ICA-based ECG removal from Surface Electromyography and its effect on Low Back Pain Assessment

Abstract

Surface electromyography (SEMG) has been used for muscle function examination in neuromuscular disorders. The utility of SEMG in low back pain (LBP) assessment was questioned because of low sensitivity. Artifacts and noise contamination may distort the SEMG measurement in LBP assessment. The purposes of this study were to develop an ICA-based ECG removal method to obtain clean SEMG signal from back muscles, and to demonstrate the relative effect of ECG on back muscles SEMG parameters and their sensitivity on low back pain (LBP) assessment. This study compared surface EMG measurements on paraspinal muscles from 10 normal and 10 LBP patients during sitting and standing. The raw SEMG signal was processed by independent component analysis (ICA) to remove the ECG contamination. Then, median frequency (MF) of both raw and denoised paraspinal SEMG were calculated respectively. The MF of healthy and LBP groups before and after ECG removal were compared separately to evaluate the effect of ECG contamination. Also, difference between MF in subject with and without LBP were compared in raw and denoise condition to study the ECG effect on LBP assessment sensitivity. Significant MF increases (p<0.05) were founded after ECG noise removal in all tests. For LBP assessment, improvements in discriminative ability, in terms of parametric difference, were seen in MF parameter during sitting (mean difference between normal and patient increase from: Left: 8 to 45Hz; Right 11 to 53Hz) and standing (mean difference between normal and patient increase from: Left: -10 to 6Hz; Right 8 to 14Hz) respectively. ECG contaminations showed significantly influence on SEMG measurements in both normal and LBP patients. Our study has demonstrated the ability of the proposed ICA-based technique in ECG removal, which leads to an improvement in LBP assessment sensitivity