Smartphone motor tracking detects early Parkinson’s signals in REM sleep behavior disorder patients
A smartphone app that measures subtle changes in finger tapping, hand movements, and speech has proven sensitive enough to distinguish people with idiopathic REM sleep behavior disorder (iRBD) who are about to develop Parkinson’s disease from those who remain stable, according to a multicenter study published June 27 in NPJ Parkinson’s Disease. The digital bradykinesia measure showed a large effect size (Cohen’s d = 1.10) in separating future converters from non-converters at baseline alone, before any clinical diagnosis could be made. The findings suggest that inexpensive, widely available smartphone sensors could replace costly clinic-based testing in clinical trials for prodromal Parkinson’s, lowering the barrier to testing new treatments designed to slow or stop the disease before classic motor symptoms appear.
What they found
The study enrolled 162 participants across Canadian and Swiss centers: 51 polysomnography-confirmed iRBD patients, 89 patients with early Parkinson’s disease, and 22 healthy controls. All used the Roche PD Mobile Application v2 on a study-provided smartphone, completing brief active tasks (finger tapping, walking, voice recordings) each day for 12 months. Key results included:
- Adherence was high. Participants completed active tasks on 73% of study days over the full year, demonstrating that daily smartphone monitoring is feasible even in older adults with iRBD (mean age around 67 years).
- Baseline digital bradykinesia scores discriminated all three groups. The smartphone-derived bradykinesia measure distinguished iRBD patients from healthy controls and from early PD patients at the first assessment, with p < 0.001.
- The measure distinguished converters from non-converters. Among iRBD patients, those who phenoconverted to a synucleinopathy during follow-up had significantly higher (worse) baseline digital bradykinesia scores than those who did not convert (p = 0.003, Cohen’s d = 1.10, a large effect size).
- Longitudinal worsening was detectable over less than a year. Over approximately 50 weeks of follow-up, digital bradykinesia worsened with a Cohen’s d of 0.50, and digital speech measures worsened with a Cohen’s d of 0.79.
- Sample size estimates favor digital measures. To detect a 50% treatment effect in a prodromal PD trial, 132 participants per arm would be needed using the digital bradykinesia endpoint. This is substantially fewer than would be required using standard clinical rating scales, which typically need several hundred participants per arm.
Why it matters
Idiopathic REM sleep behavior disorder is the strongest known clinical predictor of Parkinson’s disease and related synucleinopathies. Most iRBD patients will eventually convert, but the timeline is unpredictable: some convert within months, others remain stable for decades. Clinical trials testing potentially neuroprotective agents in this population have been hampered by the need for large sample sizes and long follow-up periods using conventional clinical scales that were designed for manifest Parkinson’s rather than prodromal stages.
Smartphone-based digital measures offer a practical alternative. If a five-minute daily task on a phone can track progression as sensitively as the data suggest, trial designers could run smaller, shorter, and far less expensive studies. For drug developers, this could mean the difference between a feasible phase 2 trial and an unfundable one. For patients living with iRBD, it could mean faster access to therapies aimed at preventing or delaying the motor and cognitive decline of Parkinson’s disease.
The study also demonstrates that digital measures capture a motor signal that is detectable even before a clinical diagnosis of Parkinson’s is warranted. The fact that baseline bradykinesia scores already separated future converters from non-converters with a large effect size (d = 1.10) suggests the window for intervention may open earlier than previously measurable. This is especially relevant given that iRBD carries a remarkably high conversion rate: longitudinal studies estimate that more than 70% of individuals with polysomnography-confirmed iRBD will develop a synucleinopathy within a decade. With such a high pretest risk, any tool that refines the timeline and tracks the trajectory of early motor decline could transform both clinical trial design and, eventually, clinical counseling.
Limits
The study is limited by its moderate sample size (51 iRBD patients, with only a subset converting during follow-up), and the 12-month observation period captures only early conversion events. The smartphone platform was provided to participants, which may not reflect real-world adherence if patients use their own devices. The study was funded by an unrestricted grant from Roche to R. Postuma, and several authors are Roche employees, representing a potential conflict of interest.
Bottom line
Daily smartphone-based digital motor assessments, particularly bradykinesia measures, can track progression in iRBD with effect sizes and sample size requirements that outperform standard clinical rating scales. If replicated in larger trials with longer follow-up, these measures could become primary endpoints in prodromal Parkinson’s disease clinical trials, accelerating the search for neuroprotective therapies.
Source
Bouhadoun S, Poulin H, Pelletier A, et al. Smartphone-derived digital motor measures to monitor progression in idiopathic REM sleep behavior disorder. NPJ Parkinson’s Disease. 2026 June 27. DOI: 10.1038/s41531-026-01440-6. PMID: 42362553.