Assessing the Weighted Adaptive Filtering to Attenuate Eye-Blink Artefact by Means of Simulation for Brain-Computer Interface Application

In Brain-Computer Interface (BCI) systems, the signal acquisition of brain’s electrical activity, when non-invasive, is usually made with the electroencephalogram (EEG). EEG signals are naturally contaminated by artefacts which can significantly distort signals, altering neurological events, therefore compromising BCI control. In order to be applied in BCI systems, the method of artefact attenuation should be automatic, online, and ideally performed with few EEG recording channels. A previous work proposed a procedure for eye-blink artefact reduction based on adaptive filtering when electrooculogram data is not available. This method satisfies all the essential conditions for application in BCI systems and also addresses the bidirectional interference issue. In order to apply this technique in BCI systems, the present work aims to proceed with its assessment through changes in simulation in order to make the reproduced environment more realistic and therefore prove the reliability and effectiveness of the method. Results show that satisfactory artefact reduction is achieved even when its time occurrence overlaps the desired ERP (event-related potential). The lowest overall RMS error was achieved using a 2nd-order filter and adaptation factor set in 10–5. Furthermore, weight thresholds have a slight influence on filter performance when using plausible values. In conclusion, the proposed approach resulted in a considerable reduction of the eye-blink artefact, preserving ERP morphology and allowing easy component identification, all this in an online and automatic fashion, with few EEG recording channels and without the need for the reference channel, hence being a great choice to attenuate eye-blink artefacts in BCI systems.