Integration of virtual reality and eye-tracking-based feedback system improves rehabilitation observation skills

Rehabilitation education has traditionally relied on an apprenticeship model, where the acquisition of practical skills, such as the ability to observe subtle patient movements to detect abnormalities, is highly dependent on the tacit knowledge of experienced practitioners. This lack of objective evaluation and standardized education presents a significant challenge. To address this, we integrated virtual reality (VR) and eye-tracking technologies to objectively measure and educate behavioral observation skills. This study aimed to quantitatively verify the educational effectiveness of this approach in novice therapists.

We tested two hypotheses: first, that the integration of VR and eye-tracking can objectively quantify and differentiate the observation skills of expert and novice therapists and second, that feedback education based on this objective data can effectively improve novice skills to an expert level. Using a single-case, multiple-baseline design (A-B design), four novice physical therapy students observed a VR video of a patient with right-sided hemiparesis performing a reaching task. During the intervention, each novice received personalized feedback using their own visual scanning heatmaps compared to the expert average, along with detailed verbal explanations.

The novice baseline visual scanning patterns, which showed a concentrated focus on the patient's upper body, were quantitatively distinct from the broad, stable visual scanning patterns of four expert therapists. Through the intervention, data rate (PND) exceeded 70% in many participants and phases, indicating an immediate and powerful effect of the intervention. This improvement was sustained by one-month post-intervention. Concurrently, participants' self-efficacy, feasibility, and social validity scores showed notable increases and high median ratings.

These results provide the first scientific evidence that an integrated VR and eye-tracking system, offering data-driven feedback, can effectively and lastingly improve rehabilitation observation skills, which have previously been difficult to teach and evaluate.