Abstract
Purpose: It is well-known that reducing the contrast of a slow moving stimulus reduces its apparent speed. Thompson's (1982 Vis Res) report of this finding also suggested that at speeds above 8 c/s reducing contrast increased perceived speed. However in a later report, Stone & Thompson (1990 Vis Res), using a more rigorous, forced-choice procedure, failed to collect reliable data at these higher speeds. I now report a thorough re-examination of the dependence of speed on contrast at higher rates of movement.
Method: Subjects viewed pairs of brief (500mS) sine-wave gratings, foveally presented in a 2 interval forced choice procedure. The speed of one grating (Variable) was stair-cased to match the perceived speed of the other (Standard). The gratings either shared the same contrast (10% and 70%) or differed in contrast, Variable 10%, Standard 70% & vice versa. 2 spatial (2 & 8 c/deg) & 5 temporal (2, 4, 8, 12 & 16 c/s) frequencies were investigated. 4 subjects completed all conditions 4 times.
Results: 1. At 2 c/deg, 10% contrast gratings appeared to move significantly slower than 70% contrast gratings at Standard speeds of 2 and 4 c/sec (by 27% & 12% respectively), but appeared to move faster at Standard speeds of 8, 12 & 16 c/s (by 12%, 21% & 20%). The cross-over point between a reduction and an increase in perceived speed with contrast reduction occurred around 6 cycles/s (3 deg/s). 2. At 8 c/deg the comparable figures are speed reductions at 2, 4 & 8 c/s (by 42%, 32%, 11%); speed increases at 16 c/s (17%) The cross-over point was much higher, around 12 c/s (1.5 deg/s).
Conclusions: These results, including the shift in crossover point at 8 c/deg, can be understood within a model that postulates that speed is computed from the ratio of a ‘fast speed’ (magno?) and a ‘slow speed’ (parvo?) mechanism. These results may present problems for Bayesian models that assume a slow velocity ‘Prior’ to explain the effects of contrast on speed perception.