Elias B. Issa - US grants

DESCRIPTION (provided by applicant): The brain's remarkable ability to recognize and remember objects depends on a series of visual processing stages culminating in inferotemporal cortex (IT). As the highest purely visual processing stage, IT contains dedicated domains for processing specific object classes, such as faces. Little is known, however, about the detailed organization and properties of neurons within these domains. The goal of this proposal is to provide such information by precisely targeting neural recordings to face and object selective regions of IT. An integrated system will be developed that uses fMRI to localize regions of interest and a novel x-ray based technique to return neural coordinates within these regions. Our first aim will be to measure the selectivity and tolerance properties of neurons throughout the 3D extent of an fMRI defined object patch. Different hypotheses about fine scale organization will be tested, and the underlying selectivity of neural populations compared to selectivity measured by fMRI. The second aim of our proposal will compare the organization in a face patch with a patch for novel objects. Face selective cells are already known to exist in face patches, but the question to be addressed is how the number, organization, or strength of tuning of cells in face patches differs from that in regions selective to unfamiliar objects. Such information will provide insights into how IT is organized across objects and whether strong familiarity with objects like faces refines neural maps. By bridging the gap between human fMRI and underlying neural structures, this work directly impacts our understanding of human object selective cortex. In case of loss of visual function, object level domains at the interface of memory and perception are prime targets for visual neural prosthetics. Knowledge of their precise locations and structures will guide repair and lay groundwork for the development of next-generation prosthetics.

DESCRIPTION (provided by applicant): Face recognition is an important part of everyday human social behavior, yet the neural circuits underlying face processing are only beginning to be understood. In particular, it remains unclear how feedback recurrently modifies face representations at the neural level despite evidence that face processing is a dynamic process. Here, I will study the role of feedback in the face patch system in primate inferotemporal cortex (IT). In the mentored phase of the project (Aim 1), I will probe the dynamics of neural responses across the face pathway under challenging image conditions (noise, occlusion) designed to maximally engage recurrent processing. Preliminary data that I have collected show that neural responses in the middle face patch evolve over time to resolve ambiguous information in a manner that correlates with activity in the anterior patches. The next logical step, which is goal of the second half of the mentored phase (Aim 2), is to measure how the anterior face patches causally influence activity in the middle face patch, and pharmacological inactivation of anterior regions will be used to systematically map the locus of feedback. In the independent phase of the project (Aim 3), I will use newly developed optogenetic silencing tools to test how rapidly and precisely projections arising from within the feedback locus can modulate middle face patch activity. An innovative feature of this project is the use of a novel stereo microfocal x-ray syste for registering the precise spatial position of all feedback sources and targets allowing for reconstruction of large scale (Aim 2) and fine scale (Aim 3) maps of source-target interactions. This endeavor will eventually yield an engineer's version of a neural circuit diagram and constrain computational models proposing different roles (gain control, resonance, and prediction) for feedback. The proposed work will be initiated in the laboratory of Jim DiCarlo at MIT. Dr. DiCarlo is a committed mentor that leads a highly collaborative, interdisciplinary research group. During the mentored phase of the project (K99; Aims 1 and 2), I will acquire training in techniques for population neural recording, pharmacological inactivation, and optogenetics providing an experimental platform for transition to an independent position in the second phase (R00; Aim 3) of the project. PUBLIC HEALTH RELEVANCE: How the collective action of distributed neural systems leads to our rich percept of the visual world is not well understood, and disorders of the circuits involved in perception, especially those involved in face recognition, can impair normal social function. Understanding the neural circuitry underlying face processing will provide valuable insights into how humans see, will improve next generation brain prosthetics for restoring visual function, and will inspire artificial vision systems.