Robust k-means-based Clustering of Independent Components Estimated from the EEG data.

In this study, we present a comprehensive investigation into the robust clustering of independent components (ICs) related to anticipatory postural control tasks in individuals with traumatic brain injury (TBI) using EEG data. Given the significance of accurately clustering the neural sources, our research evaluates the performance of various k-means clustering algorithms, including traditional, our modified approach of repeated k-means, and global k-means. This study aims to identify the optimal clustering approach that accurately locates the cortical sources germane to balance dysfunction and is computationally efficient. Our results highlight the superior performance of the global k-means algorithm in terms of clustering quality and computational runtime, demonstrating its application in a real-world dataset with noise.