Purpose:
Natural vision involves sequential eye movements that bring the fovea to locations selected by peripheral vision. We examine how the location and extent of peripheral visual field loss (PVFL) affects eye movement behavior in a naturalistic visual search task.
Methods:
9 patients with PVFL and 8 normally-sighted control subjects completed up to 30 monocular visual searches. Subjects located a 2°x2° target, pseudo-randomly selected from a 22°x9° natural image. Gaze points were collected at 50 Hz. Saccades were classified as eye movements that exceeded 30°/s or the mean+2SD of all eye movements for the trial. Remaining eye positions were considered fixation clusters.
Results:
Search duration, fixation duration, saccade size and number of saccades/trial were not significantly different between control and patient groups (p > 0.1). A chi square test showed that distributions of saccade directions were significantly different for patients and controls (Fig A, p < .01). Humphrey Visual Field Pattern Deviations for each subject (interpolated across the field, Fig B) were compared with the spatial distribution of eye movement vectors (Fig C). Interpolated visual impairment plotted against eye movement vectors (Fig D) were positively correlated in 7 out of 9 patients (mean=58.7% positively correlated) and negatively correlated in 1 patient.
Conclusions:
PVFL has a predictable effect on eye movement behavior. Most patients made significantly fewer eye movements into less sensitive areas of their visual field, indicating a maladaptive neglect of visually impaired locations. Although PVFL may not affect common measures of visual search, a comparison of eye movement distributions between patients and controls can reveal characteristic deviations that may be useful in screening, diagnosis and rehabilitation outcomes assessment.
Keywords: eye movements • visual search • low vision