Abstract
In Experiment 1 the reversal point of a moving object (12°/s) was compared with a stationary fixation cross. For five out of six participants the reversing object's trajectory had to extend significantly past the fixation point (0.1°) in order for it to appear to reverse on the fixation point, apparently in agreement with temporal integration theories of spatial localization (e.g. Lappe & Krekelberg, 1998). In seeming contradiction to this, in a second experiment, no reliable illusion was in evidence when eight participants compared the point of disappearance of a moving object with the fixation cross. For seven of these latter participants the moving object had to traverse the fixation cross for it to appear to start moving co-incident with it (a Fröhlich effect). This suggests that the illusion generated by a reversing object is most similar to that of a suddenly appearing object. If we consider a reversing trajectory to consist of two components, it would seem that the component after the direction change masks the parts of the first component just prior to the direction change, which would have been visible had the object not reversed. The Fröhlich effect in the reversing case is an active masking, and may be triggered by unnatural infinite accelerations. Returning to Experiment 1, when we added an irrelevant flash to the condition described above, the magnitude of the illusion was significantly increased (cf. Müsseler, Stork, & Kerzel, 2002). To further investigate this, a double-judgement (first the flash, then the moving object) condition was introduced — this did not reliably increase the illusion. In the irrelevant flash and double-judgement conditions, the presence of the flash had the effect of masking a portion of the moving object's trajectory around its reversal point. We propose that this is a result of attentional capture (Baldo & Klein, 1995) by the flash and consequent masking of the moving object, with clear ramifications for accounts of the flash-lag effect (cf. postdiction, Eagleman & Sejnowski, 2000).
We wish to acknowledge Paul Bretherton and Amanda White for help with data gathering and analysis. Support was provided by a Griffith University Applied Cognitive Neuroscience Research Centre Grant.