Purchase this article with an account.
Alexandra R Bowers, Kevin E. Houston, Robert B Goldstein, Eli Peli; Peripheral prisms and training improve detection of pedestrians by drivers with hemianopia. Invest. Ophthalmol. Vis. Sci. 2014;55(13):2155.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Peripheral prism (p-prism) glasses for hemianopia provide visual field expansion that patients in multicenter trials reported to be helpful for detection of blind side obstacles when walking; however, hazard detection performance has never been objectively verified either for walking or driving. We tested the hypothesis that p-prism glasses and training would improve detection of blind side pedestrian hazards in a driving simulator.
Ten patients with complete hemianopia (without neglect or significant cognitive decline) were fitted with oblique 57Δ p-prisms. After 2.4 weeks (median) wearing the prism glasses, they received 6 in-lab training sessions (each 1 hour). The computerized training involved repetitive touching of stimuli appearing in prism-expanded vision and seeing-field areas, presented over static images and video driving scenes (Houston et al. ARVO 2013). Detection in the driving simulator was evaluated at baseline without p-prisms and then with p-prisms immediately before and after training, and 3 months later. Pedestrians (n = 52 per assessment) walked or ran toward the road on a collision course; participants pressed the horn to indicate detection.
Blind side pedestrian detection rates improved significantly from 33% without prisms to 47% with prisms before training, and 73% after training (p = 0.02 and p = 0.03, respectively). This improvement was maintained 3 months later. However, there was wide between-subject variability: blind side detection rates ranged from 13% to 81% at baseline, and 27% to 96% with prisms after training. Blind side response times improved significantly from 3.2s without prisms to 2.6s with prisms before training (p = 0.02); but there was no further reduction following training (2.5s, p = 0.96). Seeing side detection performance (100% and 1.2 s) was significantly better than blindside performance even after training (p = 0.005).
P-prism glasses improved detection of potential blindside pedestrian hazards by drivers with hemianopia, and computerized training further improved performance. These preliminary results are promising and provide the first objective evidence of the effects of the p-prisms and training on detection in a realistic simulated driving task.
This PDF is available to Subscribers Only