Evaluation of visual augmented feedback on manual wheelchair (MWC) users' wheelchair propulsion pattern in a MWC...

Efficient manual wheelchair (MWC) propulsion is essential for long-term mobility but is associated with upper limb overuse injuries. Clinical guidelines recommend the semicircular (SC) pattern, yet novice users often adopt suboptimal techniques and receive limited feedback. This study evaluated whether visual augmented feedback delivered through the miWe virtual reality MWC simulator improves SC propulsion, contact angle, and cadence in novice users. Eight adults with no prior MWC experience participated in an ABA single-subject design with 6 experimental participants and 2 control participants. Baseline and post-test were conducted with a real MWC and SMARTWheel, while intervention sessions occurred in a gamified simulator equipped with a convolutional neural network (CNN) propulsion classifier. Results showed that experimental subjects improved SC propulsion usage, with Tau-U values ranging from 0.33 to 0.67 and 100% non-overlapping data (PND) across all experimental participants, indicating consistent improvement and retention at post-test. In contrast, control participants showed deterioration, with negative Tau-U values (-0.67 and -0.33) and 0% PND, indicating no improvement. Contact angle improved in experimental participants (Tau-U up to 0.67, PND = 100%) but not in controls (PND = 0%). Cadence showed variable changes across participants, with Tau-U values ranging from -0.67 to 0 and mixed PND results (0-100%), indicating no consistent reduction at post-test. These results support the use of automated visual feedback in virtual reality (VR)-based MWC training to promote efficient propulsion patterns aligned with clinical recommendations, with the strongest and most consistent effects observed for SC pattern adoption.