Abstract
Adaptation to an isotropic pattern of random dots, moving uniformly in one direction, elevates the luminance detection threshold for patterns of test dots moving in various directions. Threshold elevation is maximal whenever test and adaptation directions are the same, and falls to zero for opposite directions. The “tuning curve” for directional selectivity measured in this way is rather broad, with some threshold elevation being present even when test and adaptation directions differ by 45°. This direction-specific adaptation is only found when “absolute” detection thresholds are measured: if a higher, “pattern” criterion is used for detection, threshold elevation is non-specific for direction of movement.