| Abstract |
Most of the understanding regarding visual contrast perception has been achieved using simple gratings, which stimulate spatially tuned neuronal mechanisms. In order to study any interaction between these channels more elaborated stimuli need to be used. The incentive of the present study was to assess the contrast sensitivity using a wide range of spatial frequencies at the high spatial frequency side of the contrast sensitivity function. The description of the form and the extent of the underlying mechanisms of the human visual system was another issue studies in the present work. Moreover, their recognition was evaluated at sub-threshold and supra-threshold combinations.
Methodology
For the experiments a high resolution FD Trinitron® SONY© GDM F520 CRT monitor, a VSG 2/5 (Cambridge Research Systems©) graphic card, a CB6 wireless remote control (Cambridge Research Systems©) and a specially developed software for stimuli generation (NewRT, Dr Neil Parry©) were used. Five observers, all with prior psychophysical experience, participated in the experiments.
Three sets of experiments were conducted. The first two experiments were designed to evaluate threshold performance of compound gratings. The main spatial frequency component was always 4c/deg and the spatial frequency of the secondary component was 4.5, 5, 6, 8, 12, 16 and 24 which corresponds at a spatial frequency ratio of 1.125, 1.25, 1.5, 2, 3, 4 and 6. The third experiment was conducted to examine the interaction between the spatial frequency tuned channels while the contrast of the two components of the compound gratings ranged from sub-threshold to supra-threshold.
Results
Compound gratings were found to be slightly more sensitive by a factor of x1.15 (which corresponds to an average 1.2dB difference) compared to
the sensitivity of individual gratings (while the two components were at equal contrast above their respective thresholds). For spatial frequency ratio equal to one (the components were of same frequency), the sensitivity was higher (compared to the single grating) by 6dB (maximum summation), decreasing to 2-4 dBs for ratios between 1 and 2. When the spatial frequency ratio was > 2 there were no channel interactions and the sensitivity is solely increased due to probability summation, by a factor between 1.15 and 1.345.
The bandwidth of the spatial frequency tuned channel selective at 4c/deg was found to be at FWHM about 0.4 to 0.6 octaves for the two subjects. Moreover, when the secondary component was of spatial frequency lower than 4c/deg, the magnitude of neuronal summation was higher, suggesting a non-symmetric neuronal channel.
When the two components were at supra-threshold contrast, the high spatial frequency component depresses the low spatial frequency component and appears more detectable.
Discussion
Compound gratings appear more detectable. The amount of increased sensitivity depends on the ratio between the spatial frequencies of the components, being maximum (6dB), when the components are of equal spatial frequency, decreasing to 2-4 dBs when the ratio is about 2. The increased sensitivity is due to neuronal summation. For ratios higher than 2 the sensitivity is increased about 1-2dB, corresponding to probability summation. There is evidence for asymmetric spatial-frequency tuned neuronal channels. Finally, there are indications that the high spatial frequency component is more detectable when the two components are at supra-threshold contrast.
|
Evaluating threshold and supra-threshold performance using compound gratings
