Understanding human vision through the lens of MEG multivariate pattern classification
Description
Multivariate pattern classification has grown rapidly in popularity in recent years, and has now become a central tool in the analysis of MEG signals. This novel methodological framework promises new perspectives and solutions in studying human vision, empowered by the ability of decoding algorithms to capture neuronal patterns. In this talk, I will offer novel insights in the duration and sequencing of visual cognitive processes obtained through multivariate analysis methods and MEG/fMRI data fusion. I will 1) characterize the time course of object recognition in the human brain; 2) demonstrate a clear dissociation between feedforward and feedback early visual processes, with well-defined temporal signatures for both mechanisms; and 3) link human vision with deep convolutional neural networks.
UPCOMING COG LUNCH TALKS:
3/6/18 - Leyla Isik, PhD. Postdoctoral Associate in the Kanwisher Lab
3/27/18 - Josh Rule, Graduate Student in the Tenenbaum Lab
4/3/18 - Paula Rubio-Fernandez, PhD. Visiting Scientiest in the Gibson Lab
4/10/18 - Eli Pollock, Graduate Student in the Jazayeri Lab
4/17/18 - Nicolas Meirhaeghe, Graduate Student in the Jazayeri Lab
4/24/18 - Rachel Magid, Graduate Student in the Schulz Lab
5/1/18 - Ilker Yildirim, Ph.D. Research Scientist in the Tenenbaum Lab
5/8/18 - Kelsey Allen, Graduate Student in the Tenenbaum Lab
5/15/18 - Tuan Le Mau, Graduate Student in the Brown Lab
Speaker Bio
Dimitrios Pantazis, who joined the McGovern Institute in 2010, oversees the operation of the Magnetoencephalography (MEG) Laboratory within the Martinos Imaging Center at MIT. Before moving to MIT, he was research assistant professor at the University of Southern California from 2008-2010. His research focuses on the development of novel MEG methods to holistically capture spatiotemporal brain activation and the study of visual brain representations. He has over 15 years of experience in developing methods for the analysis of MEG data and has published prominent articles in Nature Neuroscience, Proceedings of the National Academy of Sciences, Scientific Reports, Cerebral Cortex, and NeuroImage. His work has been featured in Science News; Scientific American Mind Magazine; Boston Magazine; MIT Technology Review; Elekta’s Wavelength Magazine; MIT News; the front-page of the MIT website and several Greek media outlets. He is a key developer of Brainstorm, an open-source environment dedicated to the analysis of brain recordings (MEG, EEG, NIRS, ECoG, depth electrodes, animal electrophysiology) with 13,000+ registered users and 400+ related publications. He is involved in numerous professional and outreach activities, including contributions to MIT and Harvard educational courses, and organization of tours to introduce imaging technology to local communities, high school science teachers, local biotechnology companies, and legislative and judicial representatives.