Further evidence on dimension-specific lateral inhibition in visual search

Abstract

Last year, we reported preliminary evidence that suggests that representations of neighboring search items on the same perceptual dimension inhibit each other (Chan & Hayward, VSS 2009). This year, we report further results from 4 experiments to strengthen and refine this notion. To tease apart a distance effect from a hemifield effect, we always put the target and the singletons in the same hemifield. We studied dimension-specific spatial interaction between two ‘special’ search items - a target and a singleton in Experiments 1 and 2, and two singletons in Experiments 3 and 4, by manipulating inter-item distance. For the target-singleton experiments, since mutual inhibition between target and singleton reduces target signal, search should be slowed and result in a stronger “attentional capture”. For the two-singleton experiments, mutual inhibition between singletons reduces their distractions, causing a weaker capture. As such, inter-item inhibition can be measured in terms of capture. In the experiments, we found stronger signs of mutual inhibition for close items than for mid/far items, when they were defined by the same dimension. No such signs were found for different-dimension items. This confirms that lateral inhibition occurs on a dimension-specific map. However, even at mid/far distances where lateral inhibition was at floor, capture was still larger for two singletons of different dimensions. This suggests that the lack of spatial inhibition across dimensional maps is not the only reason that two singletons of different dimensions produce stronger capture. We speculate that non-spatial dimension-specific inhibition also takes place, so that signals of irrelevant dimensions are inhibited. However, inhibition is limited to one dimension at a time, and so is not effective for avoiding capture from two dimensions simultaneously.