Title: Statistical methods for modeling heritability of EEG connectivity
Authors: Hernando Ombao - King Abdullah University of Science and Technology (KAUST) (Saudi Arabia) [presenting]
Abstract: A number of recent studies have found evidence that characteristics of functional brain connectivity are significantly associated with various genetic markers, however, the majority of work in this area has been restricted to resting state fMRI data. We develop novel measures of connectivity that capture complex dependence structures and present new models that quantify heritability in EEG connectivity. We present the results from a novel study of EEG spectral-based connectivity measures during a working memory task from 350 healthy university students. Using recently developed statistical methods for testing associations between high-dimensional feature sets (which improves upon existing statistical methods through the use of non-Euclidean metrics), we identify specific sets of channels for which the coherence measures in the delta, theta, alpha, and beta frequency bands are significantly associated with a set of genetic markers previously implicated as risk factors for Alzheimer's disease. Additionally, we compare these heritability estimates with genome-wide heritability and with estimates from a set of neurotransmitter genes related to dopamine regulation. These results suggest that some genetic factors linked to Alzheimer's disease may also play a role in working memory performance in healthy individuals.