Validation of semiautomatic scoring of REM sleep without atonia in patients with RBD

Abstract

Objective/background: To evaluate REM sleep without atonia (RSWA) in REM sleep behavior disorder (RBD) several automatic algorithms have been developed. We aimed to validate our algorithm (Mayer et al., 2008) in order to assess the following: (1). capability of the algorithm to differentiate between RBD, night terror (NT), somnambulism (SW), Restless legs syndrome (RLS), and obstructive sleep apnea (OSA), (2). the cut-off values for short (SMI) and long muscle activity (LMI), (3). which muscles qualify best for differential diagnosis, and (4). the comparability of RSWA and registered movements between automatic and visual analysis of videometry.

Patients/methods: RSWA was automatically scored according to Mayer et al., 2008 in polysomnographies of 20 RBD, 10 SW/NT, 10 RLS and 10 OSA patients. Receiver operating characteristic (ROC) curves were used to determine the sensitivity and specificity of SMI and LMI. Independent samples were calculated with t-tests. Boxplots were used for group comparison. The comparison between motor events by manual scoring and automatic analysis were performed with "Visual Basic for Applications" (VBA) for every hundredth second.

Results: Our method discriminates RBD from SW/NT, OSA and RLS with a sensitivity of 72.5% and a specificity of 86.7%. Automatic scoring identifies more movements than visual video scoring. Mentalis muscle discriminates the sleep disorders best, followed by FDS, which was only recorded in SW/NT. Cut-off values for RSWA are comparable to those found by other groups.

Conclusion: The semi-automatic RSWA scoring method is capable to confirm RBD and to discriminate it with moderate sensitivity from other sleep disorders.

Keywords: REM sleep behavior disorder; REM sleep without atonia; Semiautomatic algorithm; Validation.