Immersive VR with Synchronous Neurostimulation Outperforms Conventional Rehabilitation in Chronic Stroke Trial
Key Insights
A randomized feasibility trial published in Nature Medicine demonstrates that the MultiSensy (search) platform combining immersive VR with synchronized transcutaneous electrical nerve stimulation (search) significantly improved upper-limb motor function in chronic stroke (search) survivors compared to conventional physiotherapy.
At follow-up, the MultiSensy (search) group showed significantly greater improvements on the Action Research Arm Test (ΔARAT > 9) and Fugl-Meyer Assessment (ΔFMA-UE > 13), with 12 of 12 patients achieving clinically meaningful FMA-UE gains versus 8 of 13 in the control group.
The platform also produced significant improvements in tactile acuity and body representation, with arm length ratio perception normalizing after six sessions and sustained at two-week follow-up, while no adverse events were reported.
A multimodal immersive platform that pairs virtual reality rehabilitation games with precisely synchronized peripheral nerve stimulation has demonstrated clinically meaningful improvements in upper-limb motor function, sensory acuity, and body representation among chronic stroke (search) survivors, according to results from a randomized feasibility trial published in Nature Medicine.
The platform, called MultiSensy (search), integrates custom immersive VR scenarios with congruent transcutaneous electrical nerve stimulation (search) (TENS) of the median nerve (search), delivering electro-tactile feedback perceived at the site of virtual contact and precisely time-synchronized with interactions with virtual objects. In a two-stage investigation involving 34 stroke (search) survivors, researchers demonstrated not only the platform's usability but also its superiority over dose-matched conventional physiotherapy across multiple outcome domains.
Pilot Study Establishes Usability and Embodiment
In the initial pilot phase, nine stroke (search) survivors (mean age 59 ± 8 years) completed a single-day protocol comparing the MultiSensy (search) platform against a real-world control task. Participants reported a strong sense of embodiment within the virtual environment, with an embodiment score of 1.64 ± 0.50 compared to a control score of −2.04 ± 0.41 (P < 0.001, Cohen's d = 2.64). Seven of nine participants rated the platform's usability as "excellent" (System Usability Scale > 80.8), with the remaining two rating it as "good."
Kinematic analysis during the pilot revealed significantly higher movement velocity (0.24 ± 0.02 m/s versus 0.17 ± 0.03 m/s, P = 0.025) and greater forearm pronosupination range (0.25 ± 0.06 versus 0.09 ± 0.01, P = 0.021) in the MultiSensy (search) condition. Notably, the Body Landmark Test showed that arm length ratio perception improved after the MultiSensy condition (0.783 ± 0.039) compared to both baseline (0.571 ± 0.026, P < 0.001) and control (0.499 ± 0.026, P = 0.0004) conditions.
Randomized Feasibility Trial Design
Twenty-five chronic stroke (search) survivors (nine females, 16 males; mean age 59 ± 10 years) were randomized to either the MultiSensy (search) group (n = 12) or conventional rehabilitation (n = 13). Both groups underwent three weeks of dose-matched upper-limb rehabilitation comprising 12 sessions (four per week). The MultiSensy group completed six to eight rehabilitation games per session, each lasting at least four minutes, while the conventional group participated in therapist-supervised sessions matched in duration and movement types.
Assessments were conducted at baseline (T0), mid-intervention (T1, day 10), end of intervention (T2, day 19), and at two-week post-intervention follow-up (T3, day 33). Primary outcomes included motor function assessed by the Action Research Arm Test (ARAT) and Fugl-Meyer Assessment of Upper Extremity (FMA-UE), as well as self-body representation measured by the Body Landmark Test.
Superior Motor Recovery with MultiSensy (search)
The linear mixed-effects model for ARAT revealed a significant Group × Time interaction (F3,74 = 3.81, P = 0.013). Post hoc comparisons demonstrated significantly greater ARAT improvements in the MultiSensy (search) group at both T2 (MultiSensy: 8.25 ± 1.96; Conventional: 2.44 ± 1.08; P = 0.029) and T3 (MultiSensy: 9.17 ± 2.37; Conventional: 3.11 ± 0.98; P = 0.049).
For FMA-UE, the model similarly showed a significant Group × Time interaction (F3,89 = 4.86, P = 0.0035) and a significant main effect of Time (F3,89 = 20.50, P < 0.001). Post hoc comparisons confirmed significantly greater improvements in the MultiSensy (search) group at T3 (MultiSensy: 13.17 ± 1.30; Conventional: 7.54 ± 1.48; P = 0.01).
Single-subject analysis underscored these findings: eight of 12 patients in the MultiSensy (search) group achieved a clinically meaningful ARAT improvement (>5.7 points) by T3, compared to three of nine in the conventional group. For FMA-UE, all 12 MultiSensy patients achieved clinically meaningful improvement (>5.25 points) by T3, whereas eight of 13 in the conventional group showed similar progress.
Body Representation and Sensory Gains
Participants in the MultiSensy (search) group demonstrated significant improvements in arm length ratio perception as early as T1 (T0: 0.709 ± 0.039; T1: 0.916 ± 0.028; P < 0.001), with further gains at T2 (0.979 ± 0.021; P versus T1 = 0.02) that were sustained at T3 follow-up (0.965 ± 0.027; P versus T0 < 0.001). By contrast, the conventional rehabilitation group showed no improvements in body representation at either the arm or hand level.
Tactile acuity, assessed via the two-point discrimination (2PD) test, also favored the MultiSensy (search) group. Between-group analysis revealed significantly higher 2PD performance improvements at T2 (MultiSensy: 0.162 ± 0.044; Conventional: −0.006 ± 0.038; P = 0.008) and T3 (MultiSensy: 0.208 ± 0.046; Conventional: 0.030 ± 0.028; P = 0.006).
No adverse events were reported during either the pilot study or the randomized feasibility study.
Kinematic Analysis and Machine Learning Insights
During rehabilitation games, the platform continuously captured hand and arm pose data at 50 Hz. In the MultiSensy (search) group, participants significantly improved velocity (P < 0.001), score rate (P < 0.001), and smoothness (P < 0.001) across sessions. Between-group analysis revealed significantly higher velocity improvement in the MultiSensy group compared to conventional rehabilitation (P = 0.011).
Principal component analysis (PCA) of kinematic data from the Shapes and Baseball games revealed that the first principal component accounted for 41% of total variance, with efficiency and smoothness features emerging as key drivers. PC1 showed significant correlations with both FMA-UE (Pearson's ρ = 0.66, P < 0.001) and ARAT (ρ = 0.60, P < 0.001) scores. K-means clustering using the first eight principal components stratified patients into three clinically distinct impairment clusters (severe, moderate, mild) with significantly different FMA-UE and ARAT scores (ANOVA P < 0.001), with between-cluster differences exceeding the minimal clinically important difference for both scales.
Limitations and Future Directions
The authors acknowledge that the follow-up period was limited to two weeks, providing only partial insight into the persistence of observed effects. Additionally, the study cannot disentangle whether improvements were driven primarily by the VR-based games, the targeted sensory neurostimulation, or their synergistic combination. Future studies incorporating single-component conditions (VR-only or neurostimulation-only) are planned to isolate each modality's contribution. While the platform is inherently suitable for home-based rehabilitation, the current trial was conducted in clinical settings, and follow-up trials will evaluate its effectiveness in home-based environments.
The randomized feasibility study was registered with ClinicalTrials.gov (NCT06400823) and received approval from the Ethical Committee of the Clinic for Rehabilitation 'Dr. Miroslav Zotovic' in Belgrade, Serbia.