Removing Non-Invasive Brain Stimulation Artifacts via Joint Blind Source Separation

Transcranial Direct Current Stimulation (tDCS) is a non-invasive technique that modifies vestibular input by feeding the vestibular system electrical currents. Consequently, Electroencephalography (EEG) recordings may contain artifacts introduced by TDCS, making it difficult to interpret brain activity. Joint Blind Source Separation (JBSS) is a signal processing tool widely used to separate TDCS-induced artifacts from EEG data. In the present paper, We compare three JBSS methods, Multi-Canonical Correlation Analysis (MCCA), Independent vector analysis based on the multivariate gaussian model (IVA-G), and Independent component analysis by entropy rate bound minimization (ICA-ERBM), to remove tDCS artifacts from artifacted EEG signals. The results show that TDCS and clean EEG sources are distinguished better by the ICA-ERBM method.