Actigraphy-based detection of isolated REM sleep behavior disorder: multicenter validation across devices and populations - Summary - MDSpire

Actigraphy-based detection of isolated REM sleep behavior disorder: multicenter validation across devices and populations

  • By

  • Li Zhou

  • Andreas Brink-Kjaer

  • Katarina Gunter

  • Salonee Marwaha

  • Ambra Stefani

  • Birgit Högl

  • Michele T. Hu

  • Emmanuel Mignot

  • Ankit Parekh

  • Qi Tang

  • Merve Aktan-Süzgün

  • Bei Huang

  • Shi Tang

  • Siyi Gong

  • Yuhua Yang

  • Xie Chen

  • Jianzhang Ni

  • Ningning Li

  • Zhixuan He

  • Yun Kwok Wing

  • Emmanuel During

  • October 29, 2025

  • 0 min

Share

Objective:

To assess the generalizability of a fully automated actigraphy-based model for detecting isolated REM sleep behavior disorder (iRBD) across different devices and populations, highlighting its potential to improve early diagnosis.

Key Findings:
  • The sleep model achieved AUCs of 0.838–0.865 across centers, indicating strong performance, while the RAR model ranged from 0.520–0.818, suggesting variability in effectiveness.
Interpretation:

The study demonstrates that the actigraphy-based detection model for iRBD generalizes well across different cohorts and devices, potentially enabling scalable screening for synucleinopathies, which is crucial for early intervention.

Limitations:
  • The study's design may not fully account for variations in demographic factors across different populations, which could affect the applicability of findings.
  • The reliance on self-reported prodrome symptoms may introduce bias, potentially impacting the accuracy of screening results.
Conclusion:

The findings support the use of actigraphy combined with prodromal symptoms for effective population-level screening of iRBD, emphasizing the need for scalable methods in clinical practice.

Original Source(s)

Actigraphy-based detection of isolated REM sleep behavior disorder: multicenter validation across devices and populations

  • by Li Zhou, Andreas Brink-Kjaer, Katarina Gunter, Salonee Marwaha, Ambra Stefani, Birgit Högl, Michele T. Hu, Emmanuel Mignot, Ankit Parekh, Qi Tang, Merve Aktan-Süzgün, Bei Huang, Shi Tang, Siyi Gong, Yuhua Yang, Xie Chen, Jianzhang Ni, Ningning Li, Zhixuan He, Yun Kwok Wing, Emmanuel During

  • October 29, 2025

Related Content

Electrophysiological signatures predict the therapeutic window of deep brain stimulation electrode contacts

Deep learning-enabled accurate assessment of gait impairments in Parkinson’s disease using smartphone videos

Clinical AI is Not (Yet) Trustworthy-But It Could Be