Abstract
Purpose: :
We, previously, showed that amblyopes can integrate global spatial and motion information and they only exhibit selective performance deficits on global tasks that contain noise as well as signal elements. This finding is not consistent with the common belief that amblyopes have deficits for global form and motion integration. Here, we compare the results of amblyopic and normal eyes for a task that contains both integration and segregation with specific models of these two processes.
Methods: :
We used two complementary tasks, one involving global mean orientation and the other, global motion direction discrimination. We measured the performance when controlled amounts of pedestal noise were introduced. We applied a standard equivalent noise model to the threshold results, which provided us with two parameters of internal noise and number of samples. The integrator model was designed as to integrate both signal and noise, as it simply combined all available information. However, the segregator/integrator model was designed with perfect knowledge of the form of the signal distribution. Therefore it segregated all information outside the signal bandwidth and only integrated information within the signal distribution (2 SD about the mean). The threshold ratio of these two models was compared to that of 6 normal and 6 amblyopic observers.
Results: :
The threshold ratio of the integrator and segregator/integrator model was similar to that of amblyopic and normal eyes. Amblyopic thresholds were worse for stimuli that required minimal integration but maximal segregation by the normal visual system.
Conclusions: :
Our results suggest that the reported global integration deficit in amblyopes is due to a deficient noise segregation rather than deficient signal integration.
Keywords: amblyopia • motion-2D • shape, form, contour, object perception