*********************************
There is now a CONTENT FREEZE for Mercury while we switch to a new platform. It began on Friday, March 10 at 6pm and will end on Wednesday, March 15 at noon. No new content can be created during this time, but all material in the system as of the beginning of the freeze will be migrated to the new platform, including users and groups. Functionally the new site is identical to the old one. webteam@gatech.edu
*********************************
Amrit Kashyap
BME PhD Proposal Presentation
Date: 11/18/2019
Time: 2 PM
Location: HSRB E160
Committee Members:
Shella Keilholz, PhD (advisor)
Madeleine Hackney, PhD (co-advisor)
Lena Ting, PhD
Christopher Rozell, PhD
Bruce Crosson, PhD
Chethan Pandarinath, PhD
Title: Understanding the Neural Basis of Whole Brain Activity through Biological Circuitry and Resource Constraints
Abstract:
Large scale cortical architecture has not only evolved in order to fulfil organisms’ behavioral functions but also to maximize circuit performance under physical resource constraints. Neural circuitry is expensive with most of ATP consumption going directly into maintaining the action potential. The locations of the neurons in the smallest known nervous system, C. Elegans, have been shown to roughly minimize total wiring costs and in humans most axons are short and run to adjacent gyri. While most of these resource constraints problems have developed to explain static circuitry and the spatial localization of neural processing, they can also be used to understand how the brain functionally operates in real time. Large scale patterns of whole brain activity seen in fMRI are thought to arise from neural populations interacting with each other through the structural fiber network and mediated by task input. The structural network, therefore is a constraint that aids understand which regions coordinate together, and in turn the dynamics are reproduced by a network that utilizes the fewest possible fiber connections. I propose using one modality as a constraint in order to infer information about the other modality, both inferring the structural network from measured functional activation and using structural network as a constraint in a Machine Learning model to fit a dynamical system to the observed functional data. Since the central nervous system is optimized for the utilization of resources, these constraints improve our understanding of the relationship between large structural circuitry and how they are used by the organism for behavior. I then aim to translate this technique to address clinically relevant questions, where our constraints can aid as a prior to help understand gaps in knowledge on how changes in structural connectivity translate into observed deviations in functional activity.
