Title: Robust and Gaussian spatial functional regression models for analysis of event-related potentials
Authors: Jeff Morris - MD Anderson Cancer Center (United States) [presenting]
Hongxiao Zhu - Virginia Tech (United States)
Philip Rausch - Humboldt University (Germany)
Abstract: Event-related potentials (ERPs) are times series with both spatial correlation across electrodes and nested correlations within subjects. Commonly used analytical methods focus on pre-determined extracted components and ignore the correlation among electrodes or subjects, which can miss important insights, and tend to be sensitive to outlying subjects, time points or electrodes. We introduce a Bayesian spatial functional regression framework that models the entire ERPs as spatially correlated functional responses and stimulus types as covariates, relying on mixed models to characterize stimuli effects while accounting for the multilevel correlation structure, including both Gaussian and more robust models using heavier-tailed likelihoods. The spatial correlation is captured through basis-space Materns that are separable or nonseparable over time. We induce both adaptive regularization over time and spatial smoothness across electrodes via a correlated normal-exponential-gamma prior. Our proposed analysis produces global tests for stimuli effects across entire time (or time-frequency) and electrode domains, plus multiplicity-adjusted pointwise inference based on EER or FDR to flag spatiotemporal (or spatio-temporal-frequency) regions that characterize stimuli differences, and can also produce inference for any prespecified waveform components. Our analysis of the smoking cessation ERP data set reveals numerous effects across different types of visual stimuli.