StaR: An EEGLAB framework for the measure projection toolbox (MPT) statistical analyses to be performed in R

Abstract

EEGLAB, a widely used toolbox in MATLAB (The Mathworks, Inc.), uses Independent Component Analysis (ICA) to decompose the EEG signal into sub-signals, and localizes brain sources of those sub-signals prior to independent component (IC) clustering for group study. In 2013, the Measure Projection Toolbox (MPT) was introduced as a new data-driven IC clustering toolbox for EEGLAB. Despite the numerous features and advantages offered by EEGLAB and the MPT, they both have limitations for statistical analyses with more than two independent variables. In order to work around those limitations, this paper introduces StaR, an EEGLAB framework for the MPT statistical analyses to be performed in R. StaR initially exports the data from different clusters generated by the MPT for different measures of interest (e.g., Event-Related Potentials (ERPs) and Event-Related Spectral Perturbations (ERSPs)) and formats the data such that further statistical analyses can be performed in R. Once in R, StaR uses linear mixed models as its default method to better handle missing values and intra-subject variability. Finally, StaR brings the results back into MATLAB to plot the results with the well-known and easy to interpret EEGLAB graphics. To make the whole process easy, StaR also offers an intuitive user interface that integrates into EEGLAB’s menu.