Effects of immersive virtual reality on upper limb function in subjects with multiple sclerosis: A cross-over study

Background: Upper limb dysfunctions are common in people with multiple sclerosis (PwMS) and lead to limi- tations in activities of daily living. In this study, we investigated the feasibility and effects of an immersive commercial virtual reality system for upper limb bilateral rehabilitation. Method: A total of 20 participants were included in a cross over study with two arm sequences: Treatment- Waiting List (T-WL; 􏰳 = 9) and Waiting List-Treatment (WT-T; 􏰳 = 11). T-WL sequence performed 12 ses- sions of bilateral UL rehabilitation over a 4-week period, based on the use of a commercially VR immersive platform (Oculus Rift), followed by a 4-week wash-out period and a 4-week waiting list period. WL-T sequence followed the protocol in the reverse order. Participants were tested at baseline (T0), after the end of the first 4-week period (T1), at the end of the wash-out period and finally at the end of the third 4-week period (T2). The primary outcome was the Box and Blocks test (BBT). Secondary outcome measures were: Nine Hole Peg Test (NHPT), Maximal isometric handgrip strength, Manual Ability Measure-36 (MAM-36), Modified Fatigue Impact Scale (MFIS), and the System Usability Scale (SUS). In absence of carryover effects, we analyzed primary and secondary outcome measures with mixed linear effect models. Treatment efficacy was assessed on the within-subject differences. Specifically, we used the intra- individual differences at the end of treatment and waiting-list periods (T1 and T2) as dependent variables and sequences (T-WL or WL-T) as independent variable. In presence of carryover effects (p-value <0.05), we assessed between sequence differences by an unpaired t-test considering T0 and T1 as time points, and sequence as group factor. Results: We observed clinical and statistical improvements for BBT, with an overall between-sequence difference of 􏰴.6 ± 2.6 blocks (p < 0.01) favoring treatment period in the less affected side, and a not significant change of 3.0 ± 2.6 blocks (p = 0.2􏰴) in the most affected side. Small and not significant between-sequence differences were found for 9HPT, and handgrip strength in both sides. Similarly, no differences were found for patient re- ported outcomes, MFIS and MAM-36. Finally, mean SUS score was 45.9 ± 11.1 points, representing a moderate usability of the system. Conclusion: An immersive VR-based approach resulted useful to improve gross manual dexterity in the less affected limb in PwMS. However, such improvement did not translate into modifications in terms of self-reported ability to carry out activities of daily living nor went along with improvement in fine hand dexterity, strength or fatigue. Finally, usability of this technology was overall judged moderate, with lower scores assigned to items representing user-friendliness.