Artifacts Removal from Low-Density EEG Measured with Dry Electrodes

Artifacts removal is an important preliminary step in electroencephalographic signals processing. This study particularly addresses the challenges of artifacts removal in a low-density scenario, i.e. when few acquisition channels are used. The ex-ploited hybrid method combines artifact subspace reconstruction (ASR) with multivariate empirical mode decomposition (MEMD). This recent technique was here tested, for the first time, on signals acquired by means of dry electrodes. Five types of artifacts were considered: eye closing, eye blinking, eye movement, clenching, and head flexing. In addition to the widely used visual inspection, the relative root mean square error (RRMSE) was used as a metric to assess the effectiveness. Ideally, the RRMSE should be 0 when comparing a segment of pure baseline EEG before and after artifact removal. The analysis was carried out for both MEMD-ASR and ASR alone. The results confirm that MEMD-ASR is more effective than ASR alone from one to four channels even in case of dry electrodes. These findings enhance the potential for practical, real-world applications of dry electrodes in EEG research.