Does the Temporal Impulse Response Cause the Flash Lag Effect?

Abstract

A moving object appears to lead a briefly flashed object, even when the two are aligned. This illusory phenomenon is called the Flash Lag Effect (FLE). Previously, we introduced a stimulus configuration designed to rigorously define the stimulus and allow us to understand the illusion in the frequency domain (Cantor & Schor, VSS 2002). We then used this paradigm to test previous claims that the magnitude of the FLE is linearly dependent on the velocity of the stimulus object. Instead, our results suggested that a more complex frequency tuning underlies the FLE (Cantor & Schor, VSS 2003). We now demonstrate that a low-level computational model based on the temporal impulse response (TIR) can account for our data. The model is adapted from Fu, Shen, & Dan (2001), and incorporates temporal contrast gain control. Specifically, in our model the TIR profiles are compressed or shortened as a function of stimulus frequency as well as contrast. As a result, the output of our model makes predictions for the contrast and frequency dependence of the FLE. We present data that confirms these predictions. It bears noting that our model can account for the full magnitude of previous measurements of the FLE (including delays on the order of 100 ms). We believe this to be evidence that the FLE occurs very early in the visual pathways. Because the perceptual delays predicted by this model occur so early, the model may also explain perceptual phenomena in other modalities, such as the Pulfrich effect in Stereo.