Abstract
Illusory motion reversals can happen when looking at a repetitive pattern of motion, such as a spinning wagon wheel. To date these have been attributed to either a form of motion after-effect seen while viewing the adapting stimulus or to the visual system taking discrete perceptual snapshots at a rate of ~10Hz. Here we present evidence that is inconsistent with both proposals. First we explore what adaptation TF is optimal for inducing illusory reversals. We find that this conforms to a low-pass function that is not tuned for direction, ruling against a motion after-effect account. We also show that the optimal test TF for illusory reversals is stimulus dependent, being lower for colour relative to luminance-defined motion. This is inconsistent with an account based on a constant rate of discrete perceptual snapshots. Instead we posit that illusory reversals happen when an attention tracking system intermittently fails to keep up with a stimulus, due to adaptation of a low-pass TF channel. According to this, the maximal rate at which repetitive elements can be tracked via attention should be subject to adaptation. We show that this is true, with participants better able to track elements after adapting to relatively fast motion and worse after adapting to slower movement. Overall, our data are consistent with human motion perception being driven by two relatively independent mechanisms, which at times can provide conflicting signals.
Meeting abstract presented at VSS 2012