Abstract
Vernier acuity thresholds represent minimal detectable spatial offset between two closely placed targets. We previously showed that Vernier thresholds for a Poisson-limited ideal observer with access to the cone excitations are determined jointly by duration and contrast through the quantity duration x contrast squared. Here we measured thresholds in 7 human observers for combinations of stimulus contrast (100%, 50%, 25%, and 12.5%) and duration (16.7 ms, 66.7 ms, 266.7 ms and 1066.7 ms), while fixing other stimulus properties (foveal viewing; two achromatic vertical bars; length 10.98 arcmin; width 4.39 arcmin; vertical gap 0.878 arcmin). The combinations of duration and contrast were chosen to form four groups of constant duration x contrast squared. Thresholds were a decreasing function of duration x contrast squared. A one-way between observers ANOVA does not reject the hypothesis threshold duration and contrast are integrated through the quantity duration x contrast squared, but the residuals obtained by predicting threshold within each of the four groups by its mean varied systematically with duration, indicating that duration x contrast squared does not fully summarize the information integration. This difference between ideal and human performance indicates that post-receptoral factors not included in the ideal observer model, such as temporal filtering, affect human performance. These factors will be included in future modeling.
Funding: This work received a UKRI’s Physics of Life grant funded by the Engineering and Physical Sciences Research Council and the Wellcome Trust [grant code: EP/W023873/1].