Applying Support Vector Regression Analysis on Grip Force Level-Related Corticomuscular Coherence

Yao Rong,Xixuan Han,Dongmei Hao,Liu Cao,Qing Wang,Mingai Li,Lijuan Duan,Yanjun Zeng
DOI: https://doi.org/10.1007/s10827-014-0501-0
IF: 1.4529
2014-01-01
Journal of Computational Neuroscience
Abstract:Voluntary motor performance is the result of cortical commands driving muscle actions. Corticomuscular coherence can be used to examine the functional coupling or communication between human brain and muscles. To investigate the effects of grip force level on corticomuscular coherence in an accessory muscle, this study proposed an expanded support vector regression (ESVR) algorithm to quantify the coherence between electroencephalogram (EEG) from sensorimotor cortex and surface electromyogram (EMG) from brachioradialis in upper limb. A measure called coherence proportion was introduced to compare the corticomuscular coherence in the alpha (7-15Hz), beta (15-30Hz) and gamma (30-45Hz) band at 25 % maximum grip force (MGF) and 75 % MGF. Results show that ESVR could reduce the influence of deflected signals and summarize the overall behavior of multiple coherence curves. Coherence proportion is more sensitive to grip force level than coherence area. The significantly higher corticomuscular coherence occurred in the alpha (p < 0.01) and beta band (p < 0.01) during 75 % MGF, but in the gamma band (p < 0.01) during 25 % MGF. The results suggest that sensorimotor cortex might control the activity of an accessory muscle for hand grip with increased grip intensity by changing functional corticomuscular coupling at certain frequency bands (alpha, beta and gamma bands).
What problem does this paper attempt to address?