Estimation of Brain Dynamics Under Visuomotor Task using Functional Connectivity Analysis Based on Graph Theory

Abstract

Network studies of brain connectivity have demonstrated that the highly connected area, or hub, is a vital feature of human functional and structural brain organization. Hubs identify which region plays an important role in cognitive/sensorimotor tasks. In addition, a complex visuomotor learning skill causes specific changes of neuronal activation across brain regions. Accordingly, this study utilizes the hub as one of the features to map the visuomotor learning tasks and their dynamic functional connectivity (dFC). The electroencephalogram (EEG) data recorded under three different behavior conditions were investigated: motion only (MO), vision only (VO), and tracking (Tra) conditions. Here, we used the phase locking value (PLV) with a sliding window (50 ms) to calculate the dFC at four distinct frequency bands: 8-12 Hz (alpha), 18-22 Hz (low beta), 26-30 Hz (high beta) and 38-42 Hz (gamma), and the eigenvector centrality to evaluate the hub identification. The Gaussian Mixture Model (GMM) was applied to investigate the dFC patterns. The results showed that the dFC patterns with the hub feature represent the characteristic of neuronal activities under visuomotor coordination.