UC San Diego

This dissertation contains several projects, each providing a solution contributing to an aspect of group- level source-level EEG analysis. I explore different methods to extract better EEG measures from individual subjects: regression to reduce confounds originated from temporal proximity of cognitive events, optimal low-pass filtering to calculate better ERPs and collaborative averaging to obtain better measures from small numbers of trials. I introduce two methods for combining source-based EEG information, calculated with ICA and equivalent dipole localization, across subjects in a study: Measure Projection Analysis (MPA) allows study-level analysis for measures, such as ERP and ERSP, that are associated with single brain areas while Network Projection Analysis enables combining network measures, such as effective connectivity, associated with an ordered pair of brain area. The last two chapters of the dissertation are dedicated to discussing meta-analysis, i.e. combing information across multiple studies. This is a subject that is well developed in the fMRI field but is new in the field of source-based EEG analysis. I introduce a user- friendly schema (Hierarchical Event Descriptors, or HED), based on established cognitive ontologies, to describe cognitive event and states in a hierarchical and machine readable manner. HED facilitates automated meta-analysis and can benefit researchers by simplifying statistical designs and streamlining event information handling. The current EEG analysis-publication workflow mostly documents qualitative descriptions of event-related EEG dynamics. This makes it difficult to look for comparable results in the literature since search options are limited to textual descriptions and/or similar-appearing results depicted in the paper figures. In the final chapter I demonstrate a method for quantitative comparison of source-resolved results (e.g., ERPs, ERSPs) across different EEG studies. The proposed source-resolved EEG measure search engine receives search queries composed of event-related EEG measures, each associated with an estimated brain source location to be compared using Measure Projection Analysis (MPA) to all records in the search engine database accumulated by automated data analysis workflows applied to data of multiple studies. A similarity-ranked list of events from other studies that have elicited similar EEG dynamics in nearby source-locations is then returned to the user along with their experiment and event metadata