Abstract
The recent finding that a neuron's area of spatial summation can vary depending on stimulus contrast has been interpreted as a trade-off between sensitivity (important in low contrast conditions) and resolution (important in high contrast conditions). In these experiments we examined spatial resolution psychophysically by measuring subjects' ability to discriminate fine orientation differences underlying grouping tasks as a function of contrast. This was done using contour stimuli embedded in distracters (i.e. Field, Hayes and Hess 1993; Kovacs and Julesz 1993). In our task, subjects had to discriminate between a “pure” contour and a “noisy” one that had orientation noise added to the elements. Thresholds were taken as the amount of noise required for subjects to perform the discrimination task at 80% correct levels. We find that when the stimuli were equated for detectability, the low contrast stimuli required the addition of substantially more orientation noise than the high contrast ones in order to be discriminated from a pure (noise free) stimulus. This suggests that judgments on fine orientation structure are less precise at low contrast than at high contrast. This may be due to the internal representation of the stimuli at low contrast being noisier than at high contrast, or because selectivity of the mechanism is reduced at low contrast. These psychophysical results support the physiological findings in suggesting a major re-organization of spatial vision at low contrast.
Supported by NEI grants EY01472 and EY08300.