigger mirror-like responses in primary motor cortex (MI) that are similar to the responses that occur during action. Using these mirror-like responses, we show that effective real-time brain control can occur without concurrent arm movement. In the second part of my talk, I will present a set of experiments showing how proprioceptive feedback can be used to augment control in a BMI context. To date, most BMIs) have relied solely on visual feedback to guide the movement of an external device such as a cursor or robotic device. However, it is well known that patients suffering loss of proprioceptive feedback can move by relying on vision of their limbs, but their movements are typically slow, poorly coordinated, and require great concentration. Our approach involves guiding the animal’s own arm, along with the visual cursor, using a robotic exoskeletal device driven by command signals from the recorded ensemble of MI neurons. The animal’s arm indirectly provides information about the state (i.e. position and velocity) of the device (i.e. the cursor) being controlled and, thus, becomes an afferent channel to assist in the control of the cursor. Our data suggest that this form of “naturalistic” proprioception can improve performance by reducing the time required to acquire a target.

Sliman Bensmaia, PhD

02/09/2009 4:00pm
Alessandro Graziano, Ph.D.
Professor Center Neuroscience
University of California, Davis

“Mechanisms of cortical plasticity after massive somatosensory deafferentation”

The mechanisms responsible for long-term, massive reorganization of representational maps in primate somatosensory cortex after deafferentation are poorly understood. Sprouting of cortical axons cannot account for the extent of reorganization and withdrawal of axons of deafferented neurons permitting expression of previously silent synapses has not bee directly demonstrated. In monkeys, deafferented for two years, extensive withdrawal of axon terminals from thalamus and cortex can be detected a decade before visible atrophy of their parent neurons. Slow but inexorable progression of axonal withdrawal is a neurodegenerative phenomenon likely to be a powerful inducement to compensatory long-term plasticity. By acting on a widely divergent ascending projection system, this mechanism can explain the extent and long-term evolution of cortical reorganization and, with it, phantom sensations in spinal patients and amputees.

Steve Hsiao & Alfredo Kirkwood

02/20/2009 4:00pm
Israel Nelken, PhD
Professor of Neurology
Hebrew University, Jerusalem

“The Coding of Surprise in Auditory Cortex”

Neural responses in the auditory system depend not only on the current stimulus but also on the history of past stimulation. One form of this history-dependence is stimulus-specific adaptation (SSA), the reduction in the responses to a common sound relative to the same sound when rare, which has been described both in inferior colliculus and in auditory cortex. I will show that the trial-by-trial responses to such sequences are well described by a model in which responses depend on the surprise (the logarithm of the estimated probability of the stimulus), where stimulus probability is estimated from an internal representation of the past stimulation sequence. Initial results suggest that the duration of the past sequence that is necessary to account for the responses of single neurons is typically long (>10 stimulus presentations), but that the internal representations of the past are typically coarse. Furthermore, I will show that these responses cannot be accounted for by simple stimulus-specific adaptation models in which adaptation occurs in narrow frequency bands. Instead, the responses to a rare tone have a component which is sensitive to the regularity in the rest of the tone sequence.

Eric Young

02/23/2009 4:00pm
Adina Roskies, PhD
Department of Philosophy
Dartmouth College

"Freedom Despite Mechanism"

Many people think that advances in neuroscience will settle the question of whether or not we have free will. I argue that neuroscience will not be in a position to settle this question, and that suggestions that it can misunderstand both the limits of neuroscience and the nature of questions about freedom. However, when appropriately wedded to a philosophical picture, neuroscience can still be informative about freedom.

Drs. Gross and Hsiao

03/02/2009 4:00pm
Geoffrey Schoenbaum, PhD
Asst Prof Anatomy & Neurobiology
University of MD School of Medicine

Cancelled due to inclement weather. No resceduled date as yet. “The Orbitofrontal Cortex and Adaptive Behavior”

Damage to the orbital area of the frontal lobe disrupts adaptive, flexible behavior. These deficits have been attributed to impairments in response inhibition and or to the idea that orbitofrontal cortex is a rapidly-flexible associative learning area. Yet recent data contradict these accounts. In my talk, I will suggest an alternate account whereby orbitofrontal cortex supports flexible responding due to its more recently described role in signaling information about expected outcomes.

Marshall Shuler, PhD

03/23/2009 4:00pm
Tatiana Pasternak, Ph.D.
Prof. Neurobiology & Anatomy
University of Rochester

”What does prefrontal cortex ‘know’ about visual motion used in discrimination tasks?”

Perceptual decisions during visual discrimination tasks often require subjects to compare two or more sequentially presented stimuli. During such tasks the stimuli not only have to be processed, but also retained in memory and the comparison between the remembered and the current stimulus must be performed. To characterize the cortical circuitry sub-serving successful execution of such discrimination tasks we focused on speed and direction of visual motion and compared neuronal activity during such tasks in two interconnected cortical regions, motion processing area MT and a region associated with executive control and working memory, the prefrontal cortex (PFC). In this talk I will show that responses in PFC to motion during discrimination tasks are direction and speed selective, resembling activity recorded in MT. While both MT and PFC neurons carry reliable motion signals during the memory delay, such signals in individual neurons were largely transient, suggesting that the contribution of these neurons to stimulus maintenance could only be accomplished at the population level. During the comparison between the current and remembered stimuli, responses in both areas were modulated by the remembered stimulus, with signals in PFC trailing those recorded in MT, pointing to MT as a more likely source of sensory comparisons. Thus, neurons in areas MT and PFC make unique contributions to motion discrimination tasks and are likely to be functionally linked, a hypothesis supported by recent studies of the effects of unilateral PFC lesions on motion discrimination and on neuronal activity in the ipsilateral MT.

Ed Connor

03/25/2009 4:00pm
Brian Knutson
Prof Psychology & Neuroscience
Stanford University

“Anticipatory affect: Neural correlates and consequences for choice”

"Anticipatory affect" refers to emotional states that people experience while anticipating significant outcomes. Historically, technical limitations have made it difficult to determine whether anticipatory affect influences subsequent choice. Recent advances in the spatio-temporal resolution of functional magnetic resonance imaging, however, now allow researchers to visualize changes in neural activity seconds before choice occurs. I will review evidence that activation in specific brain circuits changes during anticipation of monetary incentives, that this activation correlates with affective experience and that activity in these circuits may influence subsequent choice. Together, emerging findings support a neurally plausible framework for understanding how anticipatory affect might influence choice.

Veit Stuphorn, Ph.D.

03/30/2009 4:00pm
David Freedman, Ph.D.
University of Chicago

“Encoding of the Non-Spatial Behavioral Significance of Visual Stimuli in Parietal Cortex”

We have an impressive capacity to recognize the behavioral significance, or category membership, of a wide range of sensory stimuli. This ability is critical because it allows us to respond appropriately to stimuli that we encounter in our interactions with the environment. Recently, we recorded from lateral intraparietal (LIP) neurons during a categorization task in which 360º of motion directions were grouped into two arbitrary categories that were divided by a learned category boundary. These recordings revealed that LIP neurons robustly encoded stimuli according to their learned category membership, suggesting that parietal visual representations can reflect abstract information about the learned significance of visual stimuli. More recent work has also revealed strong neuronal encoding of the learned associations between pairs of static visual-shapes in LIP, suggesting that parietal cortex likely plays a role in encoding the significance of both spatial (e.g. motion or space) and non-spatial (e.g. shape) stimuli.

Sliman Bensmaia, Ph.D.

04/06/2009 4:00pm
Jaynie F.Yang, PhD
Professor, Physical Therapy
University of Alberta

“Neural Control of Human Walking as Seen Through the Study of Infant Stepping”

Considerable advances have been made to elucidate the spinal and brainstem mechanisms underlying quadrupedal locomotion in the last century. Yet the applicability of these findings to the bipedal human remains unclear. One of the reasons for the difficulty with studying the human is that the influence of the higher centres in the nervous system (i.e., cerebrum, cerebellum) on walking cannot be easily separated from that of the lower centres (brainstem and spinal cord). To overcome this problem, my colleagues and I have used the stepping behaviour in young infants as a window to the organization of the spinal/brainstem control for human walking. Human infants are born with a relatively immature cerebrum and cerebellum compared to the spinal cord. They exhibit stepping behaviour from birth. I will discuss how we have used simple mechanical perturbations to study the sensory control of stepping and the relationship between pattern generating circuits for each leg. More recently, we have used crawling to study the limitations imposed by the immature circuitry for controlling coordination during quadrupedal locomotion. Overall, we have found remarkable similarities in the control of locomotion between human infants and quadrupeds.

Amy Bastian, PhD

04/20/2009 4:00pm
Zygmunt Pizlo, Ph.D.
Prof., Psychological Sciences
Purdue University

“A New Approach to 3D Shape Perception”

Perception of 3D shapes is one of the most fundamental, as well as one of the most difficult problems in perception. The difficulty is related to the fact that the visual system has to produce veridical interpretation of 3D shapes (i.e., interpretation that agrees with the 3D shapes “out there”) based on the 2D information provided by the retinal images. Conventional theories assumed that the 3D shape percept is produced by reconstructing visible surfaces of an object from depth cues. None of these theories, however, could actually explain veridicality of shape perception in the presence of changes of the viewing direction (i.e., the shape constancy phenomenon). The new approach is based on the assumption that human shape perception critically depends on a priori knowledge (aka priors) about shapes of physical objects. We identified four general-case shape priors: symmetry, maximal 3D compactness, minimum surface area, and planarity of contours (maximal compactness refers to maximum volume for a given surface area). These priors are applied to 2D contours extracted from a single 2D image of the 3D shape. Finding contours in the image is referred to as figure-ground organization and it is still an unsolved problem. A computational model based on our shape priors leads to veridical recovery of a 3D shape and its performance is very similar to the performance of human subjects. I will show that shape priors are very effective because shape, as a visual characteristic, is very complex. This is different from all other perceptual characteristics, such as lightness, color, size and motion. The lecture will be concluded with a brief discussion of the interaction of symmetry prior with binocular disparity. This interaction leads to a new model of binocular shape perception.

Rudiger von der Heydt, Ph.D.

04/27/2009 4:00pm
Lloyd B. Minor, M.D.
Andelot Professor
JHU School of Medicine

“Learning and Plasticity in the Balance System: The Physiology of Vestibular Compensation”

Processes of vestibular compensation are responsible for improvement in the performance of the vestibuloocular reflex (VOR) after inputs from one labyrinth are eliminated. The symmetry of the VOR is restored for responses to low-frequency, low-velocity head movements. Asymmetries in the VOR are noted for responses to high-frequency, high-acceleration rotations. Responses to these stimuli return to relatively normal levels for rotations that are excitatory with respect to semicircular canals on the intact side but remain markedly diminished for inhibitory rotations. The improvement in the performance of the VOR elicited by excitatory rotations is related to contributions from phasic inputs from the intact side to the reflex pathways. The long-term goal of this research is to understand, at the level of single neuron physiology, the processes that mediate vestibular compensation. Our studies have shown that the proportion of highly phasic, irregularly discharging afferents is increased on the intact side following unilateral labyrinthectomy. There are also changes in the responses of position-vestibular-pause (PVP) neurons, interneurons in the VOR pathway, following unilateral labyrinthectomy. These PVP neurons respond to neck proprioceptive inputs during the acute stage of compensation and carry signals indicative of an efference copy signal of the motor command. These extravestibular inputs have an important role in the compensation process. (Supported by NIH R01 DC02390).

Steven Hsiao, Ph.D.

05/04/2009 4:00pm
Jonathan D. Wallis, Ph.D.
University of California, Berkeley

“Reward Processing in Prefrontal Cortex”

The frontal cortex comprises at least 18 cytoarchitectonically distinct regions and accounts for about 30% of our total cortical area, yet despite this anatomical complexity, little progress has been made in linking specific functions to specific anatomical subregions. Single-unit neurophysiology provides us with excellent spatial and temporal resolution to dissociate the functional properties of these regions. Early neurophysiological studies emphasized the importance of orbitofrontal cortex in representing reward information, but reward-related signals have subsequently been discovered in all frontal areas. The current challenge is to identify the role that reward plays in these different areas. I will describe several experiments that have enabled us to differentiate the role of reward in lateral prefrontal cortex from that in orbital and medial areas. Our results are consistent with reward signals in lateral prefrontal areas serving a role in the allocation of attentional resources, while reward signals in orbital and medial areas have properties that may underlie decision-making. In closing, I will discuss experiments that we are currently conducting aimed at dissociating the differential contribution of orbital and medial areas to decision-making.

Veit Stuphorn, Ph.D.

05/18/2009 4:00pm
Alipasha Vaziri, Ph.D.
Research Scientist
Howard Hughes Medical Institute

“Sculpting Light for Non-Linear Excitation beyond the Gaussian Optics; Enabling 3D Super-Resolution Imaging and More”

Recent advances in optical microscopy have enabled imaging of biological samples beyond the diffraction limit at nanometer resolution. A general feature of most of the techniques based on photoactivated localization microscopy (PALM) or stochastic optical reconstruction microscopy (STORM) has been that the imaging depth is limited to a fraction of an optical wavelength. However, to study whole cells, the extension of these methods to a 3D-super-resolution technique is required. We have overcome this limitation by using a two-photon illumination technique called temporal focusing in which the spectral properties of the pulse are used to control its axial intensity distribution in space. Using temporally focused beams we have generated super-resolution images in 3D over an axial range of ~10μm in various biological samples. As discussed the extension of temporal focusing to more general spatial light distributions in 3D (“light sculpting”) has a number of import applications in structural and function bio-imaging.

Josh Vogelstein

06/11/2009 4:00pm
Barry E. Stein, Ph.D.
Prof. Neurobiology & Anatomy
Wake Forest School of Medicine

“How the brain integrates information from different senses to produce adaptive behavior”

Midbrain neurons in the superior colliculus (SC) are able to synthesize information from different senses, thereby dramatically altering their responses to external stimuli (visual, auditory and somatosensory) and the behaviors that depend on them. They are an excellent model for understanding multisensory integration. It is through this process that cross-modal stimuli whose properties make it likely that they are derived from the same event markedly enhance the physiological responses of these neurons. This enhances the salience of the event. As a result, behavioral performance on orientation and localization tasks is also enhanced. Often, these performance benefits are quite striking, substantially increasing the probability of responding appropriately to biologically relevant events. In contrast, cross-modal stimuli that are more likely to be associated with different external events either yield no multisensory integration or degrade physiological responses and, thus, the salience of their signal in the brain. They also degrade behavioral performance. However, contrary to some theories of sensory development, these multisensory integrative abilities are not present in the newborn’s brain. Furthermore, their fundamental characteristics are not pre-specified. Rather, the acquisition of multisensory integration capabilities is a gradual postnatal process that depends heavily on at least two factors: the development of a cooperative interaction between descending projections from different sensory (visual, auditory and somatosensory) subdivisions of association cortex, and the acquisition of extensive postnatal experience with cross-modal cues. These factors are used by the brain to craft the underlying neural circuit and the fundamental principles that guide multisensory integration in order to adapt them to the environment in which they will be used. The anatomical and physiological properties of this circuit, its developmental antecedents, and the likely component that acquires the relevant cross-modal experience will be discussed.

Manuel Gomez-Ramirez, Ph.D.

09/14/2009 4:00pm
Friedrich T. Sommer, Ph.D.
Adjunct Associate Professor
University of California, Berkeley

“Whole-cell recordings in thalamus reveal how retinal gamma oscillations convey additional visual information”

A recent study will be presented that explores early visual coding by reexamining the widely held assumption that visual input is encoded by firing rate changes that are locked to stimulus events. The project was motivated by work that had shown that retinal firing patterns are not only influenced by external stimuli but also by dynamics of intrinsic neuronal networks such as oscillations. We developed methods to assess whether retinal neurons are able to encode information by using spike timing relative to ongoing oscillations and if this information is relayed to the cortex by the thalamus. Then, we used our method to analyze data obtained in collaboration with the Hirsch lab at USC who used whole-cell recordings in vivo to measure retinal inputs and the spikes they evoke from single thalamic relay cells. Our analysis showed that relay cells transmit information using two separate channels. The first channel, as has been described before, transmits information localized to the neural receptive field by using changes in spike rate evoked by the stimulus; this channel relays visual signals slower than 30 Hz. The second, novel, channel encodes information by using spike timing relative to intrinsic retinal oscillations and occupies a separate frequency band, in the gamma range (50-70 Hz). Remarkably, the amount of information in the second channel could match or even exceed that conveyed by the first, a result that we were able to reproduce with a simple model. Because retinal oscillations involve large-scale networks, the novel channel could convey "gist", or contextual aspects, of the visual scene.

Ernst Niebur

09/21/2009 4:00pm
John Gale, Ph.D.
Research Fellow Dept. Neurology
Harvard University

“Encoding the How and Why of Behavior: The Role of the Anterior Striatum in Associative Learning”

The process of associative learning, wherein the brain links sensory stimuli with specific motor behaviors and expected rewards, is fundamental to survival; this is true for simple activities, such as learning to reach for a container of water to quench thirst, as well as for complicated patterns of behavior, such as learning to swing a bat to hit a ball. Recent evidence suggests a critical portion of this process is encoded in the anterior striatum and mediated through the action of the neurotransmitter dopamine. The anterior striatum consists of a dorsal region that includes the caudate nucleus and a ventral region that includes the nucleus accumbens. Anatomical, neurochemical and brain lesion data suggest that the dorsal and ventral regions of the anterior striatum serve different roles in associative learning; however, the precise physiological mechanisms at the level of single neurons are unknown. This study examines single neuronal activities from the dorsal and ventral striatum in non-human primates as they perform a visual-motor associative learning task. The findings suggest that the dorsal and ventral striatum encode distinct but complementary types of information, both fundamental to the process of visual-motor associative learning.

Ernst Niebur

09/28/2009 4:00pm
Maurice Chacron, Ph.D.
McGill University

“The contribution of short-term depression and subthreshold membrane conductances to directional selectivity in midbrain neurons”

Directional selectivity to moving stimuli has been observed in a variety of organisms ranging from insects to primates and is thought to provide a neural correlate of motion perception. As originally proposed by Reichardt, directional selectivity requires differential temporal filtering of at least two inputs from different spatial locations followed by nonlinear integration. Here we show how different time constants of short-term depression across the receptive field can give rise to a directional bias that is then enhanced by subthreshold calcium channels in midbrain neurons of weakly electric fish. We furthermore show that bursts of action potentials can selectively carry directional information and may be used by higher order neurons to decode this information. Both subthreshold conductances and burst firing are ubiquitous within the CNS and may be used to convey directional information in other systems.

Sarah Stamper

10/26/2009 4:00pm
Chou Po Hung
Assistant Professor
National Yang Ming University

“The visual shape alphabet and its neural population code”

One of the main bottlenecks to understanding information processing in higher cortical areas is that we do not know the key dimensions which underlie coding in these areas. At the end of the macaque ventral visual pathway, the anterior inferotemporal cortex (AIT) has been hypothesized to encode anywhere from 36 to a nearly infinite number of shape dimensions via multiple scales of functional organization. Using multi-electrode array recordings and multivariate analysis methods, we are examining the organization and coding within ~2 mm2 sections of AIT. We show that neurons are loosely clustered by key dimensions in shape space and that it is possible to extract these dimensions while controlling for investigator bias. These shape dimensions are marked by continuities and fractures in the topographic map. Neurons on opposite sides of a fracture appear to encode opposite features, suggesting a functional organization which differs from traditional columnar models of AIT and may resemble direction maps in cat area 18. We are currently integrating these results with computational models of visual object recognition, with the goal of obtaining a universal alphabet of visual shape dimensions.

Steve Hsiao

11/02/2009 4:00pm
Scott Thompson
Professor, Dept. of Physiology
University of Maryland, Baltimore

“Excitatory synapses get the blues: dysregulation of serotonin signaling in depression”

Although the serotonin hypothesis of depression has been with us for 50 years, we still don't know what serotonin does in the brain, what is wrong in the depressed brain, and how serotonin dysfunction can affect cognitive function. I will present unpublished results indicting that serotonin engages the cellular machinery of learning to regulate the strength of excitatory synaptic transmission and that this process is altered in depression and key to the therapeutic actions of antidepressants.

Alfredo Kirkwood

11/09/2009 4:00pm
Andreas Burkhalter, PhD

Washington University

“Maps, Streams and Circuits in Mouse Visual Cortex”

At the initial stage of visual processing, images are decomposed into fragments that are represented by small retinal receptive fields. The task of the visual cortex is to put these image fragments back together again and to create percepts of distinct objects. It is well established that this process involves the integration of information that is distributed across many functionally specialized areas, however, the underlying synaptic mechanisms by which image features are grouped are incompletely understood. We are studying these processes in the visual cortex of mice. Although mice are better known for their exquisite sense of touch than their vision, we found that they have a highly complex visual cortex whose organization is remarkably similar to that of primates. For example, our studies have shown that mice have 10-12 functionally distinct visual areas. This is less than the three dozens found in primates, but many more than depicted in previously published maps. Our studies have further shown that these areas are connected to a cortical network which is subdivided into dorsal and ventral subnetworks. We have performed single unit recordings in different dorsal and ventral stream areas and have found that neurons in ventral areas have higher visual acuity than neurons in dorsal areas. In contrast, neurons in dorsal areas neurons are selective for higher speeds of visual motion and higher temporal frequencies than neurons in ventral areas. The topographical distribution of these properties resembles that found in primates in which the dorsal processing stream is specialized for object motion and the dorsal stream is concerned with texture discrimination. Our detailed studies of laminar projections have shown that areas are interconnected by feedforward and feedback circuits and that these connectivity patterns can be used to derive an areal hierarchy. Recordings in different areas have further shown that this structural hierarchy corresponds to a functional hierarchy in which receptive fields in higher areas are larger than in lower areas. Similar to more highly visual mammals, we have found that visual responses not only depend on bottom up feedforward inputs, but responses are influenced by intraareal lateral connections and by top-down feedback inputs from higher to lower cortical areas. Whole cell patch clamp recordings in slices have shown that these modulatory feedback influences are generated by synaptic circuits that generate weaker inhibition than feedforward circuits. These differences suggest that feedforward circuits are specialized for high temporal resolution, whereas feedback circuits increase the gain and width of the integration window which enables the association of inputs and supports perceptual grouping.

Rudiger von der Heydt

11/12/2009 4:00pm
John Phillips, Ph.D.
Professor Deparment Optometry
University of Auckland

“Optical inhibition of human eye growth and myopia progression”

The abnormal eye enlargement which causes myopia (short-sight) also increases the risk of ocular morbidity. This lecture first reviews the optically-guided processes controlling eye growth, which normally ensure that the developing eye remains in focus as it grows. It then reports results of recent clinical studies in children with progressing myopia. These studies employ lenses, including soft contact lenses which incorporate a novel optical design aimed at inhibiting abnormal eye growth and thus myopia progression.

Dr. Steven Hsiao

11/16/2009 4:00pm
Ralph Etienne-Cummings, Ph.D.
Prof. Electrical & Computer Eng.
Johns Hopkins University

“Implementing Models of the Primate Visual Cortex in Silicon”

Imagine having a car that is able to drive itself or a robot that is able to perform tasks ranging from the tedious (e.g. housekeeping) to the dangerous (flying military aircrafts) and even to the difficult (building space stations) without human supervision. Today’s robots make certain tasks easier but still require remote supervision and control by humans. Intelligent robots need to be able to interact with objects in their surroundings with minimal human involvement. This involves three steps: (1) detecting the presence of the object, (2) recognizing the object – determining whether it is an obstacle to be avoided, an item to be retrieved, or perhaps a tool required for a particular task, and (3) tracking the trajectory of the object – determining how and when to react to it. While these steps are computationally difficult, humans and other primates are able to perform them easily. They are able to rapidly and effortlessly identify and categorize diverse objects in cluttered scenes under widely varying viewing conditions, such as changes in position, rotation and illumination. Although the density of processing elements in engineered systems have approximated, in some cases surpassed, those of biological systems, they are still unable to match the level of proficiency and speed of biological visual systems. Substantial research is still needed at the computational architecture level to move the performance of engineered systems towards that of biology. The work I will present takes some strides in that direction. We are working towards the development of an autonomous, continuous-time visual system that emulates visual information processing in the primate visual cortex. This multi-stage system will utilize large-scale arrays of identical silicon neurons to build a biologically-plausible model of its biological counterpart. In this talk, I will describe (1) the design a neural array transceiver with neurons and synapses that more closely mimic biology, (2) implement silicon facsimiles of cortical simple cells, complex cells, and composite feature cells according to the hierarchical model of primate visual cortex of proposed by Riesenhuber and Poggio, and (time permitting) (3) implement neural algorithms analogous to cross-correlation and Kalman filtering for object detection and tracking respectively.

Dr. Steven Hsiao

02/22/2010 3:30pm
Jeremy Wolfe & Christof Koch
Harvard University & Cal Tech

3:30 p.m "Visual Search goes Scenic" 5:00 "On the Relationship Between Consciousness and Attention"


Ernst Niebur/Howard Egeth

03/01/2010 4:00pm
Dean Buonomano

“The neural basis of timing and the processing of time-varying stimuli”

The brain’s ability to seamlessly assimilate and process spatial and temporal information is critical to most behaviors, from speech and music recognition, to motor coordination. We must understand the mechanisms that allow the brain to process and tell time if we are to develop general theories of sensory and motor processing, as well as of cognition. Our work suggests that cortical circuits are inherently capable of telling time as a result of the interaction between incoming stimuli and the current internal state of a network – defined both by ongoing activity and time-dependent neural properties, such as short-term synaptic plasticity (we refer to these models as state-dependent network, SDN). Computational, in vitro, and psychophysical studies aimed at examining SDN models of spatio-temporal processing will be presented. Additional findings that suggest that in vitro cortical networks can ‘learn’ the temporal patterns to which they are exposed will also be presented

Drs. Wang & von der Heydt

03/29/2010 4:00pm
Kathleen Cullen, Ph.D.
Professor Dept of Physiology
McGill University

“How actions alter sensory processing: active sensation in the vestibular system”

Our vestibular organs are simultaneously activated by our own actions as well as by stimulation from the external world. The ability to distinguish sensory inputs that are a consequence of our own actions (vestibular reafference) from those that result from changes in the external world (vestibular exafference) is essential for perceptual stability and accurate motor control. A main focus of research in our laboratory has been to understand how the brain distinguishes between vestibular reafference and exafference. In our experiments, single-unit recordings are made in alert rhesus monkeys during passive and voluntary (i.e., active) head movements. We find that neurons in the first central stage of vestibular processing (vestibular nuclei), but not the primary vestibular afferents, distinguish between active and passive movements. Notably, neuronal sensitivities to head motion are reduced by 70% during active motion. To better understand how neurons differentiate active from passive head motion, we systematically tested neuronal responses during different combinations of passive and active motion resulting from rotation of the head-on-body and/or head-and-body in space. We found that during active movements, a cancellation signal is generated when the activation of proprioceptors matched the motor-generated expectation. We hypothesized that this match is computed in the cerebellum, which in turn sends a cancellation signal to the vestibular nuclei during active motion to suppress reafference. To test this hypothesis, we have performed a parallel series of experiments in the cerebellum. Consistent with our hypothesis, neurons in the rostral fastigial nucleus, the most medial of the deep cerebellar nuclei, show a decrease in sensitivity during active as compared to passive motion. In addition, our most recent results suggest that vestibular reafference is no longer effectively cancelled in the vestibular nuclei following reversible inactivation of the rostral fastigial nucleus. Thus, taken together our results favor the proposal that the cerebellum plays a causal role in the cancellation of reafference, and provide novel insights into how reafference is suppressed in mammalian sensory systems at the level of single neurons.

Dr. Lloyd Minor

04/26/2010 4:00pm
Charley Della Santina
Associate Professor
Johns Hopkins School Medicine

“Restoring the Sixth Sense in 3D: Progress Toward a Bionic Vestibular Labyrinth”

Unlike sight, hearing, taste, touch and smell, vestibular (inner ear balance) sensation seldom rises to the level of conscious perception. However, bilateral loss of vestibular sensation chronically disables those who fail to compensate through sensory substitution. Without sensory input to reflexes that normally stabilize the eyes and body, affected individuals suffer blurry vision during head movement, postural instability and chronic disequilibrium. While some retain enough residual sensation to compensate for their loss through rehabilitative training paradigms that recruit adaptive mechanisms, profound loss condemns others to chronic imbalance for which no adequate treatments exist. Supplanting damaged hair cells by directly transducing head movement into vestibular afferent activity, a “bionic vestibular labyrinth” could restore balance and quality of life to otherwise perpetually dizzy patients. This presentation will briefly review the impact of bilateral vestibular sensory loss, discuss physiologic and anatomic studies supporting the feasibility of prosthetic vestibular nerve stimulation, and describe a multichannel vestibular prosthesis (MVP) designed to restore vestibular sensation of head rotation in all directions. Similar to cochlear implants in concept and size, the Johns Hopkins MVP includes miniature gyroscopes to sense head rotation, a microcontroller to process and remap inputs into desired afferent firing rates, and stimulation circuits to drive multiple electrodes implanted within the vestibular labyrinth. In rodents and nonhuman primates rendered vestibular-deficient via treatment with gentamicin (an antibiotic that often damages vestibular hair cells in humans), the Johns Hopkins MVP restores the vestibulo-ocular reflex for head rotations about any axis of rotation in 3-dimensional space. Looking ahead, our efforts now focus on two parallel paths. First, we are addressing bioengineering issues prerequisite to human implantation. Second, we seek to use the MVP as a tool to study the neural and behavioral mechanisms that underlie motor learning after loss, restoration and reorientation of vestibular sensation.

Dr. Lloyd Minor

05/17/2010 4:00pm
Sergio Neuenschwander, Ph.D.
Max Planck Inst. Brain Research

"Building expectations: New vistas for gamma oscillations"

Gamma oscillatory responses have been linked to visual processes demanding selective attention. In my talk, I will present our findings about gamma responses induced by expecting a behaviorally relevant stimulus event. For this study, spiking activity and local field potentials were recorded from area V1 of awaking behaving monkeys. Temporal expectation effects were studied in two tasks. In a first task, fixation point color change (behaviorally relevant event) occurred at a fixed time, enabling the monkey to estimate the proximity of the visual event requiring the behavioral response. In a second task, the proximity of change was signalized by a slight luminance increase of the fixation point prior to its change in color. Expectancy led to clear enhancement of gamma responses to stimuli presented over the receptive fields. These effects were spatially widespread, since comparable magnitudes of change occurred for both central and peripheral receptive fields. As I will discuss, these results support the existence of a new gain mechanism capable of widespread, spatially non-selective modulation of gamma activity.

Rudiger von der Heydt

05/24/2010 4:00pm
Andreas Tolias
Baylor College of Medicine



Veit Stuphorn

06/21/2010 4:00pm
David Ginty, Ph.D.
Professor of Neuroscience
JHU School of Medicine

"A molecular-genetics strategy for understanding the sense of touch"


Dr. Steven Hsiao

09/20/2010 4:00pm
Jeremy Wolfe, Ph.D.
Professor of Ophthalmology
Harvard School of Medicine

"Visual Search Gets Real: From the Lab to the Airport to the Radiology Suite”

We are built to search. Our ancestors foraged for food. We search for pens, coffee cups, and cars in parking lots. Even when a desired object is in plain view, we search because we can only recognize one (or a very few) objects at one time. We are adapted to these limits and we manage pretty well, using attentional processes that guide us to likely targets and that enable us to abandon search when the target is not present. Modern civilization has developed a number of visual search tasks that pose interesting challenges to our search abilities. We ask radiologists to find cancer and airport screeners to find threats under conditions where the targets are rare, ambiguous, and really important. In this talk, I will discuss some of the processes that govern everyday search tasks and I will consider what can happen when these processes confront the demands of those modern, socially important search tasks.

Dr. Rudiger von der Heydt

09/27/2010 4:00pm
Nelson Spruston, Ph.D.
Prof. Neurobiology & Physiology
Northwestern University

"Neural signal integration in pyramidal neurons and inhibitory interneurons in the hippocampus"

Pyramidal neurons process information in neural circuits by integrating thousands of excitatory and inhibitory synaptic inputs. I will describe experiments that probe the properties of dendrites and the synaptic inputs they receive. A combination of experimental and computational modeling has provided us with insights into how these inputs are integrated in the dendritic tree of hippocampal pyramidal neurons. I will also describe a new series of experiments that have revealed novel integrative properties of interneurons in the hippocampus and neocortex.

Dr. Alfredo Kirkwood

10/04/2010 4:00pm
Barbara Landau, Ph.D.
Professor Cognitive Science
Johns Hopkins University

“Genes, brains, and spatial representation: Evidence from Williams Syndrome”

Our experience of the spatial world is a unitary one—we perceive objects and layouts, we remember them and act on them, and we can even talk about them with ease. Despite this impression of seamlessness, spatial representations in human adults appear to be specialized in domain-dependent manner, engaging different properties and computational mechanisms for different functions. In this talk, I will present evidence that this domain-specific specialization in cognitive function emerges early in development and is reflected in patterns of breakdown that occur under genetic defect. I will offer evidence from Williams syndrome—a relatively rare genetic syndrome that gives rise to an unusual profile of severely impaired spatial representation together with spared language. Results from a variety of spatial domains – including object representation, motion perception, action, navigation, and spatial language-- show a strikingly uneven profile of sparing and deficit within spatial representations, consistent with the idea that specialization of function drives development and breakdown. These findings raise a crucial question: Can specific genes target specific aspects of cognitive structure? I will argue that, to answer this question, we need to take seriously the idea of cognitive structure as well as the mechanisms underlying normal development. I will present a speculative hypothesis that does so, explaining the WS cognitive profile in terms of the normal development of spatial representations, together with specific abnormal outcomes resulting from the WS gene deletion.

Dr. Rudiger von der Heydt

10/11/2010 4:00pm
Julio Martinez-Trujillo
Assistant Professor of Physiology
McGill University

"Physiological Mechanisms of Attention in the Primate Brain"

The primate nervous system is the result of an evolutionary process during which environmental factors have exerted selective pressures on brain structure and function favoring adaptations and mechanisms that make information processing efficient. One example is the cognitive mechanism of attention through which the brain selectively enhances the processing of behaviorally relevant signals while suppresses irrelevant ones. Attention is ubiquitous to many species, and likely evolved as a solution to a mismatch between the limited processing capacity of sensory and cognitive brain systems, and the overwhelming amount of signals entering the brain at any given moment. In this talk I will concentrate on the results of two main experiments addressing two main questions concerning the physiological mechanisms of attention: 1) how attention modulates sensory processing in visual neurons within area MT of awake macaques, and 2) how neurons in the prefrontal cortex of these animals enhance the processing of behaviourally relevant stimuli while suppress the processing of irrelevant ones.

Ernst Niebur

10/18/2010 4:00pm
Matthew Shapiro, Ph.D.
Associate Professor
Mt. Sinai School of Medicine

"Memory Retrieval Mechanisms"

The purpose of memory is to guide adaptive behavior. To do so, memory retrieval is informed by means, motives, and opportunities. In familiar situations, memory integrates biological and psychological goals with behavioral strategies to realize available resources. This talk will describe some of the neuronal mechanisms that may contribute to memory retrieval, focusing on coding by hippocampal and prefrontal cortical circuits. Neurons in both regions signal temporally extended sequences directed toward expected goals, as well as the rules or strategies that guide ongoing behavior. Circuits in both regions oscillate in similar frequencies, and the coherence of the oscillations change during learning and memory performance. The data suggest that means, motive, and opportunity are integrated through the bidirectional interactions between prefrontal and hippocampal circuits. Future experiments will test if, in familiar situations, hippocampal "recognition codes" activate prefrontal circuits that signal expected outcomes or adaptive strategies, which in turn guide hippocampal coding toward goal-related, adaptive, memory retrieval.

Dr. James Knierim

10/25/2010 4:00pm
Diego Contreras
Associate Professor Neuroscience
University of Pennsylvania

”Cellular Mechanisms of Sensory Driven and Spontaneous Gamma Oscillations in Sensory Neocortex”

The seminar will show data in vivo on the cellular responses and the distribution in the depth of the cortex of spontaneous and sensory driven gamma oscillations. The data was obtained in mouse primary visual cortex and rat barrel cortex. In mouse V1, the oscillations are dependent on contrast but not on stimulus orientation and are greatly enhanced by ketamine. In rat barrel cortex, gamma oscillations emerge as a transformation of naturalistic high frequency (90-200 Hz) inputs that result from "slip-stick" ressonant motion of the whiskers. The frequency transformation in barrel cortex is dependent on GABAergic inhibition in cortical Layer 4 since it is blocked by picrotoxin. Finally, intracellular data in cat in vivo, demonstrate the presence of various types of intrinsic cellular generators of gamma oscillations in cortex and thalamus, as well as synchronized thalamocortical oscillations.

Alfredo Kirkwood

11/01/2010 4:00pm
Roozbeh Kiani
Department of Neurobiology
Stanford University

“Computation and representation of choice certainty in the parietal cortex”

The degree of confidence in a decision provides a graded and probabilistic assessment of expected outcome. Although neural mechanisms of perceptual decisions have been studied extensively in primates, little is known about the mechanisms underlying choice certainty. In my talk, I will show that the same parietal cortex neurons that represent formation of a decision encode certainty about the decision. I will briefly explain the neural computations that underlie a direction discrimination task and will propose a model for the calculation of choice certainty.

Manuel Gomez-Ramirez

01/24/2011 4:00pm
Josh Neunuebel
Postdoctoral fellow
MBI and Janelia Farm

“Tracking the Flow of Information through the Hippocampal Formation in the Rat”

The hippocampus receives input from upper levels of the association cortex and is implicated in many mnemonic processes, but the exact mechanisms by which it codes and stores information is an unresolved topic. This work examines the flow of information through the hippocampal formation while attempting to determine the computations that each of the hippocampal subfields performs in learning and memory. The formation, storage, and recall of hippocampal-dependent memories theoretically utilize an autoassociative attractor network that functions by implementing two competitive, yet complementary, processes. Pattern separation, hypothesized to occur in the dentate gyrus (DG), refers to the ability to decrease the similarity among incoming information by producing output patterns that overlap less than the inputs. In contrast, pattern completion, hypothesized to occur in the CA3 region, refers to the ability to reproduce a previously stored output pattern from a partial or degraded input pattern. Prior to addressing the functional role of the DG and CA3 subfields, the spatial firing properties of neurons in the dentate gyrus were examined. The principal cell of the dentate gyrus, the granule cell, has spatially selective place fields; however, the behavioral correlates of another excitatory cell, the mossy cell of the dentate polymorphic layer, are unknown. This report shows that putative mossy cells have spatially selective firing that consists of multiple fields similar to previously reported properties of granule cells. Other cells recorded from the DG had single place fields. Compared to cells with multiple fields, cells with single fields fired at a lower rate during sleep, were less likely to burst, and were more likely to be recorded simultaneously with a large population of neurons that were active during sleep and silent during behavior. These data suggest that single-field and multiple-field cells constitute at least two distinct cell classes in the DG. Based on these characteristics, we propose that putative mossy cells tend to fire in multiple, distinct locations in an environment, whereas putative granule cells tend to fire in single locations, similar to place fields of the CA1 and CA3 regions. Experimental evidence supporting the theories of pattern separation and pattern completion comes from both behavioral and electrophysiological tests. These studies specifically focused on the function of each subregion and made implicit assumptions about how environmental manipulations changed the representations encoded by the hippocampal inputs. However, the cell populations that provided these inputs were in most cases not directly examined. We conducted a series of studies to investigate the neural activity in the entorhinal cortex, dentate gyrus, and CA3 in the same experimental conditions, which allowed a direct comparison between the input and output representations. The results show that the dentate gyrus representation changes between the familiar and cue altered environments more than its input representations, whereas the CA3 representation changes less than its input representations. These findings are consistent with longstanding computational models proposing that (1) CA3 is an associative memory system performing pattern completion in order to recall previous memories from partial inputs, and (2) the dentate gyrus performs pattern separation to help store different memories in ways that reduce interference when the memories are subsequently recalled.

Jim Knierim

01/31/2011 4:00pm
Matthew Botvinick
Associate Professor, Psychology
Princeton University

“Hierarchical Reinforcement Learning”

Research on human and animal behavior has long emphasized its hierarchical structure, according to which tasks are comprised of subtask sequences, which are themselves built of simple actions. The hierarchical structure of behavior has also been of enduring interest within neuroscience, where it has been widely considered to reflect prefrontal cortical functions. In recent work, we have been reexamining behavioral hierarchy and its neural substrates from the point of view of recent developments in computational reinforcement learning. Specifically, we've been considering at a set of approaches known collectively as hierarchical reinforcement learning, which extend the reinforcement learning paradigm by allowing the learning agent to aggregate actions into reusable subroutines or skills. A close look at the components of hierarchical reinforcement learning suggests how they might map onto neural structures, in particular regions within the dorsolateral and orbital prefrontal cortex. It also suggests specific ways in which hierarchical reinforcement learning might provide a complement to existing psychological models of hierarchically structured behavior. A particularly important question that hierarchical reinforcement learning brings to the fore is that of how learning identifies new action routines that are likely to provide useful building blocks in solving a wide range of future problems. Here and at many other points, hierarchical reinforcement learning offers an appealing framework for investigating the computational and neural underpinnings of hierarchically structured behavior. In addition to introducing the theoretical framework, I'll describe a first set of neuroimaging and behavioral studies, in which we have begun to test specific predictions.

Veit Stuphorn

02/07/2011 4:00pm
Helen Scharfman
Prof Psychology, Physiology, Neuroscience
New York University Langone Medical Center

“The role of hilar neurons of the dentate gyrus in normal hippocampal function and disease”

The dentate gyrus is critical to normal cognitive function, contributing to learning and memory functions of the entorhinal-hippocampal circuitry in many ways, some of which are still debated. In addition, it has been suggested that the dentate gyrus contributes to mood, emotion and anxiety, and that pathology in the dentate gyrus contributes to diseases such as Alzheimer’s disease, schizophrenia and temporal lobe epilepsy. Most investigators consider the functions of the dentate gyrus to be mediated by its primary cell type, the granule cell, but there are neurons in the adjacent hilar region that may be just as important, although they are far less numerous. This seminar will discuss the types of hilar neurons, their role in dentate gyrus circuitry, and evidence that hilar neurons play a critical role in the normal animal, as well as pathological conditions.

Dr. James Knierim

02/27/2012 4:00pm
Soojin Park, PhD
JHU Cog Sci

“The integrative nature of scene representation”

A central question in human visual cognition is how we construct a coherent scene percept, rapidly recognizing scene structure while integrating multiple views that constantly change over time. Past studies have shown that a number of brain regions are involved in scene understanding, including the parahippocampal place area and the retrosplenial cortex. Here I examine the nature of the representations in these networks of scene processing regions, using fMRI repetition attenuation and multi-voxel pattern analysis. First I show how the brain represents orthogonal but complementary scene properties, such as spatial layout and object content. Second, I show the functional architecture of neural scene representation that supports the integrated scene percept that encompasses multiple views over time. Finally, I will present some of my recent experiments on the representation of the size of space in scene-selective areas of human brain. Altogether, my results show how multiple scene processing regions play distinct but complementary roles in constructing the rich and integrated scene representation that underlies our everyday experience of the visual world.

Ernst Niebur

02/28/2011 4:00pm
Jeremy Barry
SUNY Downstate

"Rapid loss of long-term, stable spatial firing patterns of place cells by inhibiting PKMζ”

PKMzeta, an atypical PKC isoform, is both necessary and sufficient for the maintenance of LTP. Injections of the PKMzeta pseudosubstrate inhibitor ZIP reverse established late-phase LTP at hippocampal synapses and also disrupt several forms of hippocampus-dependent spatial memory. Because hippocampal place cells are believed to play a crucial role in spatial memory, we asked if ZIP infusions would interfere with an existing place cell representation of a familiar environment. To examine if injecting a volume of 1.0 microliter disturbs single cell waveforms recorded from tetrodes, we used Tris-buffered saline and muscimol as test agents. Saline injections do not affect unit activity. In contrast, 4.4 mM muscimol leads to a gradual silencing of both pyramidal cells and interneurons. Units closer to the injection site are inactivated earlier than units further from the injection site. Simultaneously recorded, putative axon fibers remain active, convincingly demonstrating tetrode stability. Bilateral injections of 10 nmol ZIP in 1 microliter saline severely perturb the characteristic stability of place cell firing fields in a familiar environment; even 5 hours after the injection there is no sign of reversal to the pre-injection firing pattern. Many pyramidal cells that had discharged as place cells lose their spatial firing specificity. The firing fields of cells that continue to show location-specific firing are no longer stable across recording sessions. The results indicate that the persistent activity of PKMzeta is necessary for the stability of hippocampal place cell firing fields and may explain how inhibition of the kinase causes the loss of stored spatial information in the hippocampus.

Jim Knierim

03/12/2012 4:00pm
Yanxi Liu
Associate Professor
Penn State University

“Computational Symmetry”

Symmetry is an essential mathematical concept, as well as a ubiquitous observable phenomenon in nature, science and art. Either by evolution or by design, symmetry imparts an efficient coding that makes it universally appealing -- recognition of symmetry and regularity is the first step towards capturing the essential structure of a real world problem while minimizing computational redundancy. Automatic symmetry detection from real world (digital) data turns out to be a surprisingly challenging problem that has puzzled researchers in machine intelligence, computer vision, robotics, and computer graphics for the past four decades. Recognizing the fundamental relevance and potential power that computational symmetry affords, we explore a formal and computational characterization of real world symmetry using a group theoretical model. Such a formalization simultaneously facilitates: (1) a robust and comprehensive algorithmic treatment of the whole regularity spectrum; (2) an effective detection scheme for real world symmetries and symmetry groups; and (3) a set of well-defined bases for measuring and discriminating quantified regularities on diverse data sets. In this talk, I will summarize the theoretical background on crystallographic groups, and will illustrate recent results of applications of computational symmetry in: texture analysis/synthesis, tracking, and manipulation; perceptual grouping and 3D modeling of urban scenes from a single view; automatic geo-tagging; and image ‘de-fencing’. I will also discuss evaluation of computer algorithm performance on real world symmetries, and future work suggesting intertwined relations with human perception.

Ernst Niebur

03/14/2011 4:00pm
Zijiang He, Ph.D.
Psychological & Brain Sciences
University of Louisville

“Push-Pull Perceptual Learning Reduces Sensory Eye Dominance and Improves Stereopsis”

Intensive perceptual training can improve visual performance in adults. Although the improvements are likely to be attributed to modifications of the excitatory and inhibitory neural networks, it is unclear what their relative contributions are. We recently designed a novel push-pull training protocol to reduce sensory eye dominance (SED), a condition that is mainly caused by an unbalanced interocular inhibition in the visual cortex. During the training, an attention cue presented to the weak eye precedes the binocular competitive stimulation. The cue stimulates the weak eye (push) while causing interocular inhibition of the strong eye (pull). We found that this push-pull protocol reduces SED (shifts the balance toward the weak eye) and improves stereopsis more so than a push-only protocol, which solely stimulates the weak eye without inhibiting the strong eye. The stronger learning effect with the push-pull training than with the push-only training underscores the crucial involvement of a putative inhibitory mechanism in neural plasticity of the mature brain. Our psychophysical observations also suggest that the learning effect of SED reduction reflects the plasticity of the primary visual cortex, whose neurons have small receptive field sizes and receive modest top-down attention modulation.

Rudiger von der Heydt

03/16/2011 4:00pm
Kristina Nielsen, Ph.D.
Systems Neurobiology Laboratory
Salk Institute Biological Studies

“Shape Encoding in Monkey Extrastriate Cortex”

The ability to recognize objects is a crucial component of any interaction with our visual environment. In the non-human primate, shape recognition depends on the ventral pathway, which extends from the primary visual cortex V1 to the inferotemporal cortex (IT). Along this pathway, shapes are encoded with increasing complexity: whereas neurons in V1 are tuned to the orientation of lines, IT neurons respond to complex parts of objects or entire objects such as faces. One interesting aspect of the parts-based representation found in IT is how it incorporates the behavioral relevance of different shape parts. In my talk, I will present results from my previous work investigating the influence of behavioral relevance on the responses of single neurons and the LFP in IT. While past research has established the overall organization of the ventral pathway, many open questions remain about how exactly shapes are represented in it. Most importantly, we need to better understand the organization of the neural circuits that encode shapes in extrastriate cortex. Dissecting the contribution of individual circuits to shape encoding will only be possible by studying brain function with much higher detail than most current techniques allow. However, two novel techniques – two-photon microscopy and viral vector-based approaches – promise to make these experiments feasible. Both techniques will allow us to study brain function at an unprecedented level of detail: Two-photon microscopy, combined with calcium imaging, makes it possible to obtain functional maps with single cell resolution. Subsequent post-mortem histology can correlate function with cell types. Viral vector-based approaches, on the other hand, allow us to selectively control the activity of neurons of a certain cell type. While both techniques are now routinely used in rodent models, they are still fairly new to research in the non-human primate, the animal model of choice when investigating complex visual processes such as shape recognition. My recent experiments have focused on establishing both two-photon microscopy and viral vector-based approaches in the macaque monkey. During the talk, I will describe how we use two-photon microscopy in the anesthetized monkey to study the functional micro-organization of area V1. I will also talk about viral vector-based approaches in the awake, behaving monkey. Our experiments show that we can use viruses to introduce Channelrhodopsin-2 (ChR2) into monkey superior colliculus, and that stimulation of the ChR2-expressing neurons is sufficient to alter the monkey’s behavior in an eye movement task.

Dr. Steven Hsiao

03/21/2011 4:00pm
Ed Vessel
Assistant Research Scientist
New York University

“Dissecting Neuroaesthetics”

What happens in the brain when a person is moved by a piece of artwork? The neural basis of aesthetic experience and judgment has recently been investigated by a number of research groups, with rather mixed findings. Recent work in our lab has sought to clarify these findings by investigating an aesthetic experience as an integration of information from multiple sources, including stimulus-triggered preference and a variety of emotional responses. Importantly, out method seeks to separate the brain response to subjective aspects of aesthetic experience from aspects tied to particular stimulus characteristics. We identified three distinct subnetworks of response entailed in aesthetic reactions: a posterior network associated with stimulus-triggered perceptual & semantic processing, subcortical regions associated with reward and decision-making, and a frontal network (including parts of the default mode network) likely reflecting aesthetic emotional responses and personal relevance. We also find that both aesthetic judgments and emotional reactions to artwork are highly individual, and that aesthetic reactions can be triggered by non-positive emotional experiences. These results give us a fuller picture of the ways in which aesthetic processing engages and integrates distinct neuro-cognitive processes.

Ed Connor

03/28/2011 4:00pm
Ben W. Strowbridge, Ph.D.
Professor Neuroscience Dept.
Case Western Reserve University

“Representing information in neuronal cell assemblies: Persistent activity in the dentate gyrus mediated by semilunar granule cells”

While most CNS neurons fire only transiently in response to synaptic input, some neurons in the neocortex and hippocampus can fire persistently following brief stimuli, reflecting a potential cellular substrate for short-term memory. The origin memory-linked persistent activity in cortical neurons is not known. Using rat hippocampal slices, we found an in vitro model of persistent activity that can encode biological information in the spontaneous synaptic inputs of dentate gyrus neurons. Persistent activity in this system appears to reflect divergent synaptic connections between a small population of displaced granule cells that are intrinsically bistable and a larger population of downstream projection neurons in the dentate hilus.

Dr. Steven Hsiao

04/11/2011 4:00pm
Simon Kelly. Ph.D.
Prof. Biomedical Engineering
City College of New York

"Parietal selection signals guiding the acquisition of reliable information"

Adaptive decision-making is enabled by brain mechanisms that rank or select among locations and objects. Within the cortical oculomotor system, both attentional and motivational factors have been shown to drive stimulus-selective neuronal activity in single-unit neurophysiology studies. To date, the focus has been restricted to the neural representation of locations to which a targeted action potentially brings immediate reward. However, the bulk of selections made in real life comprise intermediate steps that are not immediately rewarding but are important for gathering information to guide later target selection for reward. In my talk I will present recent data from an ongoing study in which we examine how monkeys - and neurons in the lateral intraparietal area (LIP) - select among cues based on their reliability in informing a later reward-target choice. We show that LIP neurons encode cue validity in both a relative and absolute manner. Perhaps more surprisingly, these signals appear not to reflect merely an encoding of expected reward but rather may track changes in expected reward, akin to temporal difference signals characterized in reinforcement learning work. These and other effects in the data are discussed in terms of LIP’s role in representing attended locations on one hand, and neuroeconomic decision variables on the other.

Dr. Manuel Gomez-Ramirez

05/02/2011 4:00pm
Naoki Kogo, Ph.D.
Professor of Psychology
University of Leuven, Belgium

The side matters: Why line drawings only confuse us in understanding figure-ground perception”

My talk will have two parts that are related each other. First, I will talk about some fundamental issues involved in investigating underlying mechanisms of a phenomenon called “completion”: filling-in of missing parts in an image. Computational models have applied contour-completion algorithms based on the detection of collinear boundary elements. However, there are examples that may indicate a fundamental problem of this approach. Importantly, for a signal to be qualified as a contour element, it has to be a part of a representation of a surface and hence the side of the boundary where the surface exists has to be specified. The same applies to interpreting neural activities at the location of completed contours. I hypothesize that completion is accompanied by a development of border-ownership signals. It means that the neural activities at the location of illusory contours not only indicate the existence of the boundary but also indicate that a surface on one side of the borderline is closer than the other surface. As the figure-ground perception is holistic, Gestalt, in nature, the global configuration of an image should be reflected in the computation of border-ownership signals. In the second part of my talk, I will talk about the computational model we developed recently (DISC model, Kogo (2010) Psy. Rev., 117, 406-439). I will show how border-ownership signals are computed in the model by global interaction between local elements in the image and I will explain an on-going research for further development of the model.

Dr. Rudiger von der Heydt

05/09/2011 4:00pm
John Krakauer, Ph.D.
Associate Professor
Johns Hopkins School Medicine



Dr. Steven Hsiao

05/16/2011 4:00pm
Frderic Theunissen
University of California, Berkeley

“Statistics of natural sounds, invariance, perception and neural representations”

We have used a combination of neuro-ethological and classical auditory neurophysiological approaches to study how behaviorally relevant sounds are processed in the avian auditory system. One of the contributions of the neuro-ethological approach has been the discovery of highly specific auditory neurons that appear to be specialized for detecting very specific behaviorally relevant sounds. On the other hand, many auditory neurons recorded in the auditory system of non-specialists do not exhibit such specificity. At the same time, animals and humans hear and process a large space of sounds and are able to categorize these in much broader perceptual terms, describing them in terms of their pitch, timbre and rhythm. By systematic analyzing neural responses to song in the ascending avian auditory system and relating receptive fields to the statistics of natural sounds, we have shown that these two approaches can be unified: we found that the spectro-temporal receptive fields of auditory neurons tile a subset of the acoustical space that is particular important for natural sounds. In addition, we found that neurons could be classified into functional clusters. Neurons in different clusters were sensitive to different song features and, we will argue, are involved in mediating distinct perceptual attributes. We have also extended our statistical analysis of natural sounds to include the degrading effects on the signals of propagation and interference from other sound sources. We show that sound features that are robust to these degradations are also the features in that show a high degree of perceptual invariance. Neurons in the avian auditory cortex that are tuned for such invariant features show robust responses to songs embedded in noisy backgrounds. Our approach shows that we need to understand both the how and the why of the computations performed by the auditory system to explain auditory perception and the processing of acoustical communication signals.

Dr. Takashi Yoshioka

05/23/2011 4:00pm
Asif Ghazanfar, Ph.D.
Assistant Professor Psychology
Princeton University

“Speech emerges and evolves through coupled oscillations”

When we watch someone speak, how much work is our brain actually doing? How much of this work is facilitated by the structure of speech itself? Our work shows that not only are the visual and auditory components of speech tightly locked (obviating the need for the brain to actively bind such information), but the statistical regularities in both modalities are also optimized to interact with rhythms in the brain. In other words, it seems that the structure of speech exploits the structure on-going brain activity, with communication emerging as this interaction unfolds in time. Similar reciprocal coupling between signal structure and on-going brain rhythms is also seen in monkey vocal communication, and the differences between this process and human speech suggest the possibility that speech may have evolved without radical changes to key brain structures or the development of new ones.

Dr. Jeffrey Yau

06/06/2011 4:00pm
Arup Roy, Ph.D.
Sr Manager Systems Engineering
Second Sight Medical Products



Ernst Niebur

06/13/2011 4:00pm
Lizabeth Romanski
Dept of Neurobiology & Anatomy
University of Rochester



Dr. Xiaoqin Wang

06/20/2011 4:00pm
Alexander Maier, Ph.D.

National Institute Mental Health



Dr. Marshal Shuler

06/27/2011 4:00pm
Sheila Nirenberg, Ph.D.
Assoc Prof Physiology & Biophysics
Cornell University

“Testing hypotheses about coding and computation in the visual system and something new about retinal prosthetics”


Dr. Ed Connor

07/22/2011 2:00pm
Cheng-Ta Yang, Ph.D.
National Cheng Kung University

“Relative saliency affects the perceptual decision process when detecting multiple feature changes”

Survival requires successful change detection. Recent studies showed that people may often fail to detect a large change. This phenomenon is called “change blindness” (CB). Extensive studies have investigated the causes of CB and they suggest that failure in encoding, retention, retrieval, comparison, or decision can result in CB. However, studies on the comparison and decision processes underlying change detection are rare. This talk will focus on how the decision mechanism combines multiple comparison outputs when detecting multiple feature changes and discuss the effect of relative salience between comparison outputs on the decision process. The systems factorial technology is conducted to design the experiments, analyze data, and make inferences regarding the process characteristics of decision mechanism. Two alternative hypotheses are considered: Co-activation model and the relative saliency hypothesis. Three series of studies are conducted to tease apart the two alternatives. The results converged to support the relative saliency hypothesis. The decision processes are flexible. The adoption of change-detection strategies that can maximize the subjective expected utility can vary depending on the relative saliency.

Dr. Rudiger von der Heydt

08/29/2011 4:00pm
Bruno Olshausen, Ph.D.
Associate Professor Neuroscience
University of California, Berkeley

“Learning Intermediate-Level Representations of Form and Motion from Natural Movies”

A key attribute of visual perception is the ability to extract invariances from visual input. How this is accomplished by neural circuits in the visual cortex is currently a mystery. Here, I describe a model for how neurons in intermediate-level areas could extract invariances from visual input, and I show how the representation of these invariances may be adapted to the statistics of dynamic natural scenes. The model is composed of two stages of processing: an early feature representation layer, and a second layer in which invariances are explicitly represented. Invariances are learned as the result of factoring apart the temporally stable and dynamic components embedded in the early feature representation. The structure contained in these components is made explicit in the activities of second-layer units that capture invariances in both form and motion. When trained on natural movies, the first-layer produces a factorization, or separation, of image content into a temporally persistent part representing local edge structure and a dynamic part representing local motion structure, consistent with known response properties in early visual cortex (area V1). This factorization linearizes statistical dependencies among the first-layer units, making them learnable by the second layer. The second-layer units are split into two populations according to the factorization in the first-layer. The form-selective units receive their input from the temporally persistent part (local edge structure) and after training result in a diverse set of higher-order shape features consisting of extended contours, multi-scale edges, textures, and texture boundaries. The motion-selective units receive their input from the dynamic part (local motion structure) and after training result in a representation of image translation over different spatial scales and directions, in addition to more complex deformations. These representations provide a rich description of dynamic natural images and provide testable hypotheses regarding intermediate-level representation in visual cortex. [Joint work with Charles Cadieu.]

Dr. Ernst Niebur

09/12/2011 4:00pm
Bartlett Mel, Ph.D.
University of Southern California

"On the close relationship between natural computations and dendritic computations"

In this talk we attempt to gain insight into the functions of individual pyramidal neurons by studying the natural computations that they are likely to be performing within their respective cortical circuits. Examples will be given of the role that dendrites play in maximizing storage capacity in online (one-shot) learning, and the role of dendrite location-dependent contextual modulation effects in the extraction of object contours in natural images.

Dr. Ersnt Niebur

09/26/2011 4:00pm
Joshua Berke
University of Michigan

“Basal Ganglia Dynamics during Action Selection and Suppression”

The basal ganglia are critical structures for motivated decision-making and reinforcement-driven learning. Distinct pathways through basal ganglia circuits have dissociable roles enabling and suppressing behavior, and in recent years the "hyperdirect" pathway to subthalamic nucleus has been proposed to mediate rapid cancellation of actions-in-preparation. I will present results from our ongoing electrophysiological investigations of these circuits during a variety of behavioral tasks, including a "stop-signal" task widely used as a test of executive function.

Veit Stuphorn

10/17/2011 4:00pm
Sridevi Sarma, Ph.D.
Assistant Professor
Biomedical Engineering, JHU

“Quickest Detection of Drug-Resistant Seizures: An Optimal Control Approach”

Epilepsy affects 50 million people worldwide, and 30% remain drug-resistant. This has increased interest in both chronic and responsive neurostimulation, which is most effective when administered at or near the foci and immediately prior to or at the seizure (ictal) onset. Precise focus localization and automatic online seizure detection (AOSD) from intracranial EEG (iEEG) recordings are therefore critical for closed-loop intervention, but remain challenging problems. Automated localization schemes have primarily been developed using univariate and bivariate features but lack consistency and accuracy. Several AOSD algorithms has been proposed thus far and though they are highly sensitive (large number of true positives), these algorithms generally have low specificity (large number of false positives), which limits their clinical use. The lack of specificity presumably occurs because (i) they compute statistics from a few channels at a time which may not capture network dynamics of the brain, (ii) the channels used may be too far from the focus, and/or (iii) the detection thresholds are not explicitly optimized to maximize AOSD performance. In this talk, we propose a novel computational framework for seizure foci localization and AOSD that involves (i) constructing multivariate statistics from all electrodes to localize foci and distinguish between non-ictal and ictal states; (ii) modeling the evolution of these statistics in each state and the state-transition probabilities; and, (iii) developing an optimal model-based strategy to detect transitions to ictal states from sequential neural measurements. This strategy is formulated as the Bayesian “Quickest Detection” (QD) of the seizure onset, and is solved via control optimization tools, and explicitly minimizes both the distance between detected and unequivocal onset times and the probability of false positives. We demonstrate our approach with intracranial EEG recordings from four patients with drug-resistant epilepsy. We find that the first singular vector of the connectivity matrix during seizure has a characteristic direction indicative of the seizure foci, and the corresponding first singular value is a robust statistic that can be used for QD of seizure onsets. Our preliminary results indicate that QD achieves high sensitivity with high specificity (low number of false positives).

Dr. Ernst Neibur

10/24/2011 4:00pm
John K. Tsotsos, PhD

York University, Toronto

“A View of Vision as Dynamic Tuning of a General Purpose Processor”

The importance of generality in computer vision systems whose goal is to achieve near-human performance was emphasized early in the field's development. Yet this generality in practice has been elusive. In this presentation, I will detail a proposal made in this direction. Using formal methods from complexity theory, we have shown what an architecture for vision might be that has two main properties: It can solve a particular class of vision problems very quickly, and, it can be tuned dynamically to adapt its performance to the remaining subclasses of vision problems but at a cost of greater time to process. Further, we have shown that a major contributor in this dynamic tuning process is the set of mechanisms that have come to be known as attention. Attentional processing has a strong formal computational foundation and this will be briefly overviewed. The result is a set of attentional mechanisms organized into three classes: selection, suppression and restriction mechanisms. The Selective Tuning model will be described as an embodiment of these mechanisms with broad predictive power for biological vision and significant experimental support. The top-down and recurrent mechanisms include both goals as well as attention mechanisms not specific to tasks or goals. The combined application of elements from the set of attentional mechanisms provides for a means to tune the general-purpose, but limited in functionality, processing network to enable the full set of visual tasks to be solved.

Dr. Rudiger von der Heydt

11/11/2011 4:00pm
Timo van Kerkoerle
Doctoral Candidate
Neuroscience Institute Netherlands

“Reciprocal mechanisms of alpha and gamma oscillations in monkey primary visual cortex”

The visual cortex is able to generate oscillations which are shown to be involved in perceptual processing. However, little is known about the in vivo cortical mechanisms of these rhythms. We find that sustained alpha (5-15 Hz) and gamma (30-100 Hz) oscillations move in opposite directions through the different layers of monkey primary visual cortex, suggesting that gamma frequency band activity is involved in the feed-forward processing of information while alpha frequency band activity is involved in recurrent interactions. These and other findings indicate that alpha and gamma frequencies serve reciprocal roles in monkey visual cortex.

Dr. Rudiger von der Heydt

11/11/2011 2:00pm
Pieter Roelfsema, Ph.D.
Neuroscience Institute Netherlands

“Neuronal mechanisms for perceptual grouping”

A fundamental task of vision is to group the image elements that belong to one object and to segregate them from other objects and the background. I will discuss a conceptual framework that explains how perceptual grouping is implemented in the visual cortex. According to this framework, two mechanisms are responsible for perceptual grouping: base-grouping and incremental grouping. Base-groupings are coded by single neurons tuned to multiple features, like the combination of a color and an orientation. They are computed rapidly because they reflect the selectivity of feedforward connections that propagate information from lower to higher areas of the visual cortex. However, not all conceivable feature combinations are coded by dedicated neurons. Therefore, a second, flexible form of grouping is required that is called incremental grouping. Incremental grouping takes more time than base-grouping because it relies on horizontal connections between neurons in the same area and feedback connections that propagate information from higher to lower areas. These connections spread an enhanced response to all the neurons that code image elements that belong to the same perceptual object. This response enhancement acts as a label that tags those neurons that respond to image elements to be bound in perception. The enhancement of neuronal activity during incremental grouping has a correlate in psychology because attention is directed to precisely those features that are labeled by the enhanced neuronal response. I will show data indicating that feedforward and feedback processing rely on different receptors for glutamate and on processing in different cortical layers.

Dr. Rudiger von der Heydt

11/17/2011 12:00pm
Hidehiko Komatsu, Ph.D.
Institute Physiological Sciences, Japan

"Neural Representation of Surface Reflectance Properties"


Dr. Ed Connor

11/17/2011 4:00pm
Ueli Rutishauser, Ph.D.
Junior Group Leader
Max Planck Institute Brain Research

“The neuronal mechanisms of declarative memory formation in the human medial temporal lobe”

A central goal of neuroscience is to understand how neuronal circuits and mechanisms enable different behaviors. One task of particular significance is learning from novel experiences. In mammals, the medial temporal lobe is crucial for this rapid form of learning. The modification of synapses and neuronal circuits through plasticity is thought to underlie memory formation. The induction of synaptic plasticity is favored by coordinated action-potential timing across populations of neurons. Such coordinated activity of neural populations can give rise to oscillations of different frequencies, recorded in local field potentials. Brain oscillations in the theta frequency range (3-8Hz) are often associated with the favorable induction of synaptic plasticity as well as behavioral memory. I will present data on single neurons recorded together with the local field potential in humans engaged in learning tasks. Recordings are from the human hippocampus and amygdala of patients with drug resistant epilepsy that were semi-chronically implanted with hybrid depth electrodes. We found that successful memory formation in humans is predicted by a tight coordination of spike timing with the local theta oscillation. More stereotyped spiking predicts better memory, as indicated by higher retrieval confidence reported by subjects. Further, I will describe a class of neurons in the human brain that selectively respond only when a stimulus, such as a photograph or a face, is seen the very first time. These neurons signal stimulus novelty, a critical prerequisite for successful learning. Demonstrating this, the response of these neurons is predictive of whether subjects remember certain stimuli. Listening to a few such neurons allows a brain-machine interface to outperform the memory retrieval performance of subjects. This supports the notion that these neurons represent memory content as such as opposed to decisions. In combination these findings provide a link between memory-related behaviors and circuit mechanisms of plasticity.

Dr. Ernst Niebur

11/18/2011 2:00pm
Rufin Vogels, Ph.D.
Leuven University, Belgium

“Mechanisms of adaptation of spiking activity and local field potentials in macaque inferior temporal cortex”

It is well known that repetition of a stimulus reduces the responses of inferior temporal (IT) cortical neurons, an area coding for object properties in the macaque visual ventral stream. Several neural models have been proposed to explain this repetition suppression or adaptation effects. We compared predictions derived from these models with adaptation effects of spiking activity and Local Field Potentials (LFPs) in macaque IT cortex. First, we compared the effect of brief adaptation on shape tuning using parameterized shape sets with predictions derived from fatigue and sharpening models. We found, as expected, suppression of spiking activity and of LFP power in the high-gamma (60-100 Hz) band with repetition. However, repetition did not affect or even could enhance the power in lower frequency bands. Contrary to sharpening but in agreement with fatigue models, repetition did not affect shape selectivity. The degree of similarity between adapter and test shape was a stronger determinant of adaptation than was the response to the adapter. The spiking and LFP adaptation effects agreed with input-, but not response-fatigue models. Second, we examined whether stimulus repetition probability affects adaptation, as predicted from the top-down, perceptual expectation model of Summerfield et al. (Nat. Neurosci., 2008). Monkeys were exposed to 2 interleaved trials, each consisting of 2 either identical (rep trial) or different stimuli (alt trial). Repetition blocks consisted of 75% (25%) of rep (alt) trials and alternation blocks had the opposite repetition probabilities. For both spiking and LFP activities, adaptation did not differ between these blocks. This absence of any repetition probability effect on adaptation suggests that adaptation in IT is not caused by contextual factors related to perceptual expectation, but instead agrees with bottom- up, fatigue-like mechanisms. Third, we demonstrate that repetition in IT not only affects the degree of activation but also the synchronized activity (LFP-LFP and MultiUnit-LFP coherence). As for LFP power, these effects depend on the frequency band. We will discuss the implications of these single unit and LFP data for the interpretation of fMRI-adaptation studies.

Dr. Rudiger von der Heydt

11/18/2011 12;00pm
Douglas Nitz, Ph.D.
University of California, San Diego

“Spaces within spaces: the different forms by which parietal cortex, premotor cortex, and hippocampus map position in three spatial reference frames”

Considerable progress has been made in understanding how visual and tactile landmarks combine with self-motion information to yield brain activity mapping position in an environment. Less well understood is how such spatial information is used to guide decision-making and motor control. Parietal cortex exhibits activity patterns compatible with a transformation of spatial cognition into action and, anatomically, sits between space-mapping structures such as hippocampus and entorhinal cortex and action-mapping structures such as premotor cortex. In this talk, I will consider the form by which parietal cortex neuronal activity maps position and compare this with what is observed in hippocampus and premotor cortex. The potential role of parietal cortex in simultaneously mapping position within multiple, coincident spatial frames of reference will also be discussed.

Dr. James Knierim

01/25/2012 4:00pm
Dietmar Heinke, Ph.D.
University of Birmingham, UK

“Explaining visual search with the Selective Attention for Identification model (VS-SAIM): Competitive interactions between/within selection and object identification”

I will present an extension of the Selective Attention for Identification model (VS-SAIM) to modelling visual search (Heinke & Humphreys, 2003; Heinke & Backhaus, 2011). VS-SAIM is a model of translation-invariant object identification in multiple object scenes. First, an early visual processing stage of the model generates feature maps of vertical, horizontal and diagonal orientations. Next, a biologically plausible competitive selection process generates a translation-invariant representation of an object in a focus of attention (FOA), thereby dealing with the issue of multiple objects. The third stage implements the object identification through similarity-based matching between stored object templates and FOA. Crucially, this bottom-up pathway is complemented by parallel top-down feedback from the object identification to the selection stage. VS-SAIM mimics a range of classical findings from search displays containing lines with different orientations; lines with different lengths; and different letters (”T”,” ” and ”L”). The simulations also include asymmetric search patterns and results from primed visual search tasks. In VS-SAIM, these experimental findings are simulated through the competitive processes which are modulated through a combination of target-distractor similarity and featural properties of the distractors. Moreover the results also suggest that the following properties are crucial for explaining visual search: close interactions between selection stage and object identification; horizontal and vertical feature being more highly weighted than diagonal features (consistent with physiological evidence); limitation of top-down influence to avoid disruption of selection processes and a ”dynamic saliency map”.

Dr. Ernst Niebur

04/02/2012 4:00pm
Ken Nakayama, PhD
Professor of Psychology
Harvard University

“Subjective Contours”

Occluding contours constitute the boundaries of objects and surfaces. They have a special status in providing information about the geometrical relations of surfaces and objects in the real world. Even though such important contours are not obviously evident to observers in the clutter of surface markings of objects and terrain, they have a special status in visual processing. In this context, subjective contours are noteworthy because they provide specific evidence for the computation of occluding contours, visible evidence for an otherwise hidden process. Rather than being relegated to mere illusion (i.e. error) or having the minor role of contour completion, subjective contours reflect a fundamental process in the interpretation of 3 dimensional scenes. As such, they provide a litmus test for occluding contour representation. Thus, phenomenology, rather than just describing the appearance of things, can at favorable moments reveal deeper substrates of vision. Accordingly, perceptual demonstrations will be used support this interpretative framework.

Rudiger von der Heydt

04/16/2012 4:00pm
Jacqueline Gottlieb, Ph.D.
Associate Professor Neuroscience
Columbia University

“Attention as a value-based cognitive selection”

Attention is a core cognitive function that has been intensively investigated in mulitple empirical paradigms. In this talk I will focus on the points of intersection between two lines of research that are each well-developed in their own right but have remained largely separate – attentional theories of reinforcement learning (RL) in rats, and the study of visual and oculomotor attention in humans and non-human primates. I will describe briefly the differences between these two approaches and highlight their intellectual intersection. I then consider the application of RL ideas to the study of target selection responses in the monkey oculomotor system, specifically the lateral intraparietal area (LIP). I will argue that LIP neurons convey an estimate of the relative utility of alternative options. However, rather than encoding the value of an action based on a physical reward neurons are better described as encoding the utility of an internal mental action - devoting cognitive resources to a stimulus or source of information. I discuss the specific requirements of an information selection mechanism. I then describe two experiments that highlight these specific requirements: a sequential decision making task where neurons show sensitivity to non-redundant information in time, and a stimulus-reward (Pavlovian) paradigm where neurons assign attentional weight to stimuli based on their Pavlovian, non-operant, associations. I discuss the implications of these findings for attentional selection and the possibility of multiple attentional mechanisms.

Dr. Ernst Niebur

05/07/2012 4:00pm
Jose-Manuel Alonso, Ph.D.
State University of New York

“Populations of ON and OFF thalamic inputs underlying the functional architecture of primary visual cortex”

At its very first synapse, the visual pathway splits into two processing channels to encode light increments (ON) and decrements (OFF) in the visual scene. While textbooks have traditionally treated ON and OFF channels as mirror copies of each other, our work demonstrates that they are not equal partners in cortical space. We demonstrate that ON and OFF thalamic afferents show pronounced asymmetries in receptive field coverage, which have a major impact in cortical functional architecture, including the mapping of visual space, stimulus orientation, spatial phase and contrast polarity. Moreover, across the cortical map, we show that the OFF channel is faster and better represented than the ON channel, which results in a pronounced OFF-dominance in visual cortex. We demonstrate that the ON and OFF thalamic inputs to the same cortical cylinder have overlapping receptive fields with a position scatter limited to < 1.5 thalamic receptive field centers, a value that matches the average receptive field size of neurons at the input layers of the cortex. At the same time, we show that the ON-OFF asymmetries in thalamic receptive field coverage generate a cortical map for ON-OFF spatial phase that is closely associated to the orientation map. Our findings suggest that ON and OFF channels play a major role in the development of visual cortical architecture and raise new questions about the role of cortical OFF dominance in visual perception.

Dr. Hey-Kyoung Lee

05/14/2012 4:00pm
Marcos Frank, Ph.D.
Associate Professor Neuroscience
University of Pennsylvania

“Experience and sleep: Partners in synaptic plasticity”

The function of sleep has long eluded scientists, but converging findings suggest that sleep may play an essential role in brain plasticity. We have been investigating this possibility by examining the roles of experience and sleep in shaping cortical circuits during early life. We find that sleep consolidates a classic form of experience-dependent cortical plasticity that occurs during a critical period of development. We also find that the underlying mechanisms include NMDAR activation and cortical protein synthesis: two processes which increase cortical synaptic strength. These findings suggest that during sleep labile plastic changes initiated during wakefulness are transformed into more permanent forms.

Dr. Alfredo Kirkwood

05/21/2012 4:00pm
Javier Medina
Assistant Professor of Psychology
University of Pennsylvania

“The Ins and Outs of Purkinje Cells: Error Signals and Motor Commands”

Neural circuits coursing through the cerebellum are at the center of current theories about how the brain learns to adapt and calibrate our movements. In particular, Purkinje cells in the cerebellar cortex are thought to receive a signal when an error is made, and use this information to adapt subsequent movements in a way that improves performance. I will present evidence from a series of recent imaging and optogenetics experiments in mice that support this alleged role of Purkinje cells in motor learning. Specifically, I will show that (1) Purkinje cells receive input signals that provide information about the size of the error in individual trials, and (2) Purkinje cells send output signals that can control movement directly.

Jim Knierim

06/04/2012 4:00pm
Peter Strick, Ph.D.
University of Pittsburgh

“Why don’t rodents play the violin? Evolution in the neural substrate of motor skills”

1) The central control of movement is faced with an evolutionary constraint: Our skeletomotor system is built on the framework of a relatively ancient spinal cord. 2) Most descending systems, including the corticospinal system, use the pattern generators and motor primitives that are built into the spinal cord to generate motor output. 3) Cortico-motoneuronal (CM) cells (i.e., cortical neurons with axons that make monosynaptic connections with motoneurons) are a relatively new phylogenetic and ontogenetic development. Furthermore, CM cells are located in a separate part of the primary motor cortex (area 4). 4) Area 4 is split into 2 regions: a rostral region we have termed “Old M1” which has disynaptic input to motoneurons; and a caudal region we have termed “New M1” which has monosynaptic input to motoneurons. In essence, Old M1 makes use of the circuits built into the spinal cord to generate motor output. This region of the motor cortex enables the motor system to avoid the “curse of dimensionality” and to solve the “degrees of freedom problem.” In contrast, New M1 uses CM cells to bypass the constraints of spinal cord mechanisms. This region of the motor cortex enables the motor system to use all of the available degrees of freedom to sculpt novel patterns of motor output. These arguments lead us to predict that the two regions of the motor cortex are differentially involved in motor learning. We speculate that Old M1 is especially important during the initial stages of learning a new skill by enabling the motor cortex to use existing spinal circuits to rapidly construct new movement patterns. In contrast, New M1 may be especially important during the later stages of learning a new skill by enabling the motor cortex to refine and precisely specify patterns of motor output.

Dr. Xiaoqin Wang

09/17/2012 4:00pm
Aaditya Rangan, Ph.D.
Assistant Prof Applied Mathematics
New York University

“Emergent dynamics in a model of the visual cortex”

A simple network model of the primary visual cortex (V1) can exhibit a variety of phenomena observed in the real visual cortex, including orientation tuning, spontaneous background patterns and surround suppression, as well as increased gamma-band oscillations and decreased trial-to-trial variance following stimulus onset. The dynamic regimes of this model which capture all of these phenomena are rich in structure and strongly shaped by temporally localized barrages of excitatory and inhibitory firing. This emergent collaborative activity has far-reaching consequences, giving rise to the phenomena mentioned above and informing many testable predictions within the real V1.

Dr. Ernst Niebur

10/08/2012 4:00pm
Natalia Dounskaia, Ph.D.
Associate Professor
Arizona State University

"Directional Preferences: A window to optimization of arm movements"

I will present results obtained in my lab from a series of recent studies of directional preferences during arm movements. This research was prompted by a prediction of the leading joint hypothesis that during horizontal arm movements, some directions are preferred and some other directions are non-preferred. We have been using a free-stroke drawing task to demonstrate directional preferences. We have also performed a number of studies to investigate factors that contribute to the directional preferences. Two methodological approaches were used with this purpose. First, we computed cost functions that represented various possible contributors to the directional preferences and compared the directions in which each cost function was optimized with the revealed preferred directions. Second, we applied experimental manipulations that could influence directional preferences. Results of these studies consistently point to a tendency to minimize neural effort for control of inter-segmental dynamics of the arm as the most influential factor underlying the directional preferences. I will also present our latest results obtained with this approach that provide implications for control of 3D arm movements with redundant degrees of freedom and for the effect of Parkinson’s disease on movement production.

Dr. Amy Bastian

11/05/2012 4:00pm
Daniel O'Connor, Ph.D.
Dept. Neuroscience
Johns Hopkins School of Medicine

THE DAVID BODIAN SEMINAR in NEUROSCIENCE THE JOHNS HOPKINS UNIVERSITY, ZANVYL KRIEGER MIND/BRAIN INSTITUTE 338 Krieger Hall 3400 N. Charles Street Baltimore, MD 21218-2685 Daniel H. O’Connor, Ph.D. Assistant Professor Department of Neuroscience School of Medicine Johns Hopkins University Host: Dr. Rudiger von der Heydt Monday, November 5, 2012 10:00 a.m. 10:30 a.m. 11:00 a.m. 11:30 a.m. 12-1:30 p.m.--LUNCH 1:30 p.m. 2:00 p.m. 2:30 p.m. 3:00 p.m. 3:30 p.m. 4:00 p.m. Seminar “Neural coding for active object localization revealed using synthetic touch”

Touch perception requires integration of somatosensation and movement, mediated in part by cortical sensorimotor circuits. However, how perception arises through activity in defined nodes within elaborate circuits is poorly understood. I will discuss experiments in which we causally linked spike trains in cortical layer (L) 4 neurons in mouse primary somatosensory cortex with whisker-based perception. In mice localizing an object using the C2 whisker, L4 spikes represented object location with both total spike count and precise spike latency. We optogenetically controlled L4 neurons in single barrels in closed-loop with whisker movement, with millisecond precision. Mimicking touch-related activity in space and time caused illusory touch perception. Scrambling the timing of spikes by tens of milliseconds did not abolish the illusion. However, illusions could only be induced by photostimulating the C2 barrel, during epochs when mice expected touch. Our results suggest that mice locate objects using spike count in somatotopically specific ensembles, gated by bouts of active exploration.

Rudiger von der Heydt

11/12/2012 4:00pm
Dario Ringach, Ph.D.
Professor of Neurobiology

“Retinal Origin of Orientation Tuning and Maps in Primary Visual Cortex”

The orientation map is a hallmark of primary visual cortex in higher mammals. It is not yet known how orientation maps develop, what function they have in visual processing and why some species lack them. I will discuss the notion that the organization of the periphery, namely the spatial statistics of retinal ganglion cell mosaics, could seed the emergence of orientation tuning and the organization of the orientation map early in development. I will present some supporting data and discuss some ongoing efforts to directly test the idea.

Dr. Kristina Nielsen

12/17/2012 4:00pm
Shiyong Huang, Ph.D.
Assistant Research Scientist
Krieger Mind/Brain Insitute, JHU

“Regulation of inhibitory circuits during critical period of ocular dominance plasticity”

Since Hubel and Wiesel first described a critical period of postnatal development for ocular dominance plasticity, the primary visual cortex has been a model system to study plasticity of neuronal circuits and the impact that altering normal development has on neural function. Critical periods for experience-dependent plasticity have been reported in many areas and are thought to be crucial for the normal CNS development. A fundamental question that remains in the field is what mechanisms control the initiation and termination of the critical period. An attractive candidate is the maturation of GABAergic inhibitory circuits. The current general hypothesis is that a developmental increase in the strength of inhibition constrains the plasticity of excitatory synapses. My research has focused on the development and plasticity of the input and output connectivity of the so called fast-spiking interneurons, the most abundant cortical GABAergic cell, which provide the bulk of somatic inhibition. My work revealed that endocannabinoids control the maturation of GABAergic efficacy and connectivity during the critical period. Importantly, an examination of the spike timing dependent plasticity of excitatory synapses onto interneurons, along with sensory deprivation studies, suggest that changes in the plasticity of the excitatory synaptic driving onto fast spiking cells might control the end of the critical period. GABAergic circuit abnormalities have been implicated in the pathogenesis of a variety of neurodevelopmental and cognitive disorders, such as schizophrenia, autism, and epilepsy. Understanding of the synaptic mechanisms underlying the development and refinement of inhibitory circuits will potentially uncover altered cellular targets in neurodevelopmental disorders.

Dr. Ernst Niebur

01/07/2013 4:00pm
Melanie Wilke, Ph.D.
Director, Cognitive Neurology
University of Gottingen, Germany

“Neural Basis of Visual Awareness and its Disorders”

While sensory responses in cortex and thalamus have been studied extensively, little is known about the role of thalamo-cortical interplay in visual awareness and decision making. I will describe a series of experiments that investigate how activity in the thalamus and cortex contribute to these cognitive operations. These experiments capitalize on a combined approach of behavioral studies, electrophysiology, pharmacology and fMRI in monkeys, and provide insight into the interaction between the cortical and thalamic circuitry under normal and pathological conditions. In a first set of experiments we investigated neural correlates of visual awareness by applying a paradigm that renders salient stimuli perceptually invisible. The neurophysiological recordings revealed widespread perceptual modulation of local field potential (LFP) activity in cortex and thalamus (V1, V2, V4, LGN and pulvinar), combined with spiking rate changes in V4 and pulvinar. At the earliest processing stages, the presence of modulation in the LFP may represent recurrent perceptual signals from extrastriate cortex or higher-order thalamic nuclei such as the pulvinar. We next tested whether the pulvinar is indeed critical for visual awareness by reversibly inactivating regions within it by means of GABA-A agonists. While pulvinar inactivation did not impair first-order visual processing, it led to a spectrum of visuomotor and visual awareness related deficits commonly seen in human patients with extensive cortical lesions, i.e. spatial neglect. Deficits included limb and eye movement bias towards the side of the lesion and reduced responsiveness to objects presented in the contralesional space. Interestingly, the inactivation-induced spatial selection bias could be alleviated by means of stimulus-associated reward manipulations, but less by visual saliency manipulations, suggesting that dPULV is less critical for perceptual processing but is involved in subsequent action selection. To investigate whether pulvinar inactivation disrupts cortical activity, we next measured whole-brain BOLD activity while monkeys performed a spatial saccade decision task in a 4.7T MRI scanner. The spatial bias was accompanied by BOLD-activity reductions in parieto-frontal cortex for contralesional cues. This activity decrease was more pronounced in the lesioned hemisphere but was also present in the intact hemisphere. During the talk we will speculate that the dorsal pulvinar contributes to communication in fronto-parietal networks and that its inactivation causes a mismatch between visual stimulus location, and eye and hand movement parameters.

Dr. Kristina Nielsen

01/28/2013 4:00pm
Manuel Gomez-Ramirz, Ph.D.
Postdoctoral Fellow
Johns Hopkins University

Advanced Research Lecture Series "Neural Mechanisms Mediating Feature-Based Attention: The Role of Spike-Synchrony and Interneuronal Noise Correlations"

Studies show that spike-synchrony is enhanced across cells encoding stimuli within the attended sensory channel. We queried the specificity of this effect by testing whether attention enhances spike-synchrony based on the sensory feature preferred by the neural population. Moreover, we assayed whether similar effects are instantiated in interneuronal noise correlations. Single-unit recordings were made from secondary somatosensory cortex while animals engaged in tactile-feature (orientation and frequency) and visual (luminance) discrimination tasks. We observed that both firing-rate and spike-synchrony were enhanced when attention matched the preferred feature of neurons. However, the incidence and magnitude of these effects were significantly greater for spike-synchrony. Importantly, we found that spike-synchrony was a reliable predictor of behavior. In contrast to the spike-synchrony attention effects, the data revealed that attention to either tactile feature resulted in reduced interneuronal correlations, suggesting that this mechanism is not feature-specific.Taken together, the data suggest that spike-synchrony plays a prominent role in feature selection, and that feature-attention operates by reducing the overall noise levels in the population and synchronizing activity across the neural cohorts encoding relevant stimulus features.

Dr. Ernst Niebur

02/04/2013 4:00pm
Pablo Celnik, M.D.
Physical Medicine & Rehabilitation
Johns Hopkins Hospital

“Neurophysiological mechanisms underlying human motor learning retention and interference”

Plasticity of synaptic connections in the primary motor cortex (M1) is thought to play an essential role in motor learning and memory. Human and animal studies have shown that motor learning results in LTP-like plasticity processes such potentiation of M1 and a temporary occlusion of further LTP-like plasticity. It has been speculated that occlusion of LTP-like plasticity following learning, indicative of the magnitude of LTP-like changes resulting from motor training, is essential for retention. However, direct evidence supporting this is lacking. I will present a series of human studies where we assessed occlusion of LTP-like plasticity after motor training. Here, we found that (1) the magnitude of occlusion is proportional to measures of motor skill retention, (2) reversal of occlusion, indicative of interference with LTP-like processes, resulted in decrease skill retention, (3) the magnitude of occlusion is proportional to anterograde interference and to the resilience to retrograde interference. Overall these studies indicate that occlusion of LTP-like plasticity is crucial for retention and point to one of the mechanisms underlying learning interference.

Dr. Steven Hsiao

02/11/2013 4:00pm
Anthony Movshon
Professor of Neural Science
New York University

“Cortical and perceptual processing of naturalistic visual structure”

The perception of complex visual patterns emerges from neuronal activity in a cascade of areas in the primate cerebral cortex. Neurons in the primary visual cortex (V1) represent information about local orientation and spatial scale, but the role of the second visual area (V2) is enigmatic. We made synthetic images that contain complex features found in naturally occurring visual textures, and used them to stimulate macaque V1 and V2 neurons. Most V2 cells respond more vigorously to these stimuli than to matched control stimuli lacking naturalistic structure, while V1 cells do not. fMRI measurements in humans reveal differences in V1 and V2 responses to the same textures that are consistent with neuronal measurements in macaque. Finally, the ability of human observers to detect naturalistic structure is well predicted by the strength of the neuronal and fMRI responses in V2 but not in V1. These results reveal a novel and particular role for V2 in the representation of naturally occurring structure in visual images, and suggest ways that it begins the transformation of elementary visual features into the specific signals about scenes and objects that are found in areas further downstream in the visual pathway.

Dr. Gomez-Ramirez

02/18/2013 4:00pm
Frank Bremmer, Ph.D.
Philips-Universitat Marburg

“Spatial Encoding During Eye Movements”

Primates move their eyes more often than their heart beats. Different from our introspection, visual perception during eye movements is not veridical. Instead, eye movements induce multi-facetted perceptual errors. It is generally agreed, that understanding the neural bases of misperceptions allows for a deeper understanding of the neural mechanisms underlying veridical perception during everyday life. Over the last years we have studied in detail the influence of eye movements on spatial perception and spatial encoding. We performed psychophysical studies in humans together with neurophysiological recordings in non-human primates. Our results unequivocally show that the functional properties of neurons in macaque extrastriate and parietal cortex are suited to explain previously described visual perceptual effects during eye movements.

Dr. Rudiger von der Heydt

03/04/2013 4:00pm
Akiko Ikkai
Postdoctoral Fellow, JHU
Psychological & Brain Sciences

“Guiding goal-directed behaviors: the locus and effects of control signals”

Attention and working memory (WM) are cognitive faculties to selectively take in relevant information and maintain it for a period time. These abilities allow us to act flexibly and appropriately in a dynamic environment where physically salient information and internal goals coexist. In this talk, I will describe experiments of spatial attention and WM that aim to understand the representation of task-relevant information that guides goal-directed behaviors. First, I will present fMRI studies that investigated roles of the prefrontal and the parietal cortices in signaling locations in space that contains task-relevant information. These “priority maps” may influence down-stream regions to produce covert and overt behaviors. The second half of the talk will focus on the effects of these spatially selective top-down signals on the sensory cortex in terms of neural oscillation. I will present MEG and EEG data demonstrating that the modulation of posterior alpha (8-13Hz) oscillatory power reflects the underlying neuronal groups’ excitatory or inhibitory state. Particularly, we used a novel spatial WM paradigm to show that the sensory cortex was inhibited during the maintenance of abstract information. Our results indicate that alpha oscillation acts as a shaping mechanism that inhibits task-irrelevant regions of the brain in order to efficiently distribute resources to the regions that are highly involved in the ongoing task. These studies illustrate an approach to understand the neural mechanisms underlying our complex cognition, which is tightly linked to behavioral outcomes.

Dr. Ernst Niebur

03/11/2013 4:00pm
Bijan Pesaran, Ph.D.
Associate Professor
New York University

"Neuronal dynamics in the posterior parietal cortex during coordination and decision"

This talk will report our findings that temporally-patterned, coherent spiking activity in the posterior parietal cortex (PPC) coordinates the timing of eye movements with arm movements and reflects decisions to look and reach. Using a spike-field approach, we identify a population of parietal neurons that fire spikes coherently with 15 Hz beta frequency LFP activity. The firing rate of coherently-active neurons predicts the reaction times (RTs) of coordinated reach-saccade movements but not of saccades when made alone. Furthermore, in area LIP, neurons that do not fire coherently with LFP activity do not predict RT of either movement type. These data indicate that coherent spiking activity in PPC controls reaches and saccades so that they are made together. We then examine decision representations in the PPC, when making choices with two different movements, a saccade and a reach. To accurately infer the valuations of each option, we develop a reinforcement learning model of choice behavior. We find that values in the model are updated more strongly in response to rewards earned for saccades than reaches. Examining the timing of decision signals in both areas, we find that coherently active neurons signal the movement choice earlier than neurons that do not fire coherently. Thus, valuations are acquired and processed differently depending on the effector, eye or arm, used to make the choice, and the neurons that signal movement choices earliest exhibit coherent activity between area LIP and PRR. Taken together, these findings indicate that coherent patterns of neural activity are important to understanding the neural computations that are performed by PPC neurons to support coordination and decision making.

Dr. Manuel Gomez-Ramirez

03/18/2013 4:00pm
Sachin Deshmukh, Ph.D.
Assistant Research Scientist
Krieger Mind/Brain Insitute, JHU

“Objects, space, and memory: how the hippocampal cognitive map comes together”

The hippocampus is thought to function as a ‘cognitive map’, which stores nonspatial information such as items and events in a spatial framework. In order to understand the computations involved in creating such conjunctive nonspatial + spatial representation, it is essential to understand the function of hippocampal inputs. Medial entorhinal cortex (MEC) is known to convey spatial information to the hippocampus. My experiments show that lateral entorhinal cortex (LEC) conveys external sensory input—both spatial and nonspatial—to the hippocampus, in contrast to the self-motion based, internal processing of the MEC. Furthermore, the landmark-derived spatial information arises de-novo in LEC, since its major cortical input, the perirhinal cortex (PRC) encodes nonspatial but not spatial information in the presence of objects. Last, the hippocampal neurons use object derived spatial information to encode vectors to individual landmarks in the same behavioral paradigm. Overall, these results show the information transformation along the PRC-LEC-hippocampus pathway.

Dr. Ernst Niebur

03/25/2013 4:00pm
Murray Sherman, Ph.D.
Professor of Neurobiology
University of Chicago

“The Thalamus and its Role in Cortical Functioning”

Glutamatergic inputs in thalamus and cortex can be classified into two categories: Class 1( driver) and Class 2 (modulator). Following the logic that identifying driver pathways in thalamus and cortex permit insights into information processing leads to the conclusion that there are two types of thalamic relay: first order nuclei like the LGN relay driver input from a subcortical source (i.e., retina), whereas higher order nuclei like the pulvinar relay driver input from layer 5 of one cortical area to another. This thalamic division is also seen in other sensory systems: for the somatosensory system, first order is VPM/L and higher order is POm; and for the auditory system, first order is MGBv and higher order is MGBd. Furthermore, this first and higher order classification extends beyond sensory systems. Indeed, it appears that most of thalamus by volume consists of higher order relays. Many, and perhaps all, direct driver connections between cortical areas are paralleled by an indirect cortico-thalamo-cortical (transthalamic) driver route involving higher order thalamic relays. Such thalamic relays represent a heretofore unappreciated role in cortical functioning, and this assessment challenges and extends conventional views both regarding the role of thalamus and mechanisms of corticocortical communication. Evidence for this transthalamic circuit as well as speculations as to why these two parallel routes exist will be offered.

Dr. Hey-Kyoung Lee

04/01/2013 4:00pm
Sridhar Raghavachari, Ph.D.
Assistant Professor Neurobiology
Duke University Medical Center

“Modeling calcium signaling in single dendritic spines”

The activation of the calcium/calmodulin-dependent protein kinase II (CaMKII) is critical for the induction of long-term potentiation (LTP), a likely cellular correlate of learning and memory. High-resoultion imaging studies have revealed that CaMKII activation during the induction of synaptic plasticity in single dendritic spines is rapid and transient and spatially restricted. However, the determinants of this circumscribed response are not clear. We developed a computational model of the signaling cascade activated in dendritic spines upon Ca2+ influx. We used a diverse and extensive set of experimental measurements to constrain model parameters that determine the Ca2+ influx, the activation of the Ca2+ sensor, calmodulin (CaM), the binding to and modulation of CaM’s affinity for Ca2+ by CaM-binding proteins and structural rearrangements within CaMKII that govern its sensitivity to CaM. Simulations of this model closely matched the experimentally observed spatial and temporal profile measured by fluorescence lifetime imaging. We show that a high concentration of CaM in the spine, maintained by the postsynaptic protein, neurogranin, is required for efficient transduction of the Ca2+ signal into CaMKII activity. We find that cooperative binding of CaM to CaMKII, which arises from allosteric interactions within the holoenzyme, is critical for the determining both the speed and magnitude of CaMKII activation. Finally, we show that the activation of protein phosphatase 1 (PP1) and mutual negative feedback between CaMKII and protein phosphatase 2A is required to explain the decay in CaMKII activation.

Dr. Ernst Niebur

04/08/2013 4:00pm
Rhonda Dzakpasu, Ph.D.
Assistant Prof. Physics Depart
Georgetown University

“How Synaptic Potentiation Balances Plasticity and Stability within an In Vitro Network of Neurons”

Long-term potentiation (LTP) is widely believed to be the physiological basis of learning and memory. Mechanisms underlying LTP have been studied extensively at the monosynaptic level, but the effects of LTP on larger scale networks of neurons remain poorly understood. We chemically induce LTP in a cultured network of hippocampal neurons and show that after synaptic potentiation, a network of in vitro hippocampal neurons returns to a homeostatic state after widespread increases in firing.

Dr. Ernst Niebur

04/15/2013 4:00pm
Carson C. Chow, Ph.D.
Laboratory of Biological Modeling

“The Dynamics of Cortical Competition”

When the visual system is presented with multiple or ambiguous stimuli, several responses can occur. In some cases, the activity of the neurons will sum sublinearly or normalize in response to multiple stimuli within their receptive field. For ambiguous stimuli as in binocular rivalry, neural activity may oscillate and the oscillations are correlated with the perception. The period of the oscillations can also change depending on the context with an infinite period corresponding to complete disambiguation. Here, I will show that a simple canonical cortical circuit consisting of recurrent excitation, lateral inhibition and fatigue mechanisms with parameters in the physiological regime is sufficient to account for all of these dynamics.

Dr. Ernst Niebur

04/22/2013 4:00pm
Peter Rapp, Ph.D.
Senior Investigator
National Institute on Aging

“The 87%: Mindspan over Lifespan in Research on the Aging Brain”

Deficits in memory and other domains of cognitive function are among the most troubling signs of aging. Alongside the devastating impairments of Alzheimer’s disease and related neurodegenerative disorders, a much larger segment of the population experiences milder decline in cognitive health that is nonetheless sufficient compromise the quality of life and capacity for independent living. An earlier view was that neuron death is an inevitable consequence of growing older and the proximal cause of age-related cognitive impairment. Pronounced and distributed neuronal degeneration, however, is now understood to be a signature of pathological aging, and considerable evidence suggests instead that the brain changes associated with normal aging are regionally selective, involving relatively subtle alterations in connectivity and a blunted capacity for dynamic modification. The magnitude of these neurocognitive changes varies substantially across individuals, and growing interest has focused on strategies to promote optimally healthy outcomes. Studies in preclinical animal models have provided a valuable window on these issues.

Dr. James Knierim

05/06/2013 4:00pm
Li Jingling, Ph.D.
Assistant Professor
China Medical University, Taiwan



Dr. Rudiger von der Heydt

08/06/2013 4:00pm
Sliman Bensmaia, Ph.D.
Assistant Professor
University of Chicago

“Spatial and temporal codes mediate the tactile perception of natural textures”

When we run our fingers over the surface of an object, we acquire information about its microgeometry and material properties. Texture information is widely believed to be conveyed in spatial patterns of activation evoked across one of three populations of cutaneous mechanoreceptive afferents that innervate the fingertips. Here, we record the responses evoked in individual cutaneous afferents in Rhesus macaques as we scan a diverse set of natural textures across their fingertips using a custom-made rotating drum stimulator. We show that a spatial mechanism can only account for the processing and perception of coarse textures. Information about most natural textures, however, is conveyed through precise temporal spiking patterns in afferent responses, driven by high-frequency skin vibrations elicited during scanning. Furthermore, these texture-specific spiking patterns predictably dilate or contract in time with changes in scanning speed; the systematic effect of speed on neuronal activity suggests that it can be reversed to achieve perceptual constancy across speeds. The proposed coding mechanism involves converting the fine spatial structure of the surface into a temporal spiking pattern, shaped in part by the mechanical properties of the skin, and ascribes a novel function to vibration-sensitive mechanoreceptive afferents. This temporal mechanism complements the spatial one and greatly extends the range of tangible textures.

Dr. Ed Connor

08/26/2013 4:00pm
Michael Yassa, Ph.D.
Assistant Professor, JHU
Psychological & Brain Sciences

“Pattern separation in the human hippocampus and medial temporal cortex”

Virtually every computational model of the hippocampus includes a description of two key processes: (1) pattern separation, i.e., the process of disambiguating similar representations to minimize interference, and (2) pattern completion, i.e., the process by which incomplete representations are filled in based on pre-existing knowledge. Pattern separation and pattern completion allow us to flexibly encode new events as distinct from old and to generalize across contexts. Convergent work from rodent studies and human neuroimaging studies has observed separation signals in the dentate gyrus and CA3 (DG/CA3) when environments or stimuli are made similar. I will present data on how these signals can be isolated in fMRI in humans, how medial temporal cortices (lateral and medial entorhinal cortex, perirhinal cortex, and parahippocampal cortex) contribute to such processes in a domain-specific way, how separation abilities change in the context of aging and in mild cognitive impairment, and furthermore how behavioral pattern separation ability is linked to (1) novelty signals in the DG/CA3 region, (2) the integrity of the hippocampal perforant path, and (3) dendritic integrity in the DG/CA3 region, all of which change with age.

Dr. James Knierim

09/16/2013 4:00pm
Asohan Amarasingham, Ph.D.
Assistant Professor
City College of New York

“Trial-to-trial variability, nonstationarity, and the interpretation of firing rate in neurophysiology”

Many experimental studies of neural coding rely on a statistical interpretation of the theoretical notion of a neuron’s firing rate. For example, neuroscientists often ask: “Do a population of neurons exhibit more synchronous firing than one would expect from the co-variability of their instantaneous firing rates?” For another example, “How much of a neuron’s observed spiking variability is due to the variability of its instantaneous firing rate, and how much to spike timing variability?” But a neuron’s theoretical firing rate is not necessarily well-defined. Consequently, neuroscientific questions involving the theoretical firing rate require additional statistical modeling choices; ignoring this ambiguity can lead to inconsistent reasoning. This observation is related to common concerns about the appropriateness of (even mild) stationarity assumptions in neurophysiological data analysis. I will illustrate these issues with examples drawn from the neural coding literature, focusing on ‘doubly stochastic’ spike train models, and describe some tools and applications that are designed with these concerns in mind.

Dr. Ernst Niebur

10/07/2013 4:00pm
William Guido, Ph.D.
Professor and Chair
University of Louisville

“Circuit dissection in the mouse visual thalamus”

The mouse visual system has proven to be a powerful experimental platform to delineate circuits underlying visual information processing. Here we will show how the power of mouse transgenics, in vitrio slice physiology and optogenetics can be used to reveal the functional and structural state of both retinal and non retinal circuits in mouse visual thalamus.

Dr. Hey-Kyoung Lee

10/28/2013 4:00pm
Farran Briggs, Ph.D.
Dartmouth University

“Attention and Early Visual Circuits”

Attention shapes how we perceive the world. In the visual system, focusing attention on a visual stimulus increases the salience of that particular stimulus. While many studies have examined how attention alters vision, the mechanisms by which this occurs remain elusive. We set out to determine whether attention changes communication among neurons in the thalamocortical circuit. In alert monkeys performing a contrast-change detection task that required covert shifts in visual spatial attention, we probed thalamocortical communication by electrically stimulating neurons in the lateral geniculate nucleus while simultaneously recording shock-evoked responses from monosynaptically connected neurons in primary visual cortex. We found that attention significantly enhances neuronal communication across this circuit by 1) increasing the efficacy of presynaptic input in driving postsynaptic responses, 2) increasing synchronous responses among ensembles of postsynaptic neurons receiving independent input, and 3) decreasing redundant signals between postsynaptic neurons receiving common input. These results demonstrate that attention finely tunes neuronal communication at the synaptic level by selectively altering synaptic weights, enabling enhanced detection of salient events within the noisy sensory environment.

Dr. Kristina Nielsen

11/06/2013 4:00pm
Marius Usher, Ph.D.
Professor Cognitive Neuroscience
Tel Aviv University

“Decision-making: from adaptive mechanisms to biases and preference reversal”

Decision making is subject to both, remarkable adaptive fits as well as surprising biases that have puzzled decision theorists. Here I will review the neural mechanism of evidence-integration to a decision criterion that is central to the standard decision framework in neuroscience, and I will show how it can be extended to more ecological situations that involve changing environments and temporal uncertainty. Then I will show how this mechanism can also help to understand decision biases and preference-reversals. The model accounts for psychophysical data from controlled experiments, with perceptual and numerical alternatives that controls the decision-variables. Finally, I will discuss possible implications for the neural substrate.

Dr. Ernst Niebur

11/18/2013 4:00pm
James Bisley
Assistant Professor, Psychology

“The role of the parietal priority map in guiding visual attention”

Visual attention is the mechanism the nervous system uses to highlight specific locations, objects or features within the visual field. This can be accomplished by making an eye movement to bring the object onto the fovea (overt attention) or by increased processing of visual information in more peripheral regions of the visual field (covert attention). We have hypothesized that neurons within the lateral intraparietal area (LIP) of posterior parietal cortex create a priority map, which is used to guide these processes. In this talk, I will use evidence from a visual foraging task and a change detection task to illustrate the role that LIP plays in targeting eye movements and to how this role may intersect with LIPs role in guiding the allocation of covert attention.

Dr. Rudiger von der Heydt

11/25/2013 4:00pm
Xiaomo Chen, Ph.D.
Graduate Assistant
Johns Hopkins University

Advanced Researchers Lecture Series “The neuronal mechanism underlying value-based decision making”

Value-based decisions could rely either on the selection of goals or of actions. We investigated this question by recording from the supplementary eye field (SEF) of monkeys during an oculomotor gambling task. SEF neurons initially encode the option and action values associated with both alternative options. Competitive interactions between the options select one of them. However, SEF encodes the chosen option ~60-100 ms before the chosen action. These results suggest that SEF neurons form a map of the competing saccade targets. Activity within this map reflects first the chosen option and only afterwards the action that is necessary to obtain it. When neuronal activity in SEF was reversibly inactivated, monkeys chose significantly more often the less valuable option. The SEF population activity is therefore causally involved in the value-based selection of saccades. Our results suggest a new cascading model of value-based decision making consisting of partially overlapping goal and action selection processes.

Dr. Ernst Niebur

12/02/2013 4:00pm
Hui-Chen Lu, Ph.D.
Assistant Professor
Baylor College of Medicine

“The roles of mGluR5 and endocannabinoid signaling in developing cortical maps”

In cortical sensory maps, thalamocortical afferents (TCAs) transmit peripheral sensations in organized arrays into distinct cortical neuronal modules to provide a topographic representation of the external sensory world. These cortical maps, which form in every individual, can be altered by exposure to abnormal sensory experience during a “critical period” of postnatal development. Glutamatergic neurotransmission plays important roles in sensory map formation and the group I metabotropic glutamate receptor 5 (mGluR5) is required for sensory map formation. mGluR5 signaling has been implicated in the pathology of several significant neurological disorders, including Fragile X syndrome, ADHD, and schizophrenia. Dr. Lu’s laboratory recently found that eliminating mGluR5 function solely in cortical glutamatergic principal neurons, not only affects the cytoarchitectural formation of cortical neurons but also alters TCA arborization. In addition, mGluR5 in glutamatergic neurons is required for the maturation of GABAergic inputs onto these neurons. To further elucidate the cell-autonomous and nonautonomous influences of mGluR5 signaling in cortical glutamatergic neurons in modulating the formation of sensory circuits, they have generated mosaic mGluR5 animals through the in utero electroporation technique. Dr. Lu will present the results from their studies investigating the anatomical and functional properties of mGluR5-deleted neurons embedded in a wild type environment. Furthermore, she will talk about the potential role of endocannabinoid signaling in modulating TC connections. Endocannabinoids are synthesized upon mGluR5 activation and are best known as retrograde messengers in regulating synaptic transmission. Her data suggest that endocannabinoids mediate, at least in part, the mGluR5 dependent remodeling of TC synapses.

Dr. Hey-Kyoung Lee

12/09/2013 4:00pm
Davi Bock, Ph.D.
Lab Head
Howard Hughes Medical Institute

"Neuronal network anatomy from large-scale electron microscopy"

In mammalian cerebral cortex, axonal and dendritic overlap alone fails to predict the connectivity of excitatory neurons. The probability that one neuron forms a synapse with its neighbor depends on their in vivo tuning properties, long-distance targets, and local microcircuit connectivity. How do these factors integrate to generate network structure in cortex? And how does this network structure pattern relate to cortical information processing? Large-scale electron microscopy (EM) of cortex-scale volumes may help to answer these questions. In recent proof-of-principle work, we have shown that in vivo calcium imaging and EM can be combined, allowing physiology and anatomical connectivity to be compared across reasonably large numbers of cortical neurons in a single cortical layer. Currently we are scaling up the imaging capacity of our transmission electron microscope camera array (TEMCA) and designing new biology experiments. Work is also underway to automate the sectioning, staining, and imaging of thousands of serial thin sections in a hands-free fashion, and to improve correlative light and EM approaches. I will summarize past work and present our current effort to determine neuronal connectivity of physiologically characterized cells at the scale of cortical columns.

Dr. Steven Hsiao

01/06/2014 4:00pm
Jeffrey M. Yau, Ph.D.
Assistant Professor
Baylor College of Medicine

“Rethinking brain organization: Evidence for supramodal perceptual networks”

Primary sensory brain areas previously thought to be dedicated to a single modality can exhibit multimodal responses. Some have interpreted these responses as evidence for crossmodal recruitment (i.e., sensory processing for inputs in a non-primary modality); however, the direct contribution of this activity to perception is unclear. What experimental evidence would support a crossmodal recruitment hypothesis? I will first describe single-unit recordings in macaque monkeys that reveal parallel transformations of shape representations in the somatosensory and visual processing pathways. Parallel tuning patterns imply analogous coding mechanisms that may facilitate cross-modal information transfer. I will then summarize results from human psychophysical experiments that reveal highly specific perceptual interactions between touch and audition. Reciprocal audio-tactile interactions in the frequency domain suggest convergent and shared pitch representations. I will conclude with results from non-invasive brain stimulation studies in humans that reveal a double-dissociation in the crossmodal contributions of visual and auditory cortex to tactile perception. Together, these results provide indirect and direct evidence for a supramodal brain organization scheme – Sensory areas process multiple modalities and collaborate in supramodal perceptual networks. I will finish my talk by describing novel methods for dissecting functional networks using combined brain stimulation and neuroimaging.

Dr. Steven Hsiao

01/13/2014 4:00pm
Alla Karpova, Ph.D.
Lab Head
Janelia Farms Reserach Campus

“Decision-making under uncertainty: Probing the neural basis of mental models”

The overall interest of my lab is to understand how model-based inference is accomplished by neural circuits. Over the past few years we have focused on the role that the rodent medial prefrontal cortex (mPFC), an area homologous to primate anterior cingulate cortex (ACC), plays in encoding the internal representation of the rules of the environment. We have designed behavioral tasks in which these rules change suddenly or evolve in a very complex manner—in some cases eliciting abrupt changes in the workings of the internal model and in others leading to the abandonment of attempts at model construction. Recordings of the activity of neuronal ensembles in mPFC revealed that moments of abrupt change in behavioral strategy are associated with sudden transitions in the pattern of neural activity across the mPFC, one interpretation of which is that such changes signify a reset of prior expectations. In addition, inactivation of mPFC by local muscimol administration revealed that the influence of mPFC on behavior is suppressed when attempts to build an internal model are unsuccessful. In combination, our observations argue that mPFC represents an animal’s beliefs about the environment’s governing rules.

Dr. Steven Hsiao

01/27/2014 4:00pm
Eunjung Hwang, Ph.D.
Assistant Project Scientist
University of California, San Diego

“Visuomotor control in posterior parietal cortex and implications for BMIs”

The posterior parietal cortex (PPC) is critical for visuomotor control as evidenced by human optic ataxia in which reaching and grasping for visual objects are impaired following parietal lobe damage. Electrophysiological examinations of primate PPC have identified an area called the parietal reach region (PRR) that is specifically engaged in reach planning. Consequently, PRR has become a candidate area for reading out a subject’s reach plans for brain machine interfaces (BMIs). However, the neural circuits underlying reach plans in PRR and the viability of extracting reach plans from PRR in various BMI task conditions have not been thoroughly studied. I examined these issues in a series of experiments in the macaque PRR using multichannel extracellular recording, online decoding, and pharmacological inactivation. Pharmacologically disrupting PRR causes misreaching, the hallmark of optic ataxia, confirming the representation of motor intention in PPC. Moreover, I found that PRR represents the intended reach goal by integrating both bottom-up visual and top-down cognitive inputs, and provides robust and adaptable signals regarding when and where to reach for BMIs. I will present these findings in detail, as well as a brief overview of my current project developing two-photon calcium imaging based BMIs.

Dr. Kristina Nielsen

02/03/2014 4:00pm
Erik Emeric, Ph.D.
Postdoctoral Fellow
Johns Hopkins University

“The Role of Motor Cortex, Prefrontal Cortex, and the Basal Ganglia in the Control of Movement”

A number of brain structures have been implicated as playing a role in the control of movement. These structures include primary motor cortex (M1), the inferior frontal cortex (IFC), and the basal ganglia have been implicated as areas critical for controlling action. However, the specific function of these structures in the control of action is debated. To elucidate the role of these structures in the control of action, we recorded single unit activity from M1, IFC, and the basal ganglia of monkeys performing the stop signal task.

Dr. Ernst Niebur

02/17/2014 4:00pm
Don Katz, Ph.D.
Associate Professor Neuroscience
Brandeis University

“The cortical ensemble dynamics of decision-making and behavior”

Most of life’s choices are made suddenly and with minimal consideration. Neural activity predicting such choices has been described to build in a slow, ramp-like manner. In this talk, however, I will present data demonstrating that standard single-neuron analyses of choice-related activity can sometimes mask much more sudden decision-making processes: by such analyses, cortical responses to tastes appear to increasingly predict consumption decisions (i.e., reflect stimulus palatability) across the 500 ms preceding the onset of ingestive behaviors; analysis of simultaneously-recorded neural ensembles, in contrast, revealed this choice-related activity to appear in nearly instantaneous, coherent transitions. Furthermore, these ensemble transitions into choice-related firing, analyzed in single trials, accurately predict the timing of naturalistic choice behavior. These data suggest a novel, dynamical characterization of perceptual neural processing.

Dr. Marshall Shuler

02/24/2014 4:00pm
Krishnan Padmanabhan, Ph.D.
Junior Fellow
Salk Institute

“The role of biophysical diversity and anatomical connectivity in shaping neural computation”

The computations performed by neural circuits are shaped by the anatomy and the physiology of individual neurons. In the olfactory system of mammals, feedforward projections from the principal cells of the bulb relay information about odors directly to cortical regions such as the accessory olfactory nucleus and the piriform. Feedback projections from these cortical areas project back to cells of the bulb, dynamically impacting the way odor information is processed. In my talk, I will discuss two features of the circuit. 1) How the biophysical diversity of the mitral cells in the bulb improve information transmission in the feedforward direction. 2) How the anatomy of connections from the piriform cortex back to the bulb shape the flow of information sent in the feedback direction. I will conclude by discussing how both these properties affect the computations performed by the olfactory system.

Dr. Kristina Nielsen

03/10/2014 4:00pm
Huizhong Tao, Ph.D.
Assistant Professor
University of Southern California

“Synaptic mechanisms for visual cortical processing and development”

Specific visual processing functions are achieved by coordinated activity of functionally and neurochemically distinct excitatory and inhibitory neurons in the neural circuits. Combining in vivo imaging, voltage-clamp recording and optogenetics, we have determined how the interplay of excitatory and inhibitory synaptic inputs shapes the visual receptive field properties, including simple/complex receptive field structures, contrast-invariant orientation tuning and direction tuning in mouse visual cortex. Our results highlight a critical role of inhibitory circuits in refining cortical functional properties in the adult and developing brain.

Dr. Alfredo Kirkwood

03/19/2014 4:00pm
Jude Mitchell, Ph.D.
Staff Scientist
Salk Institute

“Neural mechanisms of attention and the marmoset as a model system”

An understanding of information processing at the level of cortical circuits remains a key challenge for understanding the brain and how the dysfunction of its circuits contributes to human disease. It has been appreciated for a long time that most of the connectivity in cortex is recurrent and involves extensive feedback from higher to lower areas. Behavioral paradigms using selective attention highlight the importance of these feedback connections in perception. My research with Old World monkeys has shown that attention reduces the variability of neuronal responses and thereby improves perception. Much of this variability originates from ongoing activity that is shared across populations (i.e., so-called “noise” correlations). I will outline a series of experiments that illuminate the neural mechanisms underlying attention-dependent reductions in variability. First I show that spatial attention increases local inhibition. Then I show in a realistic spiking network model how increases in inhibition shunt away variability from ongoing cortical activity. This model makes predictions for the role of different neuronal classes in attention. I am now testing these predictions using a New World monkey, the common marmoset, as a model system (Mitchell et al., J Neurosci, 2014). The marmoset offers several exciting opportunities for studying neural circuits in primates, including a lissencephalic (flat) cortex for superior laminar and array recordings as well as the development of the first primate transgenic lines and new genetic models of human mental disease.

Dr. Kristina Nielsen

03/24/2014 4:00pm
Joseph Monaco, Ph.D.
Research Fellow
JHU School of Medicine

Advanced Researchers Lecture Series Postponed to 4/21/14

The hippocampus is thought to play a critical role in episodic memory by incorporating the sensory input of an experience, such as when navigating mammals stop to fixate on familiar landmarks, onto a spatial framework embodied by place cells. However, the dual roles of spatial navigation and memory formation appear to conflict: Wayfinding requires spatial stability whereas memory depends on the continual encoding of the items and events of experience. I have investigated distinct neural coding mechanisms that together may provide stability while allowing for rapid, attentive integration of sensory information into hippocampal representations. First, I will describe computational modeling studies that derive from the temporal phase relationship between place cells and the theta (6-10 Hz) rhythm. Integrating self-motion signals in the phase of theta oscillations requires high precision and is not intrinsically robust to noise. By extending previous oscillatory interference models, I will show how an attentive interaction with external cues can provide feedback to theta oscillators, allowing for retrieval of positional fixed points and correction of phase drift, thus ensuring spatial stability. Second, more recent work has focused on behavior-based analysis of place cell recordings from rats navigating closed-loop tracks. While it is known that place fields are constructed through exploration, the interaction between discrete exploratory behaviors and episodic-like modifications of hippocampal representations has not been established. This study found that increased neural activity during exploratory head-scanning behaviors predicted the formation and potentiation of place fields on the next pass through that location. This result is consistent with current theories of hippocampal rate remapping and indexing of long-term memories, in which variations in firing rate encode information about discrete experiences. Thus, these studies begin to demonstrate how multiple neural coding mechanisms may cooperate within the hippocampus to allow for robust spatial navigation in a changing world.

Dr. Ernst Niebur

04/11/2014 4:00pm
Vincent McGinty, Ph.D.
Research Associate
Stanford University

“Orbitofrontal cortex neurons express a value code that depends on the target of gaze”

Humans constantly shift the location of their gaze, often targeting objects with motivational significance or value. While many primate physiology studies have shown how single neurons encode value, almost all used tasks that suppressed natural gaze shifts by requiring prolonged central fixation. As a result we know little of how single neurons encode object value during natural, active vision – a behavior that dictates our everyday waking experience. This is a particularly pertinent question given recent findings that both choices and neural representations of value in humans are influenced by where and when we shift our gaze. In this talk, I will describe our study of value coding in single neurons in a free-gaze Pavlovian conditioning task in macaque monkeys. Our key finding is that a third of orbitofrontal cortex (OFC) neurons are sensitive to both the value of a visible target and to the location of gaze with respect to that target. When two targets of differing value are shown simultaneously, monkeys often shift their gaze between them, and we find that some neurons rapidly shift their firing to reflect the value of the currently fixated target. The existence of gaze-specific value signals at the saccadic time scale in single neurons corroborates human fMRI findings, and is consistent with behavioral data suggesting that a gaze-specific value code underlies gaze-driven biases in decision-making. The mechanism by which gaze drives OFC value coding is not known, but a strong possibility – and a fertile area for future research – is that OFC neurons receive input from and inherit the properties of object-sensitive neurons in the superior temporal sulcus, a region that projects directly to OFC.

Dr. Ed Connor

04/21/2014 4:00pm
Joseph Monaco, Ph.D.
Research Fellow
JHU School of Medicine

Advanced Researchers Lecture Series “Landmark influence: How attention to sensory cues stabilizes and updates the hippocampal cognitive representation of space”

The hippocampus is thought to play a critical role in episodic memory by incorporating the sensory input of an experience, such as when navigating mammals stop to fixate on familiar landmarks, onto a spatial framework embodied by place cells. However, the dual roles of spatial navigation and memory formation appear to conflict: Wayfinding requires spatial stability whereas memory depends on the continual encoding of the items and events of experience. I have investigated distinct neural coding mechanisms that together may provide stability while allowing for rapid, attentive integration of sensory information into hippocampal representations. First, I will describe computational modeling studies that derive from the temporal phase relationship between place cells and the theta (6-10 Hz) rhythm. Integrating self-motion signals in the phase of theta oscillations requires high precision and is not intrinsically robust to noise. By extending previous oscillatory interference models, I will show how an attentive interaction with external cues can provide feedback to theta oscillators, allowing for retrieval of positional fixed points and correction of phase drift, thus ensuring spatial stability. Second, more recent work has focused on behavior-based analysis of place cell recordings from rats navigating closed-loop tracks. While it is known that place fields are constructed through exploration, the interaction between discrete exploratory behaviors and episodic-like modifications of hippocampal representations has not been established. This study found that increased neural activity during exploratory head-scanning behaviors predicted the formation and potentiation of place fields on the next pass through that location. This result is consistent with current theories of hippocampal rate remapping and indexing of long-term memories, in which variations in firing rate encode information about discrete experiences. Thus, these studies begin to demonstrate how multiple neural coding mechanisms may cooperate within the hippocampus to allow for robust spatial navigation in a changing world.

Dr. Ernst Niebur

09/22/2014 4:00pm
Katalin Gothard, Ph.D.
Associate Professor
University of Arizona

“The eyes--a window to the social brain”

Primates explore the visual world through eye-movement sequences. Saccades bring details of interest into the fovea while fixations stabilize the image. During natural vision, social primates direct their gaze at the eyes of others to communicate their own emotions and intentions and to gather information about the mental states of others. Direct gaze is an integral part of facial expressions that signals cooperation or conflict over resources and social status. Despite the great importance of making and breaking eye contact in the behavioral repertoire of primates, little is known about the neural substrates that support these behaviors. In this talk, I will show that the monkey amygdala contains neurons that respond selectively to fixations at the eyes of others and to eye contact itself. These “eye cells” share several features with the canonical, visually responsive neurons in the monkey amygdala, however, they respond to the eyes only when they fall within the fovea of the viewer, either as a result of a deliberate saccade, or as eyes move into the fovea of the viewer during a fixation intended to explore a different feature. The presence of eyes in peripheral vision fail to activate the eye cells. These findings link the primate amygdala to eye movements involved in the exploration and selection of details in visual scenes that contain socially and emotionally salient features.

Dr. Ernst Niebur

09/29/2014 4:00pm
Jianhua Cang, Ph.D.
Associate Professor, Northwestern U.
Weinberg College of Arts & Sciences

“Critical Period Plasticity and Binocular Matching in the Mouse Visual Cortex”

Experience shapes neural circuits during critical periods in early life. My lab has demonstrated that critical period plasticity drives binocular matching of orientation preference in the mouse visual cortex (Wang, Sarnaik, and Cang, 2010). Although studies have started to reveal the genetic and epigenetic mechanisms that control the opening and closure of the critical periods, the functional significance of a properly-timed critical period in visual system development is not yet clear. In this seminar, I will present our recently-published data that address this question.

Dr. Alfredo Kirkwood

10/13/2014 4:00pm
Andrea Chiba, Ph.D.
Associate Prof of Cognitive Science
University of California, San Diego



Dr. Ernst Niebur

10/20/2014 4:00pm
Rich Krauzlis, Ph.D.
Senior Investigator
National Eye Institute

“A new framework for thinking about attention”

Attention is commonly believed to be controlled by a network of areas in the cerebral cortex, with frontal and parietal cortex regulating limited resources available in the sensory areas of cortex. However, subcortical structures like the superior colliculus also play a role in attention, and in this talk I will explain how our investigation of the superior colliculus has led us to a very different view of how attention is controlled. I will present evidence that the superior colliculus plays a crucial role in the control of spatial attention, but surprisingly, the mechanisms used by the superior colliculus appear to be independent of the well-known signatures of spatial attention in visual cortex. These recent results demonstrate that processes beyond the well-known correlates in extrastriate cortex play a major role in visual spatial attention. Furthermore, based on recent results from fMRI and physiology experiments in my lab, as well as clues from neuroanatomy and disorders of attention, I speculate that the brain mechanisms for attention are based on an evolutionarily conserved ciruit motif that predates the emergence of the neocortex.

Dr. Manuel Gomez-Ramirez

11/03/2014 4:00pm
Mark Baxter, Ph.D.
Professor of Neuroscience
Icahn School of Medicine at Mt. Sinai

“Cognitive and socioemotional development after postnatal anesthetic exposure”

Exposure to general anesthetics during the postnatal period causes neuron death and long-lasting alterations in neuronal morphology and plasticity in rodents. Some epidemiological studies in humans have found impaired cognition in individuals that underwent anesthesia as children (before the age of 4), especially associated with repeated exposure to anesthesia. But studies of neurotoxicity in rodents are not definitive and alternative explanations other than anesthetic neurotoxicity could account for the observations in humans. Thus we have turned to a nonhuman primate model, that has protracted neural development like that of humans, to address the question of whether anesthetic exposure is sufficient to alter neurocognitive development. Rhesus monkeys given 3 4-hour exposures to sevoflurane during the first 6 weeks of life display abnormal emotional behavior at the age of 6 months. Effects of anesthetic exposure on visual recognition memory appear to differ by sex. We are continuing to follow the behavior of these monkeys longitudinally to get a fuller picture of the long-term effects of early anesthetic exposure on cognition. Parallel mechanistic studies in rodents support a threshold effect of anesthetic exposure on mitochondria which sensitizes neurons to damage from repeated exposure. These studies suggest that early exposure to anesthesia may be sufficient to impair cognition in humans, and promote the search for strategies to maintain the beneficial effects of anesthesia that allow safe surgery while protecting central nervous system function.

Dr. Marshall Shuler

11/10/2014 4:00pm
Friedrich Sommer, Ph.D.
University of California, Berkeley

“Decoding the hippocampal theta wave in navigating rat”

Extracellular multi-electrode recordings allow monitoring of both spiking activity and local field potentials (LFP). The LFPs reflect a superposition of spiking, synaptic and subthreshold activity. In the hippocampus of a navigating rat, place neurons spike selectively according to location, while LFPs exhibit a powerful traveling wave at 9Hz propagating through the hippocampus. The LFPs exhibit very little place-tuning at single anatomical sites. Previously it was believed that the 9Hz wave represents the animal's behavioral state but contains little specific information about the animal's location. We recently demonstrated the opposite: the LFP structure at multiple sites contains detailed information about the place of the animal. Specifically, the location of the animal can be decoded from the LFP with comparable precision as can be done from spike trains detected in the same recordings. Further, unsupervised learning methods can extract LFP place components that tile the animals enclosure. I will present a possible explanation of this finding based on compressed sensing, as well as describe its broader implications for decoding neural activity in non-topographically arranged neurons.

Dr. Ernst Niebur

11/20/2014 12:00pm
Philip O'Herron
Post Doctoral Fellow, Neuroscience
Medical University of South Carolina

“Neural correlates of single-vessel hemodynamic responses in vivo”


Dr. Rudiger von der Heydt

11/20/2014 4:00pm
Inah Lee, Ph.d.
Assoc Professor, Brain & Cognitive
Seoul National University

“Visual object, scene, and the hippocampal system”

The hippocampus and its associated neural networks are important for remembering events in space. Among other inputs that influence information processing in the hippocampal networks, visual information has been suggested to play critical roles. For example, a visual background, or scene, exerts a powerful influence when an animal forms a memory of an environment along with the events that take place in the environment. Such scene-dependent memory relies on the hippocampus and its associated networks. In comparison to the extensive literature on spatial learning and memory, however, the visual scene-dependent memory has been investigated less in rodents especially using electrophysiological recording techniques in freely moving rats. A recent trend of using mice and rats for optogenetic techniques and other high-throughput imaging and physiological techniques have rejuvenated the interest in investigating the rodent visual systems. This trend opens a great opportunity for understanding how complex visual information such as background scene and object information is registered in the higher visual systems, and how the representation is transformed and reaches the hippocampus for allowing recognition memory. In our laboratory, we have established that rats are very good at making visual scene-based choices (both spatial and nonspatial choices) and the hippocampus is necessary for this behavior. In this talk, I will present what we have found so far in the visual scene- and visual object-dependent memory tasks in the hippocampus and in the perirhinal cortex.

Dr. James Knierim

11/20/2014 2:00pm
Kevin Fox, Ph.D.
Professor, Head of Research
Cardiff School of Biosciences

“Circuits and mechanisms for plasticity in cortical layer V RS and IB cells”

Most functional plasticity studies in the cortex have focused on layers (L) II/III and IV, whereas relatively little is known of LV. Layer V contains cells that form the major output projection from the cortex. Structural measurements of dendritic spines in vivo suggest some specialisation among LV cell subtypes. We therefore studied experience-dependent plasticity in the rat barrel cortex using intracellular recordings to distinguish regular spiking (RS) and intrinsic bursting (IB) subtypes. Post-synaptic potentials and supra-threshold responses in vivo revealed a remarkable dichotomy in RS and IB cell plasticity; spared whisker potentiation occurred in IB but not RS cells while deprived whisker depression occurred in RS but not IB cells. Modelling studies showed that subthreshold changes predicted the supra-threshold plasticity. Similar RS/IB differences were found in the LII/III to V connections in brain slices. These studies demonstrate the major functional partition of plasticity within a single cortical layer and reveal the LII/III to LV connection as a major excitatory locus of cortical plasticity. We have recently generalised these findings to the mouse barrel cortex. The major distinction in functional plasticity between RS and IB cells is conserved between species. Using knockout mice to target key molecules for homeostatic and Hebbian forms of plasticity we have begun to dissect the mechanisms involved in RS and IB cell experience-dependent plasticity.

Dr. Hey-Kyoung Lee

11/21/2014 4:00pm
Ziad Hafed, Ph.D.
Group Leader
Werner Reichardt Centre Int Neuro

“Alteration of visual perception around the time of microsaccades: mechanisms and implications”

Neuronal modulations such as response gain enhancement and reductions in response variability are classically thought to reflect the allocation of covert visual attention to behaviorally relevant stimuli. In this talk, I will describe results showing that these classic neuronal signatures of covert visual attention can occur without any attentional task at all, simply as a consequence of generating tiny subliminal eye movements. In six different macaque monkeys and two different brain areas classically implicated in covert visual attention, we found that classic neuronal signatures of covert visual attention can occur if stimuli appear immediately prior to tiny eye movements called microsaccades. Our results suggest that there is an obligatory link between pre-motor processes and neuronal or behavioral signatures of selective visual processing, even when such pre-motor processes are associated with seemingly irrelevant subliminal motor outputs. These results also form a direct neuronal correlate of recent behavioral results in which I found that the largest changes in attentional performance in classic cueing tasks occur in pre-microsaccadic intervals. Even when target selection is forced to occur without overt actions, covert selective processing may nonetheless intrinsically remain an “active perception” phenomenon.

Dr. Kristina Nielsen

11/21/2014 12:00pm
Norbert Fortin, Ph.D.
Assistant Professor, Neurobiology
University of California, Irvine

"The neurobiology of memory for sequences of events: A synergistic approach in rats and humans"

It is well established that the ability to temporally organize information is fundamental to many perceptual, cognitive, and motor processes. Temporal organization is also critical to memory. In fact, since many of our memories have overlapping elements, including specific items and locations, our capacity to distinguish individual memories critically depends on remembering their unique temporal context. Unfortunately, while our understanding of how the brain processes the spatial context of memories has advanced considerably in recent years, our understanding of their temporal organization lags far behind. The overall objective of our research is to understand the fundamental neurobiological mechanisms underlying the memory for sequences of events and the memory for elapsed time. In this seminar, I will primarily focus on our recent work on sequence memory in rodents in which we used localized brain inactivations and single-cell recordings to help elucidate the contributions of the hippocampus and prefrontal cortex. I will also present recent findings from our parallel work in human subjects, which suggests that rats and humans use similar strategies, cognitive processes and neural circuits to remember sequences of items and that this capacity is impaired in normal aging. I will conclude by discussing the importance of developing integrated, cross-species approaches to advance basic and clinical memory research.

Dr. James Knierim

11/21/2014 2:00pm
Nachum Ulanovsky
Principal Investigator

“Neural codes for 2-D and 3-D space in the hippocampal formation of bats”

The work in our lab focuses on understanding the neural basis of spatial memory and spatial cognition in freely-moving, freely behaving mammals – employing the echolocating bat as our animal model. I will describe our recent studies, including: (i) Recordings of 3-D place cells, 3-D grid cells, and 3-D head-direction cells in the hippocampal formation of freely-flying bats, using a custom neural telemetry system – which revealed an elaborate 3-D spatial representation system in the bat’s brain; and (ii) Absence of theta oscillations in the bat’s hippocampal formation – arguing against a central role of theta in spatial cognition. I will also describe our recent studies of spatial memory and navigation of bats in the wild, using micro-GPS devices, which revealed outstanding navigational abilities and provided the first evidence for a large-scale 'cognitive map' in a mammal.

Dr. Cynthia Moss

11/24/2014 4:00pm
Shawn Mikula, Ph.D.
Max Planck Institute, Heidelberg

"High-Resolution Whole-Brain Staining for Electron Microscopic Circuit Reconstruction"

Neuronal circuits are central to neural computation. Currently only electron microscopy provides the resolution necessary to reconstruct such circuits completely and with single-synapse resolution. Because almost all behaviors rely on neural computations that are widely distributed throughout the brain, a reconstruction of the entire brain is highly desirable but requires the undivided brain to be prepared for electron-microscopic observation. Here we describe a preparation, BROPA (Brain-wide Reduced-Osmium staining with Pyrogallol-mediated Amplification), that results in the brain-wide preservation and staining of all ultrastructural details at a resolution necessary for tracing neuronal processes and identifying synaptic contacts between them. Using serial block-face electron microscopy (SBEM) we tested in a number of diverse regions the ability to follow neural "wires" reliably and over long distances as well as the ability to detect synaptic contacts. Our results suggest that the BROPA method can produce a preparation suitable for obtaining a comprehensive circuit map of an entire mouse brain.

Dr. Ernst Niebur

01/26/2015 4:00pm
Doris Tsao, Ph.D.
Assistant Professor
California Institute of Technology

“Mechanisms for Object Recognition in the Macaque”

Using fMRI, one observes regions in both the human and macaque brain that appear specialized for representing specific object categories such as faces and bodies. In my talk, I will discuss the anatomical and functional organization of the macaque face and scene processing systems, including new work probing the mechanism for representing view-invariant identity.

Dr. Ed Connor

02/16/2015 4:00pm
Benjamin Hayden, Ph.D.
Assistant Professor
University of Rochester

“Neuronal Basis of Economic Choice”

My lab is interested in determining the neural operations by which reward-based choice occurs. We hypothesize that prefrontal and striatal reward regions are organized into a processing hierarchy that implements evaluation, comparison, and selection steps of choice. We record activity of single neurons in several regions, including OFC, vmPFC, VS, and dACC while macaques perform various economic choice tasks. We propose that OFC can be placed early in this process, that vmPFC and VS can be placed at an intermediate stage, and that dACC can be placed relatively late in the process. We further propose that the comparison process itself occurs through a mutual inhibition of activity of single neurons. Collectively, our data suggest that economic choice is both parallel and distributed broadly throughout the prefrontal cortex and striatum.

Dr. Veit Stuphorn

03/09/2015 4:00pm
Michael Beierlein, Ph.D.
Assistant Professor
University of Texas Medical School

“Mechanisms controlling thalamic circuit dynamics”

Synchronous neuronal activity in the thalamocortical system is critical for a number of behaviorally relevant computations, but hypersynchrony can limit information coding and lead to epileptiform responses. In the somatosensory thalamus, afferent inputs are transformed by networks of reciprocally connected thalamocortical neurons in the ventrobasal nucleus (VB) and GABAergic neurons in the thalamic reticular nucleus (TRN). These networks can generate oscillatory activity and studies in vivo and in vitro have suggested that thalamic oscillations are often accompanied by synchronous neuronal activity, in part mediated by widespread divergence and convergence of both reticulothalamic and thalamoreticular pathways, as well as by electrical synapses interconnecting TRN neurons. However, the functional organization of thalamic circuits and its role in shaping input-evoked activity patterns remain poorly understood. We show that optogenetic activation of cholinergic synaptic afferents evokes near-synchronous firing in mouse TRN neurons which is rapidly desynchronized in thalamic networks. We identify several mechanisms responsible for desynchronization: shared inhibitory inputs in local VB neurons leading to asynchronous and imprecise rebound bursting, TRN-mediated lateral inhibition which further desynchronizes firing in VB, and powerful yet sparse thalamoreticular connectivity which mediates re-excitation of TRN but preserves asynchronous firing. Our findings reveal how distinct local circuit features interact to desynchronize thalamic network activity.

Dr. Hey-Kyoung Lee

03/16/2015 4:00pm
Sliman Bensmaia, Ph.D.
Associate Professor
University of Chicago

“Spatial and Temporal Codes in Touch”

A fundamental question in sensory neuroscience is how patterns of neuronal activity convey information about objects and culminate in perception. Here, we present evidence that the tactile perception of texture relies on a variety of complementary codes, both across neurons and across time, that in combination account for perception. During texture exploration, the human fingertip undergoes complex mechanical transformations, which culminate in a spatial distribution of strains and stresses in the skin. At the same time, movement between skin and surface elicits low-amplitude, high-frequency, and texture specific surface waves that propagate across the finger skin. These different aspects of skin deformation — some spatial, other temporal — are transduced by different receptors embedded in the skin: The coarse spatial layout of textures is encoded in the spatial pattern of activation of one population of afferents, while fine textural features are encoded in precise temporal spiking patterns in two others. The perception of texture thus arises from the combination of spatial and temporal codes distributed across three afferent classes. In primary somatosensory cortex (S1), one subpopulation of neurons, whose receptive fields consist of excitatory and inhibitory sub-regions, is well suited to extract texture information from the spatial neural image that reflects coarse surface features. Another subpopulation of S1 neurons is highly sensitive to texture-like skin vibrations and thus well suited to convey information about fine textural features. In the latter population, information about skin vibrations is multiplexed in temporal and rate codes: frequency composition is encoded in the timing of the response whereas amplitude is encoded in the spike count. Like its peripheral counterpart, then, the cortical processing of texture relies on different subpopulations of neurons that implement different neural codes.

Dr. Ed Connor

03/30/2015 4:00pm
Tirin Moore, Ph.D.
Associate Professor, Neurobiology
Stanford University

“A Role for Motor Control Mechanisms in Visual Perception and Cognition”

Recent work indicates that structures principally involved in movement planning and initiation also appear to influence sensory processing and underlie aspects of cognition. For example, neurons within the frontal eye field (FEF), an area of prefrontal cortex, play a key role in the programming and triggering of saccadic eye movements, but also influence signals within posterior visual cortex. This work not only implicates the FEF in the control of visual selective attention, but it demonstrates the profound influence that motor control circuitry can have on the processing within posterior sensory representations. I will describe recent evidence of this influence and discuss its relationship to perception and cognition.

Dr. Rudiger von der Heydt

04/06/2015 4:00pm
Albert Lee, Ph.D.
HHMI, Janelia Farms Research

“Rules and mechanisms governing hippocampal representations”

The hippocampus is crucial for the formation of new long term memories of facts and events in humans as well as for spatial learning and memory in rodents. Extracellular recordings from the rodent hippocampus have revealed that a spatial map rapidly forms to represent each new environment an animal encounters. The representation for a given environment consists of place cells, each of which fires selectively whenever the animal is at particular locations (called the cell’s place fields) within the environment, and silent cells, each of which fires few or no spikes across the environment. In one set of studies, we used intracellular recordings in freely moving rats to explore the cellular mechanisms underlying individual place cell activity. Models have generally assumed that synaptic inputs from upstream neurons, many of which are themselves spatially tuned, summate to determine place cell firing as well as the silence of silent cells. However, we found that cellular excitability appears to play a surprisingly strong role in determining which neurons will be place and silent cells in a novel environment, with place cells being more excitable than silent cells. Using extracellular experiments, we investigated the detailed structure of the spatial map, searching for rules governing how place fields are distributed across space and neurons. We employed a very large environment (48-meter-long track) to facilitate statistical analyses of the distribution of fields. We found that a parsimonious, 2-parameter stochastic model could account for the observed features of hippocampal spatial representations. Briefly, each neuron could be described as having a cell-specific propensity for randomly forming a place field in any given location, and there was a specific distribution of propensities across different neurons with fewer cells having higher propensities and more cells having lower propensities of field formation. This structure has implications for how individual locations, different environments of a wide range of sizes, and experiences with a wide range of durations are encoded. Finally, we have been using intracellular recordings to explore how the hippocampal representation of a newly encountered space evolves with experience to become a stable, long-term memory. The results we will describe reveal both new phenomena as well as new questions.

Dr. James Knierim

04/27/2015 4:00pm
Anitha Pasupathy, Ph.D.
University of Washington, Seattle

“Discriminating partially occluded objects: insights from visual and frontal cortex”

Occlusions are ubiquitous in natural vision and they pose a major challenge to successful object recognition by artificial vision systems. The primate visual system, however, deals effortlessly with occlusions – we recognize partially occluded objects rapidly and accurately, but the underlying neural mechanisms are largely unknown. Visual cortical area V4, an intermediate stage of shape processing in ventral visual cortex, and the Prefrontal cortex likely play an important role in this perceptual capacity. To delineate their relative roles, we measured the responses of neurons in V4 and the ventrolateral prefrontal cortex (vlPFC) while monkeys discriminated pairs of shapes under varying degrees of occlusion. Our experimental results suggest that PFC circuits can selectively augment and clarify impoverished shape information in feedforward signals, thereby contributing to enhanced shape selectivity in visual cortex and to the successful discrimination of partially occluded shapes. I will present our data and a V4-PFC network model based on these findings, and I will discuss the role for feedback mechanisms in object recognition and image segmentation under occlusion.

Dr. Kristina Nielsen

04/28/2015 4:00pm
Wyeth Bair, Ph.D.
Assistant Professor
University of Washington, Seattle

“Modeling the ventral and dorsal visual pathways”

I will present results from our modeling studies of mid-level form processing in the ventral visual pathway and motion processing in the dorsal pathway, and I will describe our efforts to unify these models in a cloud-based framework. Since the work of Pasupathy and Connor, it is well known that many V4 neurons are tuned for the boundary curvature of simple shapes in an object centered coordinate system. Several models have been proposed to account for this type of tuning, but I will show that none of the major models, including spectral receptive field and H-max models, appear to capture the essence of the V4 representation. I will describe two paths to achieving better models, one based on modifying existing biologically plausible models, and the other based on a machine learning approach involving deep networks. In the dorsal pathway, we have developed image-computable models of the pattern-motion selectivity exhibited by neurons in area MT. I will show show how these models can be refined to explain recent results regarding binocular motion integration and 3D motion sensitivity. Our goal is to present all models in a unified, image-computable and online framework that facilitates collaborative testing and refinement of the models.

Dr. Kristina Nielsen

07/20/2015 4:00pm
Elias Issa, Ph.D.
Research Scientist
Massachusetts InstituteTechnology

“Neural signals for performing visual inference and learning in the cortical hierarchy”

Feedforward processing increases neural selectivity for objects across stages of the ventral visual hierarchy. These initial, bottom-up estimates could, in principle, be improved upon by online processing (inference) or by synaptic plasticity offline (learning). Optimal inference and learning, however, are driven by error signals, and it remains unclear whether these signals are represented by the visual system. Here, we recorded across multiple hierarchical levels of inferior temporal cortex (IT) and found that error signals are encoded in the dynamical signatures of IT neurons. This interpretation of neural activity as error coding accounted for a number of seemingly disparate neural phenomena observed in previous experiments -- sublinear input integration, temporal sharpening of responses to familiar images, and the mixed modulatory effects of feedback -- unifying them under one coherent computational framework. Future experiments will test the putative role of feedback in generating error signals and will test whether these signals drive inference, learning, or both. In particular, I will outline our current efforts to develop an emerging animal model, the common marmoset, for next generation experiments leveraging the latest neuroscientific tools to study the role of feedback in high-level vision.

Dr. Ed Connor

07/20/2015 4:00pm
Elias Issa, Pg\h.D.



Dr. Ed Connor

07/24/2015 12:00pm
Hendrijke Nienborg, Ph.D.
"Context-dependent visual processing for Perceptual decisions"
Centre for Integrative Neuroscience

“Context-dependent visual processing for perceptual decisions”

Perceptual decisions are influenced by the behavioral or sensory context in which a subject makes a decision. It is increasingly clear that such contextual influences involve the earliest stages of sensory processing. However the neural circuit elements responsible for such contextual modulation, and their computational roles are still poorly understood. The central goal of my lab is to further such understanding using a variety of approaches. During perceptual decision tasks, the activity of individual sensory neurons correlates with an animal’s decision. The origin of this decision-related activity in sensory neurons has long been explained as the causal effect of feed-forward noise in sensory neurons on perceptual decisions. My previous work using recordings in visual cortex of macaques performing a visual discrimination task challenged this dominant view. It suggested that a significant component of decision-related activity in visual neurons reflects top-down signals. In my current work we used statistical modeling to dissect decision-related activity in macaque V2. We were able to isolate a top-down component that reflects the animal’s bias going into the trial. It is compatible with the influence of a continuously updated belief about the sensory stimulus for perceptual inference. Important classes of regulatory circuit elements for these influences are inhibitory interneurons, global neuromodulators and feedback projections. My past and current work focuses on the first two types of these regulatory circuit elements. Using optogenetic techniques currently not available in the macaque I examined the contribution of different types of interneurons to the contextual modulation of the responses in mouse primary visual cortex. Our current work probes the role of the neuromodulator serotonin on the visual responses in the macaque early visual cortex. Together, our work supports the view that the diverse circuit elements responsible for contextual sensory modulation may be required to achieve perceptual inference while accounting for complex beliefs in natural task.

Dr. Kristina Nielsen

07/27/2015 4:00pm
Alden Hung, Ph.D.



Dr. Ed Connor

07/27/2015 4:00pm
Alden Hung, Ph.D.

“Investigations of neuronal mechanisms for social visual perception in macaques and marmosets”

The ventral visual pathway of the primate brain supports important aspects of high-level vision, including those pertaining to object recognition and social perception. My research goal is to understand the neural mechanisms underlying this process by studying and comparing two model non-human primate species, macaques and marmosets. My recent research has focused on the common marmoset, a New World monkey with transgenic potential. Using fMRI and electrophysiological recording, we have established the existence of multiple face-selective patches that bear a strong resemblance to those known to exist in the macaque. We are currently investigating the possible homological correspondence among visual areas based on whole-brain fMRI collected from both macaques and marmosets engaged in naturalistic viewing tasks. These studies suggest several parallel processing streams along the primate ventral pathway that contribute to high-level visual perception.

Dr. Ed Connor

10/12/2015 4:00pm
Lauren Barghout, Ph.D.
Visiting Scholar, BISC Program
University of California, Berkeley

“Top-down image segmentation via fuzzy spatial-taxon cut”

Computer vision is typically thought of as an open universe problem because every possible outcome is unknown. Since humans can recursively access multiple perceptually organized visual structures, they have unbounded creativity in the way in which they can interpret visual scenes. Matters are further complicated by the fact that the same image may convey multiple meanings depending on the context or task at hand of the viewer. These issues present a problem for automated image segmentation, because they add uncertainty to the process of selecting which objects to include or not include within a segment. In this talk, I discuss how to frame image segmentation as a closed universe problem that assumes that scenes are organized according to a standardized natural-scene-taxonomy, comprised of spatial-taxons. Spatial-taxons are regions (pixel-sets) that are figure-like, in that they are perceived as having a contour, are either `thing like', or a `group of things', that draw our attention. Defined in this way, the image segmentation problem can be operationalized into a series of iterative two-class fuzzy inferences and scene organization is viewed a recursive spatial-taxon inference. This system provides a top-down computer vision model of scene organization.

Dr. Rudiger von der Heydt

11/02/2015 4:00pm
Roozbeh Kiani, M.D., Ph.D.
Assistant Professor
Center for Neural Science New York University

“Neural mechanisms of strategic adjustment of perceptual decisions”

Slowing down following a mistake (Post-error slowing, PES) is often observed in humans but its neural mechanism and adaptive role remain controversial and less understood. Here we studied changes in the neural mechanisms of decision-making after errors in humans and monkeys that performed a motion-direction discrimination task. We found that PES is mediated by two factors: a reduction in sensitivity to incoming sensory information and an increase in accumulated evidence for the decision (decision criterion or bound). Both effects are implemented through dynamic changes in the decision-making process. Neuronal responses in the monkey lateral intraparietal area (LIP) revealed that bound changes are implemented by decreasing an evidence-independent urgency signal. They also revealed a reduction in the rate of evidence accumulation, reflecting the reduced sensitivity. These changes in the bound and sensitivity provided a quantitative account of choices and reaction times in PES. Sensory and motor delays, sensory and accumulation noise, and static parameters of the decision-making process remained unchanged. We suggest that PES is a manifestation of adaptive increase of decision bound in anticipation of potentially maladaptive changes in sensitivity to incoming evidence.

Dr. Ed Connor

02/01/2016 4:00pm
John Allman, Ph.D.
Professor of Neurobiology
California Institute of Technology

“Dendritic Spines in Alzheimer's Disease”

We have analyzed gene expression from cerebral cortical samples using RNA-Seq, which provides quantitative expression levels for the entire set of genes in the sample. Our samples were small pieces dissected from frozen fronto-insular cortex, an area of the brain involved in social cognition and decision-making. The samples were obtained from autopsy brains from elderly volunteers, who participated in a comprehensive study of aging and dementia at Rush University Medical School. See: The cognitive functioning of these individuals was carefully monitored during life, and they were classified as either normal, mild-cognitively impaired, or demented based on a battery of neuropsychological tests and clinical observations repeated annually. Their clinical histories were very well documented, and each sample had a thorough neuropathological evaluation of amyloid plaques and other hallmarks of Alzheimer's disease and other dementias. Overall, we found reduced expression of many genes expressed in the dendritic spines in normal aging, which is compatible with normal cognitive functioning. However, we found in the demented samples significantly increased expression of more than 30 genes involved in the machinery of dendritic spines. This observation is consistent with research in other labs done with electron microscopy showing an overall reduction in the number of spines but a paradoxical enlargement of the surviving spines in Alzheimer's disease. Taken together these results imply that the higher activity and enlargement of the spine machinery may be destructive to the spines and ultimately cause the death of the pyramidal cells from which they arise.

Dr. Ernst Niebur

02/15/2016 4:00pm
Howard Egeth, Ph.D.
Psychological & Brain Sciences

“The Dark Side of Attention: Ignoring and Learning to Ignore”

Current research in my lab is focused on how we manage to attend to relevant information and ignore irrelevant information even when it is highly salient. We are exploring the role of learning and are using both behavioral and neuroimaging techniques (especially EEG/ERP). Some of our recent work suggests that inhibition of irrelevant material, rather than enhancement of relevant material is what drives attentional selectivity.

Dr. Rudiger von der Heydt

02/22/2016 4:00pm
Vivek Jayaraman, Ph.D.
Group Leader
HHMI Janelia Research Center

“The neural dynamics of fly navigation”

Fruit flies, like many other animals, display a range of sophisticated behaviors including visual navigation and place learning. Behavioral genetics studies have implicated a central brain region called the central complex in such adaptive integration of sensory and motor information. My lab uses a combination of physiological and optogenetic techniques in head-fixed behaving flies to identify and understand circuit computations carried out by this intriguing brain region. I will discuss recent results from two–photon calcium imaging experiments performed during tethered walking behavior in virtual reality. Our experiments reveal population activity patterns in the central complex that resemble dynamics theorized to underlie the computation of head direction in mammals. We are now attempting to harness the power of fly genetics in combination with modeling, physiology and selective neural perturbation to understand how these circuit dynamics arise and how they relate to the fly's behavioral decisions.

Dr. James Knierim

02/29/2016 4:00pm
Mayank Mehta, Ph.D.

“Virtual reality elucidates the mechanisms of hippocampal spatial and directional selectivity”

All animals move through space. What are the sensory and biophysical mechanisms that generate mental maps of space? How do these maps contribute to memory guided behavior? While tremendous progress has been made, these questions have not been fully resolved, partly because it is difficult to precisely measure, let alone manipulate, the wide range of sensory and motor variables that change when subjects move in real space. Hence, we have developed a noninvasive, immersive and multisensory virtual reality system where precisely controlled stimuli determine the surrounding virtual space, and nonspecific stimuli are spatially uninformative. We simultaneously measured rats’ behavioral performance and the activities of thousands of neurons from the hippocampal circuit while rats performed complex tasks, including the Virtual Morris Water Maze task. We also developed computational techniques to decipher the emergent neural dynamics. This integrative, experiment-theory approach provided many surprising results which I will describe.

Dr. Ernst Niebur

03/07/2016 4:00pm
Alejandro Schinder, Ph.D.
Leloir Institute, Argentina

“Adult neurogenesis as a mechanism for hippocampal remodeling and spatial discrimination”

The dentate gyrus of the hippocampus presents unique forms of plasticity conferred by the continuous generation of granule cells (GCs) that make millions of new connections every day and, thus, remodel the existing networks. Evidence accumulated during recent years has shown that mice display impairment in spatial discrimination when adult neurogenesis is reduced or new neurons are silenced. My lab has focused on understanding how new GCs modify the preexisting circuits in a manner that is meaningful for information processing. We found that developing GCs are functionally relevant before reaching maturity due to a combination of intrinsic and network properties that make them very active and highly susceptible to activity-dependent synaptic modification, distinct from all other neurons in the circuit. Thus, integration of new neurons in the network is largely tailored by ongoing experience. In my talk I will dig into mechanisms that shape development, integration and function of new GCs and discuss their specific contribution for the discrimination of similar inputs.

Dr. Alfredo Kirkwood

03/28/2016 4:00pm
Douglas Nitz, Ph.D.
Associate Professor
University of California, San Diego

“Encoding of Axis and Analogy in the Rat Subiculum”


Dr. James Knierim

04/04/2016 4:00pm
Howard Egeth, Ph.D.
Prof. Psychological & Brain Sciences
Johns Hopkins University

POSTPONED A new date will be announced soon.


Dr. Rudiger von der Heydt

04/18/2016 4:00pm
Winrich Freiwald, Ph.D.
Assistant Professor
Rockefeller University

“On the Neural Circuits for Face Recognition: A Vision Science Perspective on the Social Brain”

Humans, like all primates, take great interest in faces. This is because faces, by structure and internal dynamics, display a plethora of social information for a visual system that can extract it. The primate brain does this through specialized hardware. The functional organization of these specializations, a network of tightly interconnected areas packed with face cells, each tuned to a different dimension of facial information, provides us with a unique model system to understand the computational principles and neural mechanisms of visual object recognition. Yet faces are special: they are not only a particularly well-defined object category, but they provide powerful inroads into the social brain as they evoke emotions, activate memories, draw and direct attention, and elicit communicative reactions. Faces trigger these processes in an automatic fashion, suggesting that just as the perceptual analysis of faces is supported by specialized hardware, these diverse cognitive functions may be as well. In his talk, Dr. Freiwald will describe recent experimental and theoretical advances towards understanding how the face-processing system encodes, transforms, and packages facial information, and how the face-processing network is embedded into the social brain in ways to suggest specific pathways for social information processing and a deep evolutionary heritage of high-level social cognition in humans.

Dr. Ed Connor

04/25/2016 4:00pm
Howard Egeth, Ph.D.
Prof. Psychological & Brain Sciences

“The Dark Side of Attention: Ignoring and Learning to Ignore”

Current research in my lab is focused on how we manage to attend to relevant information and ignore irrelevant information even when it is highly salient. We are exploring the role of learning and are using both behavioral and neuroimaging techniques (especially EEG/ERP). Some of our recent work suggests that inhibition of irrelevant material, rather than enhancement of relevant material is what drives attentional selectivity.

Dr. Rudiger von der Heydt

05/02/2016 4:00pm
Jill Leutgeb, Ph.D.
Assistant Prof of Neuroethology
University of California, San Diego

“Space, Time, and Memory Coding in the Hippocampus”

Even though there is consensus that the medial temporal lobe is required for episodic memories, it remains poorly understood how neuronal computations within its circuitry contribute to memory encoding and retrieval. Dr. Jill Leutgeb’s research is directed towards understanding the neuronal mechanisms of long-term memory storage at the systems level. Her lab seeks to determine how the various subregions of the medial temporal lobe, including the hippocampus and entorhinal cortex, contribute to memory formation. In particular, it is well established that these brain regions are specialized for spatial processing and navigation, and Dr. Leutgeb’s laboratory has described how internal representations of space are combined with information about context and time. For example, she will discuss how neuronal activity patterns in hippocampal subregions, in particular in the long overlooked CA2 region, can continuously change while also preserving information that resembles past events. Furthermore, she will describe how the entorhinal cortex jointly represents spatial and nonspatial information before these processing streams converge in the hippocampus. These studies inform our understanding of how hippocampal networks distinctly code for space, time, and context.

Dr. James Knierim

05/09/2016 4:00pm
Pawan Sinha, Ph.D.
Professor Vision & Computational
Massachusetts Institute Technology

"Learning to see late in childhood"

The hope inherent in pursuing basic research is that sometime in the future the work will prove beneficial to society. This fruition can often take many years. However, in some instances, even the conduct of basic research can yield tangible societal benefits. I shall describe an effort that perhaps fits in this category. Named 'Project Prakash', this initiative provides sight to blind children on the one hand and helps address questions regarding brain plasticity and learning on the other. Through a combination of behavioral and brain-imaging studies, the effort has provided evidence of visual learning late in childhood and has illuminated some of the processes that might underlie such learning.

Dr. Rudiger von der Heydt

05/23/2016 4:00pm
Randy Flanagan, Ph.D.
Department of Psychology
Queen's University, Ontario, Canada

“Representing objects and actions when interacting with the world”

Most manual tasks involve grasping and manipulating objects. In the real world, such tasks often involve choosing among several candidate objects to act upon, and then generating motor commands tailored to the properties of the selected object. In the first part of this talk, I will discuss evidence for the provocative idea that, when confronted with multiple potential reach targets, the sensorimotor system prepares well-specified actions for each target, prior to selecting between then. I will suggest that doing so may enable the motor system to ‘co-optimize’ competing action plans. In the second part of the talk I will discuss how the sensorimotor system represents object properties including weight, the accurate prediction of which is essential for skilled and dexterous manipulation. In addition to evidence from psychophysical experiments, I will describe recent functional neuroimaging work examining the contribution of object-sensitive ventral visual stream areas during the planning of manipulation tasks.

Dr. Ed Connor

Style DIV, please skip.

Style DIV, please skip.