Abstract
The binding problem in visual perception refers to linking features (e.g. color, shape, size, location) of a particular object to perceive it as a unified object. In this research we attempted to examine the binding mechanism underpinning location and identity in a categorical search. In a visual search task for features by Treisman and Gelade (1980), identity and location report were not tightly linked, such that one could report the identity but not the location of a target and vice versa. However, Johnston and Pashler (1990) reached a different conclusion by correcting for various kinds of errors and biases. They found that the detection of identity is highly contingent upon the correct detection of a feature's location and vice versa. It is not clear if similar effects would be obtained for targets specified at a more abstract, categorical level, such as letters among digits. Our prediction, based on findings of location-specificity in anterior portions of the ventral stream is that subjects would always have at least a general idea of the target's location if they could identify it. We used a categorical search paradigm in which five digits (distractors) and a letter (target) were presented simultaneously on the circumference of a hypothetical circle. Stimuli were presented for 50ms and then masked. Participants were asked to report the identity and the location of the letter on each trial. Contrary to our prediction, on trials in which the identity was reported correctly (chance was 3%) subjects were unable to report the general location (+/- one position around the hexagon, which was ½ of the positions) on 25% of trials. The results suggest that it is possible to detect categorical targets without being aware of their location. However, we did not find strong evidence that location can be reported without identity.
Meeting abstract presented at VSS 2018