CFE 2019: Start Registration
View Submission - CMStatistics
Title: Covariate-adjusted hybrid principal components analysis for EEG data Authors:  Aaron Scheffler - University of California, San Francisco (United States) [presenting]
Abigail Dickinson - UCLA (United States)
Charlotte DiStefano - UCLA (United States)
Shafali Jeste - UCLA (United States)
Damla Senturk - University of California Los Angeles (United States)
Abstract: Electroencephalography (EEG) studies produce region-referenced functional data in the form of signals recorded across electrodes on the scalp. The data capture underlying neural dynamics, and it is of clinical interest to model differences in neurodevelopmental trajectories between diagnostic groups, e.g. typically developing (TD) children and children with autism spectrum disorder (ASD). Valid group level inference requires characterization of the EEG dependency structure and covariate-dependent heteroscedasticity, such as changes in variation over developmental age. Resting state EEG is collected on both TD and ASD children aged two to twelve years old. The peak alpha frequency (PAF) is an important biomarker linked to neurodevelopment. It is known to shift from lower to higher frequencies as children age. We model patterns of alpha spectral variation, rather than just the peak location, regionally across the scalp and chronologically across development for both the TD and ASD diagnostic groups. We propose a covariate-adjusted hybrid PCA (CA-HPCA) for region-referenced functional EEG data. It utilizes both vector and functional PCA while simultaneously adjusting for covariate-dependent heteroscedasticity. CA-HPCA assumes the covariance process is weakly separable conditional on observed covariates leading to stable and computationally efficient estimation. A mixed effects estimation framework is proposed coupled with a bootstrap test for group level inference.