抄録
REM sleep behavior disorder (RBD) precipitates neurodegenerative disease such as Parkinson's disease. We developed a new algorithm to quantify automatically the appearance ratio of tonic and phasic mentalis EMG activity (tonic% and phasic%) during REM sleep stage. We optimized the parameter values in the algorithm by collating the classification result from the presented algorithm with the result from visual inspection by neurology medical doctor. We used the PSG data from 24 Parkinson disease patients. The verification results showed good detection accuracy (tonic%: sensitivity=89%, specificity=80%, phasic%: sensitivity=88%, specificity=85%).
