Automatic detection of Periodic Limb Movements in Sleep (PLMS) disorder by TMS320C6713 Digital Signal Processing Card

The purpose of this study is to detect automatically Periodic Limb Movements in Sleep (PLMS) disorder by TMS320C6713 Digital Signal Processing Card. For this purpose, records of 19 subjects´ polisomnograph are collected and grouped by a specialist doctor as PLMS, OSAS (Obstructive Sleep Apnea Syndrome), PLMS OSAS and NONE. Subjects´ right and left Electromyographs (EMG) are decomposed from polisomnograph (PSG) records and divided into one hour pieces by using an auxiliary program. In Matlab-Simulink, Yule-AR Power-Spectra of each 0, 5 seconds of one hour EMGs are estimated. Leg movements are found by applying a threshold value to Power-Spectra Estimate maximums. Leg movements are scored according to the criteria of 0, 5 second relaxation, 0, 5–5 second movement, 5 second relaxation. As a result, for PLMS patients 19, 8/hour, for PLMS-OSAS patients 32,39/hour, for NONE subjects 11, 35/hour, for OSAS patients 10, 20/hour average leg movements are scored. The scoring algorithm is embedded in to TMS320C6713 card by Matlab-Simulink-Real Time Workshop. The programmed signal processing card was run online.