S. Salvalaggio1, M. Agostini1, F. Baldan1, C. Berlingieri1, M.V. De Gregorio1, P. Kiper1, L. Maistrello1, S. Osto1, G. Pregnolato1, A. Turolla1
1San Camillo IRCCS, Laboratory of Rehabilitation Technologies, Venice, Italy
Background: Neuroplasticity in the motor cortex is underpinned by learning-dependent mechanisms. Internal models of the body and the environment are required to improve motor control and performance. Rehabilitation modalities based on Virtual reality (VR) provide enriched environment for patients’ training, characterised by augmented real-time feedback on motor performance and results. The aim of VR modalities is to maximise learning-dependent neuroplasticity for the recovery of motor function, after stroke.
Purpose: To compare whether the continuous visualization of an on-line teacher, during VR-based rehabilitation, is more effective than the same treatment provided without displaying the on-line teacher, for the recovery of upper limb (UL) motor function after stroke.
Methods: All consecutive inpatients, diagnosed with first-ever cortical-subcortical stroke, ischemic or haemorrhagic, admitted to the Stroke Rehabilitation Unit at the San Camillo IRRCS in Venice (Italy) were enrolled. Patients were randomized in two groups (i.e. teacher, no-teacher). In the teacher group, specific feedback called “virtual teacher” was always displayed online during every task repetition, showing the end-effector correct kinematic to be emulated by the patient. Whereas patients in the no-teacher group performed the same upper limb exercises, but without the virtual teacher assistance. In both groups, patients received the same amount of VR therapy (i.e. 1h/d, 5dd/w, 4 ww: 20h overall) and an additional hour of Conventional Therapy (CT). Clinical effects were assessed before and after treatment with the following outcome measures: Fugl-Meyer Assessment (FMA) scale for upper extremity (FMA-UE, primary outcome), FMA sensation, FMA pain/rom, Reaching Performance Scale (RPS), Functional Independence Measure (FIM) and Modified Ashworth Scale (MAS). Moreover, kinematic outcomes were assessed by: Time (s), Speed (mm/s), Smoothness (number of submovements).
Results: Overall, 124 patients were enrolled and randomized to teacher (N = 62) and no-teacher (N = 62) groups. On average, patients were 63 (± 14) years old, with a distance from stroke onset of 7.36 (± 14.65) months. All clinical outcomes improved significantly within both groups, while no differences were observed between groups. Considering a minimally clinically important difference (MCID) of FMA-UE ≥ 5, a similar amount of patients (χ2 = 0.29307, p = 0.59) “responded to therapy” in both teacher (58%) and no-teacher groups (52%). Regression models of responders showed that MAS at biceps brachii affects negatively the FIM scale (R2=0.79). With regard to kinematic data collected in 70 patients (teacher, N = 38 patients; no-teacher, N = 32 patients), all of them improved significantly in both groups after treatment, but no differences emerged between groups.
Conclusion(s): The results confirm the positive effects enhanced by VR rehabilitation in empowering recovery of UL motor function after stroke. However, a single specific feedback (i.e. on-line visualization of a virtual teacher) does not seem to induce significant differences as postulated by supervised learning paradigms. The effect of different combinations of feedbacks (e.g. visual, auditory, virtual teacher, haptic) to provide to patients during VR rehabilitation should be better investigated.
Implications: VR is an effective rehabilitation modality that can be used to maximise rehabilitation induced neuroplasticity, thus promoting better recovery of UL after stroke.
Funding, acknowledgements: Funded by the Società Italiana di Fisioterapia (Bando di Ricerca S.I.F. 2013) and by the Italian Ministry of Health (Grant-GR-2011-02348942).
Keywords: Stroke, Virtual Reality, Rehabilitation
Topic: Neurology: stroke
Did this work require ethics approval? Yes
Institution: San Camillo IRCCS
Committee: San Camillo IRCCS
Ethics number: Prot. 2013.11
All authors, affiliations and abstracts have been published as submitted.
