Abstract

Title: Estimating across-site PAC , delays, and complex network properties under source mixing

Abstract: Magneto- and electroencephalography (M/EEG) offer high temporal resolution, making it possible to study directed and frequency-resolved interactions including different types of cross-frequency coupling. However, inevitable source mixing can cause spurious FC if not properly accounted for. In this talk, I will consider the problems of detecting and quantifying Granger causality, phase-amplitude coupling, signal transmission delays and complex network properties from M/EEG data in the presence of source mixing. I will present methods and simulation studies highlighting the capabilities and limitations of M/EEG based FC estimation. Finally, I will show normative functional M/EEG connectomes derived from large pediatric and adult cohorts, including CamCAN, and their dependance on age and gender. These normative connectomes are expected to become instrumental for characterizing pathological brain communication patterns.

Bio: Stefan Haufe is a joint Associate Professor of Uncertainty, Inverse Modeling and Machine Learning at Technische Universität Berlin and Physikalisch-Technische Bundesanstalt Berlin as well as a Group Leader at Charité – Universitätsmedizin Berlin. His group develops signal processing, inverse modelling and machine learning methods primarily for analysing neuroimaging data. He is also interested in model interpretation and “explainable artificial intelligence”.

Venue: MRC CBU West Wing Seminar Room and Zoom https://us02web.zoom.us/j/82385113580?pwd=RmxIUmphQW9Ud1JBby9nTDQzR0NRdz09