Automating REM Sleep Without Atonia Scoring Methods

REM Behaviour Disorder (RBD) is a parasomnia characterized by the absence of the physiologic muscle atonia during REM stage. The diagnosis of this disorder is based on the assessment of muscle activity and REM sleep without atonia (RSWA) by means of an electromyogram during polysomnography. In recent years RBD has attracted the attention of researchers because in its isolated form it is linked to the subsequent onset of neurogenerative diseases such as Parkinson's disease, dementia with Lewy bodies and Multiple System Atrophy. This study focuses on the analysis of polysomnographic reports from three different databases containing also RBD and RSWA patient records. Two of them were provided by the Sleep Disorders Centre of A.O.U. Molinette in Turin, Italy; while the other is created from the CAP Sleep Database, made available in open access on PhysioNet by the Sleep the Disorders Center of the Ospedale Maggiore of Parma, Italy. The polysomnographies were analysed in order to evaluate the muscular activity during REM stage with the parameters most commonly used in the clinic and known in literature for the study of REM sleep without atonia. The RSWA scoring methods chosen are as follows: REM Atonia Index (automatic, submitted by Ferri in 2008 and 2010), Montréal (visual, Lapierre and Montplasir in 1992 and 2010) and SINBAR (visual, Barcelona and Innsbruck groups, from 2011 to 2013). An algorithm has been developed to calculate these methods as described in the literature. With regard to the visual methods, the algorithm developed is based on the criteria presented and translates the methods into automatic form. The RAI calculation method was developed and consolidated on the CAP database, and then applied to the others, while the automation of visual methods was developed in collaboration with the Sleep Disorders Centre of A.O.U. Molinette in Turin on the databases provided by them. This study outlines the results and the challenges of this automation process. One of the databases contains PSG recordings of subjects during the intake of an antidepressant drug and in its absence. The data obtained from the algorithm were used to compare the two conditions in terms of RSWA scoring.