Application of Independent Component Analysis for Rejection of Motion Artefact in BSPM Recorded During Exercise

Acquisitions of body surface potentials require in some studies that, patients perform exercise on cycle ergometer to increase their heart rate. Under this condition recordings are highly affected by the muscular activity derived from the pedalling (motion artefact). To reject this interference we propose the use of independent component analysis (ICA) by applying two different algorithms (FastICA and temporal decorrelation). Datasets from 13 subjects were used. For quantification of the results, the signal-to-noise ratios were compared after suppressing the motion artefact. ECG signals showed an average improvement of about 4 to10 dB.