SEMINAR UNF SERIES
|Speaker:||Mehraveh Salehi, Ph. D. Candidate|
|Title:||Individualized and state-specific human brain parcellation in multiple scales|
|Where:||CRIUGM Room M6804 (http://www.criugm.qc.ca/en/contact.html)|
|When:||Thursday November 29th, 13h-14h|
*The seminar will be presented in English
Mehraveh Salehi is a Ph.D. candidate in Electrical Engineering department at Yale University. She is currently working as a research intern at Google DeepMind in Montreal. She earned her Bachelor degree in Electrical Engineering from Sharif University of Technology, Tehran, Iran. Her research lies at the intersection of statistical machine learning and computational neuroscience. She is interested in developing models that relate human behavior to individual brain connectivity patterns using optimization and machine learning techniques. She has received a number of awards including the Young Scientist Award from the International Conference on Medical Image Computing and Computer Assisted Intervention 2017 (MICCAI), and the best poster award from the BioImaging Sciences Retreat 2018. She is also the recipient of Tananbaum Fellowship, Advanced Graduate Leadership Program (AGLP) Fellowship, and CRA-Women Graduate Fellowship.
The goal of human brain mapping has long been to delineate functionally coherent regions in the brain and elucidate the functional role of these regions. Previous work has shown great success on defining functionally coherent regions at multiple scales, by grouping voxels into nodes and further grouping those nodes to form communities or networks in the brain. While majority of previous work has assumed fixed functional units across individuals and states, we show that the parcellation of human brain is both individual and state dependent. In this talk, I will first present a recently developed individualized and state-specific parcellation technique that utilizes submodular maximization of an exemplar-based utility function. Then, I will cover the predictive models that estimate biological and cognitive characteristics of individuals (including sex and IQ) as well as their brain’s cognitive state (approximated by the task condition as well as the within-condition task performance), solely based on the features extracted from these individualized parcellations.