Abstract
Purpose: Perception is not isolated, but depends largely on recent experience. We used ERPs to isolate a cortical marker of the information gained from multiple samples of a visual stimulus over time. Specifically, we measured the extent to which the intervening delay between successive visual numerosity estimates modulated the amplitude of the P300, an index of contextual updating in working memory. Methods: We recorded EEG activity while observers estimated the number of dots in briefly presented displays. Although the positions of individual dots were randomized, numerosity was repeated one, two, three, four, or a random number of trials later (n-back). To assess the degree of dependence within response pairs, we computed the precision gained by averaging two estimates of the same numerosity and conjointly measured the amplitude of the P300 time-locked to the onset of the displays. Results: Although participants were not aware that numerosities regularly repeated, the average of two judgments of a given numerosity became more independent, and therefore more informative, as they were made further apart in time. In accordance with this behavioral effect, the amplitude of the posterior parietal P300 increased as an inverse function of n-back. Conclusions: We uncovered an inverse modulation of the P300 as a function of increasing intervening delay between two perceptual estimates. As the P300 reflects the extent to which a stimulus is consolidated into working memory, our results support the idea that an implicit memory trace of a visual pattern influences perception of subsequent retinal images, and offers a novel means of tracking the extent to which perceptions are built from multiple samples over time. This perceptual autocorrelation suggests a mechanism to mediate between the contrasting needs for (1) sufficient acuity to detect changes and represent salient stimuli, and (2) enough abstraction to maintain perceptual continuity in a noisy environment.
Meeting abstract presented at VSS 2012