Purchase this article with an account.
Concetta F Alberti, Robert Goldstein, Alex R Bowers; Gaze and pedestrian eccentricity explains differences in detection of stationary and approaching pedestrians by drivers with hemianopia. Invest. Ophthalmol. Vis. Sci. 2016;57(12):1961.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
We previously evaluated the effects of homonymous hemianopia (HH) on detection of blind side pedestrians in a driving simulator. Pedestrians were either stationary at the side of the road or approached on a collision course. Although the proportion of timely reactions (the driver would have been able to stop) was similar in the two conditions, there were relatively more missed pedestrians in the stationary condition and more late reactions (the driver would not have been able to stop to avoid a collision) in the approaching condition. We investigated gaze and pedestrian eccentricity to see if they were related to these differences.
Data from 12 HH participants were analyzed for stationary (n = 138) and approaching (n =141) blind side pedestrians initially appearing at about 14° from the car heading. When the driver reacted to the pedestrian (pressed the horn), we determined the first time after the pedestrian appearance at which gaze intersected the pedestrian (i.e., detection likely occurred). For missed pedestrians, we determined the time at which gaze came closest to the pedestrian.
After appearing, the eccentricity of the stationary pedestrians increased rapidly with respect to the car heading. By comparison, the eccentricity of the approaching pedestrians remained approximately constant at 14°. Consistent with this, eccentricity at the time of closest gaze was greater for stationary than approaching pedestrians (16.8° vs. 14.0°, p<0.01), and the time of closest /intersecting gaze occurred sooner for both missed pedestrians (0.9 vs. 2.6 s, p < 0.001) and late reactions (2.1 vs. 3.1 s, p= 0.053).
In the stationary condition, pedestrian eccentricity increased rapidly such that pedestrians on the blind side quickly went deeper into the blind hemifield requiring large scans for detection. Thus, the pedestrian was either detected early or gaze did not reach the pedestrian, accounting for the higher proportion of misses reported previously. However, the approaching pedestrians maintained an approximately constant eccentricity so that the pedestrian was within the range of a typical eye saccade for longer, providing more time for detection to occur resulting in a later time of detection and accounting for the relatively higher proportion of late reactions in this condition.
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.
This PDF is available to Subscribers Only