Abstract
Although vernier acuity is a hyperacuity in normal adults, it is worse than grating acuity during early infancy, and takes several years of normal development to achieve hyperacuity. We investigated the nature of the vernier mechanism in human infants by comparing vernier acuity and grating acuity under two luminance conditions, 20 and 120 cd/m2. Adult data and infant ideal observer simulations suggest that vernier acuity should follow approximately square root law in its dependence on luminance, while grating acuity should vary less with variations in luminance. We tested 40 infants aged 7 to 25 weeks, using steady-state visual evoked potentials (VEPs) and a within-subject design. Grating acuity stimuli were 5 Hz phase-reversing vertical sinewave gratings that varied in linear steps from low to high spatial frequency during 10-second trials. Vernier acuity stimuli were squarewave gratings containing six vernier offsets per bar; offsets appeared and disappeared at a rate of 5 Hz and offset size decreased in log steps during each 10-second trial. Thresholds were estimated by extrapolating the evoked response to zero microvolts, based on six trials per condition. In the youngest infants, vernier acuity was significantly better in the low luminance condition, compared to higher luminance. Grating acuity at low luminance was unaffected or slightly degraded compared to high luminance. In older infants, the luminance conditions had no significant effect on acuity. These data do not support infant ideal observer predictions. The results may reflect higher levels of neural noise in young infants.
Supported by NIH, EY12692