Movement Classification Method for Awake Bruxism Using sEMG Signals

This paper presents a method to classify five brux-ism movements and three daily movements using surface elec-tromyographic ($s$EMG) signals obtained from the left and right masseter muscles, which are the muscles in the cheek. Forty features, eight window sizes for extracting features from $s$EMG signals, and eight classifiers were evaluated to design a suitable classification method. As a result, the highest classification accuracy of 96.57% was achieved by using a Support Vector Machine with the following eight features: MSR, LD, MYOP, SSC, MEAN, AAC, AR, and MMAV2, extracted with a window of 0.75 [$s$] at a sampling rate of 2000 [Hz] and an analog-digital conversion resolution of 10 [bit].