Paired electrodes- and contraint independent components analysis-based denoising to alleivate motion artifacts in electroencephalogram collected at construction sites

Abstract

Despite the potential of mobile electroencephalogram (EEG) devices in managing construction workers' safety, health, and productivity, deploying mobile EEG at sites is hindered by significant motion artifacts introduced by worker movements. To address this issue, the authors propose a paired electrode-and constraint independent component analysis-based denoising that adaptively suppresses EEG motion artifacts via leveraging simultaneously collected motion artifact references. The proposed denoising was compared with an advanced benchmark on an EEG dataset collected under real human motions. Results show the proposed technique's denoising performance is statistically higher than the existing advanced benchmark. The finding of this study can improve the applicability of mobile EEG to construction sites, thereby significantly contributing to management of workers' safety, health, and productivity.