Abstract
Purpose. Murakami (Vision Res. 2004) hypothesized that normal motion sensitivity is limited by the variability of fixational eye velocity. The present study tested this hypothesis by increasing fixation variability preferentially in one meridian.
Methods. Motion thresholds were measured for 8 gaussian-modulated random dot patches, arranged symmetrically at an eccentricity of 10 deg around a central fixation spot. On each trial, the random dots in each patch translated for 500 ms to produce common motion in the horizontal, vertical, or rotary direction. The motion threshold in each meridian was the target velocity that yielded 50% correct judgments. Fixation variability was the SD of eye velocity during monocular fixation in the horizontal, vertical and torsional meridians, as measured with a search coil. Subjects (N = 10) viewed the stimulus with gaze either straight ahead or deviated temporally by 45 deg.
Results. A shift from straight ahead to temporal gaze increased the SD of eye velocity substantially more in the horizontal meridian (∼3×) than in the vertical or torsional meridians (∼1.5×). Horizontal motion thresholds increased by an average of 1.6×, much less than the SD of eye velocity. In temporal gaze, no relationship existed between the motion threshold and the SD of eye velocity in the horizontal, vertical or torsional meridians.
Conclusions. The results are not consistent with the hypothesis that fixation variability universally limits motion sensitivity, and suggest a role for extra-retinal eye-movement signals to reduce the influence of retinal image noise on motion processing.
Support: R01 EY05068, T35 EY07088, and P30 EY07551.