Cycling with hemianopia to explore road user detection and scanning behaviour in virtual reality

Cyclists with homonymous hemianopia (HH) may face challenges in detecting road users, especially in unpredictable situations or with distractors. This study examined how HH affects detection, scanning behaviour, and the relation between detection and scanning in a virtual cycling environment. Participants with real HH, simulated HH, and unimpaired vision (N = 12 per group) cycled in a virtual environment, braking upon detecting road users during predictable and unpredictable events, with and without distractors. Scanning behaviour and detection performance (time remaining at braking onset before a collision would have occurred (TTC)) were recorded. TTC was similar across groups and unaffected by distractors. Real and simulated HH participants showed larger TTC declines, i.e. lower performance, during unpredictable events than unimpaired vision participants. Real HH participants scanned similarly to unimpaired vision participants, while simulated HH participants scanned their blind hemispace more. In real HH, wider horizontal gaze distribution was associated with longer TTC. Individuals with HH can maintain detection performance in our virtual environment, but may be hindered more by unpredictability. Scanning a wider horizontal range may improve detection, but a scanning bias towards the blind hemifield, observed in simulated HH, did not.