Automatic 3D Video Analysis of Upper and Lower Body Movements to Identify Isolated REM Sleep Behavior Disorder: A Pilot Study

Abstract

Rapid eye movement (REM) sleep behavior disorder (RBD) is a parasomnia characterized by dream enactment, abnormal jerks and movements during REM sleep. Isolated RBD (iRBD) is recognized as the early stage of alphasynucleinopathies, i.e. dementia with Lewy bodies, Parkinson's disease and multiple system atrophy. The certain diagnosis of iRBD requires video-polysomnography, evaluated by experts with time-consuming visual analyses. In this study, we propose automatic analysis of movements detected with 3D contactless video as a promising technology to assist sleep experts in the identification of patients with iRBD. By using automatically detected upper and lower body movements occurring during REM sleep with a duration between 4s and 5s, we could discriminate 20 iRBD patients from 24 patients with sleepdisordered breathing with an accuracy of 0.91 and F1-score of 0.90. This pilot study shows that 3D contactless video can be successfully used as a non-invasive technology to assist clinicians in identifying abnormal movements during REM sleep, and therefore to recognize patients with iRBD. Future investigations in larger cohorts are needed to validate the proposed technology and methodology.

Improved automatic identification of isolated rapid eye movement sleep behavior disorder with a 3D time‐of‐flight camera

“…In particular, specificity and positive predictive value for iRBD identification increased when also including movements identified in the head, hands and upper body. This is in line with our previous preliminary findings [13], as well as with previous literature showing improved identification of people with RBD when recording muscle activity in the upper extremities [15–17] and with the increased frequency of neck myoclonus (i.e., head jerks) in subjects with RBD [18]. However, when considering only movements in the head, hands and upper body (thus excluding the lower body), people with iRBD were differentiated from the no‐RBD group with significantly lower accuracy and F1 score than when movements in the lower body only were considered.…”

“…In this study, the aim was to improve and advance the previously proposed automatic analysis of 3D videos recorded with a TOF sensor for identification of people with iRBD [ 12 , 13 ]. Compared to the previous approach, movements were separately identified in multiple body parts, the methodology was validated on unseen subjects in a large cohort and the performances were assessed with a 10‐fold cross‐validation approach.…”

“…In particular, automatic video analysis would be highly beneficial, as abnormal movement activity is an intuitively understandable biomarker of disease. Whilst for RWA quantification several automatic methods have been proposed [ 6 , 7 , 8 , 9 , 10 , 11 ], only two previous studies from our group have investigated automatic video analysis to identify RBD [ 12 , 13 ]. In both studies, videos were recorded with a 3D time‐of‐flight (TOF) sensor, allowing movements to be identified and quantified with high precision.…”

“…The first study showed that movements in the lower limbs could differentiate 40 people with iRBD from 64 people with other disorders with motor activity during sleep with 90.4% accuracy [ 12 ]. In the second pilot study, the combination of movements in the upper body (i.e., above the hips) and lower limbs differentiated 20 people with iRBD from 24 people with sleep‐related breathing disorders with increased accuracy compared to the analysis of movements in the lower limbs only [ 13 ]. Both these studies analyzed only movements included in manually identified REM sleep and the body regions were manually identified.…”

New Perspectives in Nonintrusive Sleep Monitoring for Neurodegenerative Diseases—A Narrative Review

“…Given its discovered importance in synucleinopathies, interest grew around prodromal RBD, also considering the difficulty in distinguishing it from mimics, i.e., other motor manifestations or parasomnias during sleep. An attempt at simplification was provided by Cesari and Waser in [109,110], respectively, which exploited 3D video analysis to evaluate limbs movements. They used custom algorithms to identify limb movements.…”