Abstract
Introduction: Crowding occurs when objects are too close together and features from several objects are combined into a jumbled percept. Two classes of mechanisms have been proposed: those involving purely perceptual factors and those pointing to a limited spatial resolution of attention. To test the attentional resolution hypothesis, we recorded ERPs and measured the N2pc component—an index of focused attention—and the SPCN component—an index of visual working memory. If crowding reflects a failure of attention, then N2pc should be reduced under conditions of crowding. Participants and methods: Thirteen healthy young adults participated in this study. Two arrays of three letters were presented, one on each side of fixation. One array was red and the other was green. Observers were told to attend either to the red array or the green array and to report whether the central letter of the attended array was a vowel or a consonant. Crowding was manipulated by using three distances between the target and the surrounding non-target letters: near, intermediate and far. The stimulus duration was 200 ms, with an SOA of 1600–1800 ms. Results: We found significant differences in the correct response rates and RTs depending on the level of crowding, with better performance (fewer errors and shorter RTs) as the distance between the target and nontarget letters increased. The key result was that the N2pc was virtually eliminated when the nontargets were near the target, indicating that attention cannot be focused onto the target under conditions of crowding. Interesting, the SPCN was largest under conditions of crowding, suggesting that observers stored all three items in working memory when the target could not be selected via attention. Conclusions: These findings suggest that crowding is a consequence of attentional failure and that working memory is recruited when selective attention fails.
Meeting abstract presented at VSS 2012