Multivariate Connectivity Analysis
PI Name: Dimitrios Pantazis
Institution: Massachusetts Institute of Technology
This project aims to develop, implement, and validate computational tools for the analysis of functional cortical networks with EEG/MEG. We propose a new approach to detecting functional interactions based on multivariate statistical inference. The main advantage of our method is the ability to automatically detect cortical networks over any combination of frequencies, unlike existing approaches that are typically limited to exploring relations such as coherence and phase synchrony within a single frequency band or between a known pair of frequencies. We will develop a method that automatically searches for functional relationships across amplitude and phase of several frequency bands, establish statistical tests that detect true activations while protecting from false positives and accounting for multiple comparisons, and validate our method on simulated data and a multi-subject audiovisual speech integration study.