Mixed Effects Spectral Vector Autoregressive Model: With Application to Brain Connectivity (2210.03017v1)

Abstract: The primary goal of this paper is to develop a method that quantifies how activity in one brain region can explain future activity in another region. Here, we propose the mixed effects spectral vector-autoregressive (ME-SpecVar) model to investigate differences in dynamics of dependence in a brain network between healthy children and those who are diagnosed with ADHD. Specifically, ME-SpecVar model will be used to formally test for significant connectivity structure obtained using filtered EEG signals in delta, theta, alpha, beta, and gamma frequency bands. The suggested model allows one-stage procedure for deriving Granger causality in common group structure and variation of subject specific random effects in different frequency oscillations. The model revealed novel results and showed more significant connections in all frequency bands and especially in slow waves in control group. In contrast, children with ADHD shared a pattern of diminished connectivity and variability of random effects. The results are consistent with previous findings about decreased anterior-posterior connectivity in children with ADHD. Moreover, the novel finding is that the most diverse subject specific effective connectivity parameters in healthy children belong to parietal-occipital region that is associated with conscious visual attention.