MIT Reports to the President 1998-99


The human brain is the most complex, sophisticated, and powerful information-processing device known. To study its complexities, the Department of Brain and Cognitive Sciences at the Massachusetts Institute of Technology combines the experimental technologies of neurobiology, neuroscience, and psychology, with the theoretical power of computational neuroscience and cognitive science.


Edward Adelson's group has devised a new set of visual illusions that serve as tools for exploring the operations of lightness perception. These experiments have exposed a set of mid-level mechanisms operating with "mid-level" representations, sensitive to statistical and configural aspects of an image. Dr. David Somers used fMRI to show that visual attention can modulate brain activity even in the first stages of visual processing in area VI. Previously it had been thought that attention's effects would only be observed at higher levels of processing.

Nancy Kanwisher's laboratory documented distinct cortical regions in the human brain involved in processing images of faces and places, and showed a complete correspondence between the neuronal activity in these regions and perceptual awareness of face versus place images.

Sebastian Seung has constructed a network model of the neural integrator of the oculomotor system based on the dynamics of intrinsic and synaptic conductances. The model maintains persistent patterns of neural activity that encode eye position in a manner consistent with single unit recordings from real integrator neurons.

Ann Graybiel's laboratory has discovered novel families of genes, some of which have preferential expression in the striatum, in a collaborative project with David Housman. Surprisingly, these novel genes link well-known second messenger pathways to Ras superfamily function. These findings are potentially of great significance both for neurobiology and for research on cancer. The second messenger pathways involved (cAMP, Ca2+, DAG) are critically involved in learning and memory mechanisms in the brain, and new evidence suggests that Ras superfamily function is as well.

Earl Miller's laboratory provided insight into the neural basis of cognition in the primate prefrontal cortex. One major discovery involved the role of the prefrontal cortex in acquiring the "rules" used to guide our behaviors. They found that as monkeys learn how certain visual cues instruct certain actions, information about the cues and actions become conjoined in neural activity. Virtually all complex behavior depends on this type of learning. They have also discovered that prefrontal neurons play a role in bringing a visual image "to mind", i.e., in recalling it from long term memory.

Matthew Wilson's laboratory has begun to identify the communication that goes on during sleep between brain regions that are involved in the formation of long-term memory. They discovered particular patterns of activity that might be involved in the process of transferring recent memory in the hippocampus into longer term stores within the neocortex. They have also recently opened a new window into the sleeping brain by identifying and reconstructing dream activity in the rodent hippocampus.

Elly Nedivi laboratory is using candidate-plasticity genes (CPGs), genes that are regulated in the brain by neuronal activity, as molecular probes to study developmental plasticity in visual and somatosensory cortex. One CPG that they have isolated and studied extensively was found to function as a cell surface molecule that promotes growth of dendritic arbors through an intercellular signaling mechanism. Recent experiments monitoring CPG15 expression and regulation during visual system development suggest that CPG15 synthesized by the presynaptic neuron functions extracellularly to promote growth of postsynaptic dendritic arbors.

Gerald Schneider's laboratory is developing methods for promoting regeneration of axons and their connections in the brain, where such regeneration is meager or non-existent in mature mammals. Recent work has shown that homographs of peripheral nerves can be used to partially restore vision in mice and hamsters blinded by transection of the optic nerve. To elicit a greater quantity of regeneration, and hence greater functional recovery, new gene therapies are being pursued. This is based on discoveries in tissue culture using transgenic mice.

Guosong Liu's laboratory has characterized postsynaptic differentiation during glutamatergic synapse formation. Understanding this process is critical for understanding synapse formation during brain development and plasticity. Using a novel biophysical approach, the laboratory determined the distribution and density of functional glutamate receptors before and after synaptogenesis, the time course of synaptic AMPA and NMDA receptor cluster formation, and the role of activity and glutamate receptor activation in the clustering of functional AMPA receptors during synapse formation.

Mriganka Sur's laboratory demonstrated that the activity-dependent development of projections from the retina to the visual thalamus is accompanied by a reduction in the number of axons that innervate thalamic neurons, but surprisingly not by changes in the strength of individual synapses. Thus, a process of normalization keeps the total synaptic input to single neurons relatively constant in the face of changing patterns of activity and inputs.

Ken Wexler and his group made a number of new discoveries in linguistic development and its biological basis. In normal development this included a new theory (the Unique Checking Constraint Theory) which integrates a large amount of data from a wide variety of languages, explaining differences in development across different languages. A number of new results in Specific Language Impairment were obtained, and a new direction was launched in the study of more complex syntax.

Reports in the past year from Edward Gibson's language comprehension lab include the finding that syntactic complexity is an important factor in ambiguity resolution, which is used in addition to word-level information and world knowledge. Furthermore, studies of reading comprehension in multiple languages (English, Spanish, Japanese, Brazilian Portuguese) show that similar syntactic complexity information is used in the processing of all languages.

Elizabeth Spelke, in collaboration with Stan Dehaene, explored two systems of representation that underlie adults' numerical reasoning. Combined behavioral and neuroimaging techniques provide evidence that reasoning about exact numerosities depends in part on a frontal brain system tied to language and verbal memory, whereas reasoning about approximate numerosities depends in part on a parietal brain system tied to spatial memory.


Two longstanding faculty members, Stephan Chorover and Alan Hein, retired after many years of service. Two new faculty members, Anthony Wagner and Pawan Sinha, were added as Assistant Professors. Nancy Kanwisher and Earl Miller were promoted to Associate Professor with tenure. Matthew Wilson was promoted to Associate Professor without tenure.


Ten new students in enrolled in the Fall 1998. One was awarded an NSF fellowship; two were funded by the new Rosenblith Fellowships. Other students are being supported by Howard Hughes Medical Institute Predoctoral Fellowships, the Merck/MIT Graduate Fellowship in Informatics, and a Poitras Predoctoral Fellowship. Of the eight students awarded Ph.D.s, five accepted postdoctoral positions at UC San Francisco, Tokyo University, NYU, the University of Pittsburgh, and the Max-Planck Institute, while two others accepted positions at Stanford University and Kinematix. Another is doing fieldwork in Guatemala. The prestigious Goodwin Medal for effective teaching was awarded to a fourth year graduate student. In addition, NIMH awarded the department another training grant, bringing our total to four.

The number of students who have chosen a major in the department again increased, as did the number of students minoring in psychology. The department also continues to draw increasingly large numbers of undergraduates into UROP positions for course credit and employment in research projects. In addition, the new course offerings in Brain and Cognitive Language Acquisition and Visual Cognition were very well received.


More information about the Deparment of Brain and Cognitive Sciences can be found on the World Wide Web at

Mriganka Sur

MIT Reports to the President 1998-99