Clusters of mu rhythm from EEG data: a comparative study between 61 and 19 channel datasets

Abstract

The EEG mu rhythm is a sensorimotor oscillation which is desynchronized by voluntary movement execution. Independent Component Analysis (ICA) allows the decomposition of recorded scalp EEG data into temporally, functionally, and spatially independent source signals. Clustering techniques applied to independent sources resolved with ICA have been proven to be successful in the identification of clusters of sensorimotor mu rhythm across different subjects. The present work deals with the issue regarding the minimum number of data channels that is recommended to find reliable clusters of mu rhythm. Left and right mu clusters were identified from high-density EEG recordings (61 channels) belonging to a publicly available EEG database. A second dataset was created by selecting a small subset of the same high-density EEG recordings. Specifically, only the 19 channels belonging to the standard 10-20 International System were used for the identification of left and right mu clusters. Quantitative parameters computed from mu clusters obtained from both the 61-channel and the 19-channel datasets were statistically compared. The obtained results suggest that clusters of mu rhythm in sensorimotor areas can be reliably found from a lower number of EEG channels compared to high-density electrodes configuration.