Integrating AI in Stroke Neurorehabilitation

Official title

Integrating AI in Stroke Neurorehabilitation (AISN)

Overview

The AISN multicenter randomized controlled trial will assess the effectiveness of a novel artificial intelligence (AI)-based clinical decision-support system integrated into the Rehabilitation Gaming System (RGS) for home-based post-stroke rehabilitation. Approximately 192 participants ≥6 months post-stroke will be recruited across several European centers and assigned to one of three groups: RGS with AI decision support, RGS without AI, or standard care. The primary outcome is upper limb motor improvement for stroke patients, with secondary measures including cognitive function, independence, quality of life, usability, cost-effectiveness, and AI-based support performance.

Detailed description

The AISN study addresses the gap in long-term, personalized stroke rehabilitation after hospital discharge by evaluating an enhanced digital therapy platform that combines the clinically validated Rehabilitation Gaming System (RGS) with a newly developed AI-based decision-support module. This AI component analyzes patient performance data to provide clinicians with diagnostic and prognostic insights, along with tailored exercise prescriptions.

The trial's key innovation is the formal validation of the AI module in real-world clinical settings, assessing its concordance with clinician decisions, predictive accuracy, and contribution to patient outcomes.

Participants will be randomized into three groups:

RGS+AI: Home-based RGS therapy with AI-driven recommendations for clinicians. RGS-AI: Home-based RGS therapy without AI support. Control: Standard rehabilitation care. The intervention phase will last 12 weeks, with daily home training for experimental groups, and follow-up at 20 weeks. In addition to standard clinical endpoints, the study will include predefined AI validation metrics, focusing on its potential as a certified medical device tool for scalable, personalized rehabilitation delivery.

Interventions

• Device AI-personalized virtual reality rehabilitation system for unsupervised home-based stroke therapy
  The personalized RGS app rehabilitation is a home-based, virtual reality therapy platform for motor and cognitive stroke recovery. Therapy tasks are gamified, task-specific, and adapt in difficulty based on real-time performance. An AI-driven clinical decision support system personalizes and updates exercise prescriptions after each session, with optional clinician adjustments. Integrated wearable sensors (RGSwear) track real-world activity and adherence. Data are securely uploaded to a cloud-ba

Primary outcome measures

• Upper limb motor change [Time frame: From enrollment to the end of treatment at 12 weeks, and follow-up at 20 weeks]

Secondary outcome measures (5)

• Cognitive function change [Time frame: From enrollment to the end of treatment at 12 weeks, and follow-up at 20 weeks]
• Disability evaluation [Time frame: From enrollment to the end of treatment at 12 weeks, and follow-up at 20 weeks]
• Emotional change [Time frame: From enrollment to the end of treatment at 12 weeks, and follow-up at 20 weeks]
• Quality of life and Health status [Time frame: From enrollment to the end of treatment at 12 weeks, and follow-up at 20 weeks]
• Therapists' qualitative evaluation of the AI-based decision support system performance [Time frame: At the end of the study, at 20 weeks.]

Eligibility criteria

Inclusion criteria

• ≥ 6 months post-stroke
• Patients presenting a first-ever ischemic or intracerebral hemorrhagic stroke
• Mild to Moderate unilateral upper limb motor impairment: Medical Research Council proximal and distal upper limb MRC ≥2; Action Research Arm Test: ARAT score < 50 (0 = no function, 57 = no functional limitation).
• Age > 18 years old
• Able to sit on a chair or a wheelchair and interact with RGS during an entire session
• Minimal experience with smartphone technology based on the clinician's opinion
• Willing to participate in the RGS therapy
• Sign the Informed Consent Form

Exclusion criteria

• Diagnosis with Epilepsy
• Severe cognitive capabilities preventing the execution of the experiment or according to clinicians' criteria.
• Severe associated impairment such as proximal but not distal spasticity, communication disabilities (sensory, Wernicke aphasia or apraxia), major pain (VAS > 75-100 mm), orthopedic devices that would interfere with the correct execution of the experiment (Modified Ashworth Scale > 3)
• Unable to use the RGS app independently according to the clinician's observations and lacking support from a caregiver to use the RGS app
• No experience with smartphone technology or based on the clinician's opinion.
• Refusal to sign the Informed Consent
• Participating or planning to participate in another trial while being part of the present study.

Criteria are shown verbatim from the registry (in English). Final eligibility is always assessed by the study center.

Healthy volunteers: No

Study design

Allocation

Randomized

Model

Parallel assignment

Masking

Double blind

Primary purpose

Treatment

Study locations

France · 1 center

• CHU de Limoges — Limoges

Italy · 1 center

• San Camillo Hospital, IRCCS — Venice

Romania · 1 center

• UMF — Cluj-Napoca

Spain · 1 center

• Parc Sanitari Sant Joan de Deu (SJDD) — Barcelona

Publications

• Rabadi MH, Rabadi FM. Comparison of the action research arm test and the Fugl-Meyer assessment as measures of upper-extremity motor weakness after stroke. Arch Phys Med Rehabil. 2006 Jul;87(7):962-6. doi: 10.1016/j.apmr.2006.02.036. PMID 16813784
• Lang CE, Edwards DF, Birkenmeier RL, Dromerick AW. Estimating minimal clinically important differences of upper-extremity measures early after stroke. Arch Phys Med Rehabil. 2008 Sep;89(9):1693-700. doi: 10.1016/j.apmr.2008.02.022. PMID 18760153
• Hsieh YW, Wu CY, Lin KC, Chang YF, Chen CL, Liu JS. Responsiveness and validity of three outcome measures of motor function after stroke rehabilitation. Stroke. 2009 Apr;40(4):1386-91. doi: 10.1161/STROKEAHA.108.530584. Epub 2009 Feb 19. PMID 19228851
• Ballester BR, Antenucci F, Maier M, Coolen ACC, Verschure PFMJ. Estimating upper-extremity function from kinematics in stroke patients following goal-oriented computer-based training. J Neuroeng Rehabil. 2021 Dec 31;18(1):186. doi: 10.1186/s12984-021-00971-8. PMID 34972526
• Cameirao MS, Badia SB, Oller ED, Verschure PF. Neurorehabilitation using the virtual reality based Rehabilitation Gaming System: methodology, design, psychometrics, usability and validation. J Neuroeng Rehabil. 2010 Sep 22;7:48. doi: 10.1186/1743-0003-7-48. PMID 20860808
• Duncan PW, Bushnell C, Sissine M, Coleman S, Lutz BJ, Johnson AM, Radman M, Pvru Bettger J, Zorowitz RD, Stein J. Comprehensive Stroke Care and Outcomes: Time for a Paradigm Shift. Stroke. 2021 Jan;52(1):385-393. doi: 10.1161/STROKEAHA.120.029678. Epub 2020 Dec 22. PMID 33349012
• Maier M, Ballester BR, Leiva Banuelos N, Duarte Oller E, Verschure PFMJ. Adaptive conjunctive cognitive training (ACCT) in virtual reality for chronic stroke patients: a randomized controlled pilot trial. J Neuroeng Rehabil. 2020 Mar 6;17(1):42. doi: 10.1186/s12984-020-0652-3. PMID 32143674
• Moher D, Hopewell S, Schulz KF, Montori V, Gotzsche PC, Devereaux PJ, Elbourne D, Egger M, Altman DG. CONSORT 2010 explanation and elaboration: updated guidelines for reporting parallel group randomised trials. BMJ. 2010 Mar 23;340:c869. doi: 10.1136/bmj.c869. No abstract available. PMID 20332511

Identifiers

NCT: NCT07138495 · AISN-2025

Primary sources (government registries)

• ClinicalTrials.gov (US NIH)

View this study on ClinicalTrials.gov ↗