Extrageniculostriate vision in the monkey. VII. Contrast sensitivity functions

Psychophysical and electrophysiological experiments have indicated the importance of spatial frequency components and their respective contrasts and orientations for the recognition of patterns. It is in the striate cortex where these types of information first converge, a fact that lends support to the accepted crucial role of this structure in pattern discrimination. Monkeys with total bilateral ablation of the striate cortex, however, retain a residual capacity for pattern discrimination and also can differentiate between a vertical and an oblique liminous bar. The present study explores their capacity for spatial frequency detection both as a function of contrast and, by extrapolation, at maximum contrast (visual acuity measure). Monkeys were presented with a forced choice between a homogeneous target and a vertically oriented sinusoidal grating in a pulling-in apparatus. Stimuli were produced by the transillumination of transparencies at spatial frequencies of 0.5, 1.0, 2.0, 4.0, 8.0, 16, and 32 cycles/deg, in 0.1-log unit steps of contrast from 0.79 to 0.006. The stimuli subtended 8° of visual angle and were matched for mean luminance at 20 cd/m². After mastering the discrimination of one spatial frequency at the highest contrast, contrast thresholds were first estimated by a staircase technique, and then determined by the method of constant stimuli. The procedure was repeated for each spatial frequency before and after histologically verified total bilateral removal of striate cortex and partial damage to circumstriate cortices. Discrimination at all spatial frequencies was mastered by all normal monkeys. Postoperatively, they could solve only problems with frequencues between 0.5 and 4.0 cycles/deg. Contrast sensitivity (threshold^-1) functions for normal and destriated monkeys have the characteristic inverted J shape. The high- and low-frequency limbs are related exponentially to spatial frequency, and the peak of the curve is about 2.0 cycles/deg. The dimensions of the functions, however, change significantly following the ablation. Sensitivity is depressed at all spatial frequencies. The mean 'visuogram' indicates a 26-dB flat loss. The mean high-frequency cutoff point is 43 cycles/deg preoperatively and 12 cycles/deg postoperatively equivalent to 0.7' and 2.5' of arc, respectively. The latter value is not worse than 20/80 on the Snellen chart. The variability of the response at each spatial frequency in the staircase method and the slope of the psychometric function derived from the method of constant stimuli provide a measure of 'instability' and 'precision', respectively, which are inversely related. Preoperatively, precision is significantly greater at high than at low spatial frequencies. Postoperatively, it is similar at all frequencies, and the values are lower than those determined preoperatively. The results demonstrate that destriated monkeys can detect gratings, although to a lesser degree than normal animals. Along with previous findings of a persistent but deficient capacity for bar orientation discriminations, these data may explain the residual pattern perception already demonstrated in such animals. Since variability of performance during contrast threshold determinations is dependent on spatial frequency preoperatively, but not postoperatively, it appears that whereas two mechansims operate in normal monkeys, only one functions in destriated animals. Based on available evidence, it may be conceivable that these mechanisms are related to the electrophysiologically indentified X- and Y-cell subsystems.