Abstract

Speaker (visiting in person): Chetan Gohil, University of Oxford, UK.

Title: Leveraging large-scale datasets in boutique MEG studies

Abstract: Many MEG studies focus on specific clinical groups or carefully designed tasks and consequently are based on relatively small datasets. While these studies are well targeted, drawing robust conclusions can be difficult when sample sizes are limited. At the same time, large public MEG datasets are becoming increasingly available. Under the assumption that features (such as spatiotemporal patterns) of brain activity are shared across populations and tasks, these datasets can be used to support analyses in smaller, boutique studies. In this talk, I will discuss ways of leveraging large-scale MEG data (e.g. Cam-CAN) to improve inference in boutique studies, including normative modelling, dynamic network approaches such as the Hidden Markov Model, and newer deep learning approaches known as ‘foundation models’. I will highlight how these approaches can increase statistical power and robustness, and discuss both their potential and their limitations.

Venue: MRC CBU West Wing Seminar Room and Zoom https://us02web.zoom.us/j/82385113580?pwd=RmxIUmphQW9Ud1JBby9nTDQzR0NRdz09 (Meeting ID: 823 8511 3580; Passcode: 299077)