REM Atonia Index Cutoffs for REM Sleep Without Atonia: A Systematic Review with Meta-analysis of Adult Diagnostic Performance.

Abstract

To synthesize evidence on REM Atonia Index (RAI) thresholds and reference values, quantify adult diagnostic performance for REM sleep behavior disorder (RBD), and distinguish diagnostic cutoffs from physiological reference limits and pediatric values. Methods: We conducted a focused systematic review of studies using automated RAI, rescreening studies from a previous review and updating PubMed/MEDLINE through July 27, 2026. Adult diagnostic accuracy was synthesized with a bivariate binomial-normal random-effects model. Continuous RAI differences were pooled as raw mean differences. Normative adult and pediatric evidence was synthesized separately because these represented different estimands. The review was retrospectively registered in OSF Registries on August 20, 2026 (https://osf.io/v3kqa). Thirty unique studies were included. Eight adult studies contributed to the diagnostic-accuracy meta-analysis; thresholds ranged from 0.80 to 0.90. Pooled sensitivity was 0.850 (95% CI, 0.766-0.908) and specificity 0.895 (95% CI, 0.814-0.944). Eight studies contributed continuous data; RAI was lower in RBD by a pooled mean difference of -0.263 (95% CI, -0.335 to -0.190). Only two methodologically non-equivalent adult studies provided physiological reference limits, 0.85 and 0.91, precluding pooling. Pediatric evidence supported age-conditioned interpretation. Adult chin-EMG RAI shows good overall discrimination of established RBD, but current evidence does not support a universal numerical boundary for REM sleep without atonia. Evidence defining physiological reference limits is substantially less developed than diagnostic evidence, and normal RAI varies with age. RAI should therefore be reported continuously, with thresholds interpreted according to their intended diagnostic or physiological purpose.