Visuospatial neglect after stroke undermines spatial awareness, mobility and daily function, often increasing fall risk and care needs. Technology-enabled training offers structured, repeatable practice that can complement physiotherapy and sustain engagement. An immersive hand-grasping task was developed to direct attention toward neglected space during a goal-oriented action, combining precisely timed auditory and visual cues. Early experiences were gathered from physiotherapists and two inpatients with visuospatial neglect over multiple supervised sessions. The programme emphasised patient-centred delivery, tolerability and feasibility within routine care. Initial signals point to high perceived usability and safety among trainee physiotherapists, encouraging yet variable performance trajectories across patients, stable standard motor scores over the short intervention window and positive subjective engagement. These observations motivate more structured evaluation while indicating how audiovisual cueing may assist attention orientation and compensatory motor strategies.
Co-Designed Task with Audiovisual Cueing
The intervention was co-created by software developers and physiotherapists to match real-time training needs for visuospatial neglect. Implemented in Unity3D and delivered through an HTC Vive headset with Leap Motion hand tracking mounted on the visor, the task presented a red ball target in an immersive scene. A directional audio cue, spatialised with head-related transfer functions, preceded the target’s appearance for 2.61 seconds, followed by a 105-millisecond interval before the ball emerged and moved along a fixed vertical path for grasping. Successful grasp caused the ball to disappear, reinforcing goal completion.
Design parameters supported accessibility and graded challenge. The field of view was set to 30 degrees horizontally and 50 degrees vertically. Targets usually appeared to the left in the neglected field at roughly 61 centimetres, with difficulty modulated through size, distance and bounce speed to adjust selection demands. Grasp detection used Unity’s collision system, while a central gaze point and trunk rotation reset the scene and initiated the next countdown. Audio cues were phased in from week seven to streamline early task demands and then standardised for patient sessions. Mounting the Leap Motion on the headset mitigated initial hand-tracking latency and improved gesture responsiveness.
Must Read: Harnessing VR for Chronic Pain Management Through Telehealth
Usability checks preceded patient work. Two trainee physiotherapists each used the system for about 10–15 minutes, then completed adapted questionnaires. Ratings indicated high ease of use, enjoyment, intention to use and safety, with low perceived learning difficulty. For patients, sessions were set at around 10–15 minutes to reduce fatigue, with therapists present to ensure safety, adjust parameters and provide immediate guidance.
Supervised Sessions and Patient Experiences
Two right-handed inpatients with visuospatial neglect following right hemisphere stroke completed 12 supervised sessions under a consistent protocol. Inclusion required medical stability, absence of hearing impairment or hemianopia and the ability to follow instructions. Sessions used uniform directions, a standardised countdown and repeated trials in which participants oriented centrally, awaited the cue, then reached to grasp the ball as quickly as possible. Trial-level response times were automatically recorded to text files for later inspection. Therapists monitored compensatory strategies, including trunk rotation and delayed reaching, initiated trials via the task menu and maintained presence throughout while patients stood. Physiotherapy goals were set beforehand and reviewed at completion.
Upper-limb performance was characterised with the Box and Block Test and the 9-Hole Peg Test to index gross and fine motor abilities relevant to grasping. These scores remained broadly stable throughout the short intervention period. Right-hand performance was consistent, left-hand performance showed mild impairment and asymmetry on both tests aligned with expectations after right hemisphere stroke. Given this stability, changes in task timing are unlikely to reflect short-term gains in general motor capacity.
Subjective experiences were positive. Physiotherapists’ perceived usability matched trainee feedback. Both patients described the task as engaging and different from usual therapy. One reported no change in overall condition despite enjoying the sessions. The other reported greater walking confidence by the end, including progression to walking without support and described the audio as calming even if not directly performance-enhancing. Both highlighted increased difficulty when targets appeared deeper in the neglected field, consistent with the attentional load inherent to visuospatial deficits. Sessions were tolerated within the brief time windows designed to limit fatigue.
Individual Performance Patterns and Feasibility Signals
Trial-level completion times across the 12 sessions were modelled to explore patient-specific trajectories using gamma-family models appropriate for positively skewed timing data, with day as the predictor to accommodate unequal session spacing. Quadratic specifications provided the best fit among the candidate curves for both patients.
For the first patient, raw timings rose around the introduction of the audio cue at session seven. Model estimates indicated a non-significant initial linear increase with slight downward curvature thereafter, suggesting potential adaptation without statistical confirmation. For the second patient, timings declined significantly over time with a quadratic profile that reached a minimum near session ten, alongside reduced intraweek variability after session six. These divergent patterns under a common protocol point to individual differences in baseline profile, recovery course and engagement with compensatory strategies.
Feasibility indicators were favourable. Trainee physiotherapists rated ease, enjoyment, safety and intention to use highly, and the 10–15 minute session length supported tolerability. Continuous therapist presence enabled safe setup, parameter adjustment and real-time support. Although qualitative patient feedback was transcribed by therapists rather than captured verbatim, reports highlighted good acceptability, one case of increased mobility confidence and general enjoyment. With standard motor scores stable, observed timing changes likely reflect task-specific adaptation rather than broader motor recovery within the short window.
An immersive hand-grasping task with precisely timed audiovisual cueing was feasible to administer as an adjunct to physiotherapy for visuospatial neglect. Trainee physiotherapists reported high usability and safety, sessions were tolerated within brief supervised windows and patient feedback indicated engagement with occasional functional confidence gains. Performance data revealed patient-specific timing trajectories, including a statistically significant improvement pattern for one participant, while standard motor measures remained stable. Together, these early signals support structured expansion with larger cohorts, richer neglect profiling and standardised outcomes to clarify how audiovisual cueing within virtual reality can aid attention orientation, trigger compensatory motor initiation and align with rehabilitation goals without overburdening patients or services.
Source: JMIR Publications
Image Credit: iStock
