Parkinson’s AI Telerehabilitation Trial Reports No Significant Difference in Primary Motor Outcome
Published in Artificial Intelligence, Brain/Neurology, Journals, Tools.
Publication date: October 7, 2026.
A study published on October 7 in npj Digital Medicine tested different durations of AI-assisted telerehabilitation in early Parkinson’s disease. The exploratory multicenter trial randomized 120 participants to one, two or three months of training, with blinded outcome assessment.
Sessions lasted 15–30 minutes, five days per week. Only 71 participants were included in the final analysis. The primary outcome—change in MDS-UPDRS Part III motor score at three months—did not differ significantly between groups.
A post hoc analysis found a higher proportion meeting an exploratory cognitive response threshold in the three-month group than in the one-month group. This finding generates a hypothesis rather than proving cognitive benefit. No falls or adverse events were recorded during the trial.
High attrition, the absence of a usual-care control group and the negative primary endpoint limit interpretation. The results do not establish that AI-assisted rehabilitation improves outcomes compared with standard care. Larger trials with prespecified analyses are needed.
Source: Chen and colleagues, npj Digital Medicine, October 7, 2026. DOI: 10.1038/s41746-026-03339-1.