Abstract
Under low visual contrast conditions, sensitivity to stimulus features is generally thought to be reduced. However, a number of studies have revealed complex deviations from this intuitive view. For example, Tadin et al. (2003) showed that direction discrimination of large, briefly presented, drifting gratings improves as contrast is reduced, and they interpreted their results as a perceptual correlate of the contrast dependence of surround-suppressed neurons in the visual cortex (Pack et al., 2005; Polat et al., 1998; Sceniak et al., 1999). Contrast dependence has also been demonstrated in cortical neurons for speed tuning (Pack et al., 2005; Livingstone & Conway, 2007), but no clear psychophysical correlate of this result has been found. Here, we investigated the ability of human subjects to estimate the direction of moving dot fields at a variety of spatial and temporal displacements under both low and high contrasts, and compared these results to neural responses recorded from cortical area MT of alert macaque monkeys under similar conditions. We observed that the estimation of motion direction depended both on the stimulus contrast and on the amount of spatial displacement undergone by the stimulus dots on each monitor refresh. Surprisingly, subjects were better at determining the motion direction of stimuli with small displacements at low contrast than at high contrast. For larger displacements this effect reversed. This result was mirrored in the activity of MT neurons. Additional experiments replicated the above interaction between contrast and spatial displacement for a variety of conditions, including those in which the mean luminance was matched between both contrast conditions. These data link contrast-adaptive responses in area MT with behavioral performance, and demonstrate that higher contrast is not better for motion direction processing at low speeds.
ARS was supported by NSF (BCS-0549036) and NIH (R21 EY017737). PKP was supported by NIH (R01-DC02852), NSF (IIS-0205271 and SBE-0354378) and ONR (N00014-01-1-0624). CCP was supported by CIHR (MOP-79352).