Abstract
‘Motion energy’ measures the statistical information for discriminating the direction of a moving stimulus. It is proportional to the area, duration, and squared contrast changes of the stimulus. Because of its generality and power, motion energy has been an influential concept in contemporary vision research and theory. Nevertheless, psychophysical evidence is needed to evaluate the visual role of motion energy. How does discriminability vary with motion energy? Do space, time, and contrast have visually interchangeable effects in determining motion energy thresholds?
We will report a series of experiments that describe motion energy thresholds for discriminating the direction (left vs. right) of briefly presented Gabor patches. Motion discriminations were evaluated as a function of spatial area, temporal duration, and contrast as well as spatial and temporal frequency and retinal eccentricity.
In general, we find that the limits on motion discrimination are not much related to motion energy. Large variations in motion energy often have no effect, and discriminability often even decreases with increased motion energy. Discriminations of local motion energy are also influenced by the global spatial pattern of motion. Many of these results correspond to the physiological characteristics of cortical neurons and to the timing variability of spike trains.