Artificial intelligence meets REM sleep behavior disorder: a new promising horizon

Also available on Spotify

RBD is a parasomnia characterised by abnormal muscle activity and dream enactment during REM sleep. The isolated form of the disorder (iRBD) is now widely recognised as an early stage of alpha-synuclein-related neurodegenerative diseases, such as Parkinson’s disease. Studies demonstrate that 80–90% of individuals diagnosed with iRBD develop neurodegenerative symptoms within 10-15 years. As such, this population represents an ideal cohort for clinical trials aimed at testing disease-modifying therapies.

Recent advances suggest that AI could play a transformative role in several key areas of RBD research and clinical management:

• Diagnosis: Current diagnostic guidelines rely on labour-intensive manual and visual analysis of polysomnographic data to detect abnormal muscle tone and behaviours during REM sleep. AI offers a promising alternative by enabling faster, automated, and objective assessment, thereby supporting clinicians in making more efficient and accurate diagnoses.
• Screening: At present, RBD diagnosis is limited to specialised neurological sleep laboratories, contributing to widespread underdiagnosis. However, the integration of AI with emerging wearable technologies could enable large-scale screening in the general population. This capability will become increasingly important as disease-modifying treatments become available, highlighting the need for early detection.
• Prognosis: AI is also being explored for its ability to detect complex patterns across sleep recordings, brain imaging, and biomarker data. These insights may help predict the timeline for progression from iRBD to overt alpha-synucleinopathies. In turn, this predictive power could allow for more personalised treatment planning and early intervention strategies.

A number of research groups are actively working in these areas, aiming to translate technological innovation into clinical utility. For instance, Dr. Matteo Cesari and his team recently demonstrated the effectiveness of a contactless 3D time-of-flight camera capable of automatically detecting nocturnal movements and accurately identifying patients with RBD. This work is now being expanded through a European Union-funded project. The role of automatic sleep stage scoring with hypnodensities for automatic detection of patients with iRBD also showed to be promising. Additionally, they have shown that AI-based analysis of sleep brain waves informs on progression and can successfully predict the time to conversion from iRBD to overt neurodegenerative disease, further supporting AI’s potential in precision medicine. Dr. Matteo Cesari has also recently reviewed the role of AI for screening of patients with iRBD.

In conclusion, advancements in AI hold great promise for transforming RBD research and clinical care. Through collaborative efforts across institutions and disciplines, the field is well-positioned to develop innovative tools that enhance diagnosis, enable large-scale screening, and improve prognostic accuracy. These developments have the potential to significantly impact clinical practice and patient care.

In the interview, Dr. Cesari also shares his advice for early-career professionals who want to work with AI in sleep.

Meet the contributors

Dr. Matias Rusamen

Postdoctoral Researcher at the Department of Technical Physics at the University of Eastern Finland, working in the Sleep Technology and Analytics research group. Under ESRS, he is a member of the Early Career Network (ECN).

Links to social media/ websites: LinkedIn

Speakers

Dr. Matteo Cesari is a biomedical engineer and senior scientist at the Department of Neurology, Medical University of Innsbruck, Austria. His work focuses on AI and advanced sleep data analysis.

Links to Papers

Cesari M, Brink-Kjaer A, During E, et al.(2025) Isolated REM Sleep Behaviour Disorder-Is Screening Possible?, J Sleep Res.

Picard-Deland C, Cesari M, Stefani A, Maranci JB, Hogl B, Arnulf I.(2025) The Future of Parasomnias. J Sleep Res.

Feuerstein S, Stefani A, Angerbauer R, et al. (2024) Sleep structure discriminates patients with isolated REM sleep behavior disorder: a deep learning approach. Annu Int Conf IEEE Eng Med Biol Soc.

Cesari M, Portscher A, Stefani A, et al.(2024), Machine Learning Predicts Phenoconversion from Poly-somnography in Isolated REM Sleep Behavior Disorder. Brain Sci.

Angerbauer R, Stefani A, Zitser J, et al. (2025) Temporal progression of sleep electroencephalography features in isolated rapid eye movement sleep behaviour disorder. J Sleep Res.

Cesari M, Ruzicka L, Högl B, et al. (2023) Improved automatic identification of isolated rapid eye movement sleep behavior disorder with a 3D time-of-flight camera. Eur J Neurol.