Abstract
We previously (VSS 2014) reported EEG and behavioral data (response frequencies) supporting the hypothesis that that perceptual learning can be obtained with multiple features and that learning involves the development of non-independence at perceptual and decisional levels. This is consistent with evidence suggesting that multiple levels of processing and representation may be involved in perceptual learning. The present study explores these notions using EEG and response times (RTs), testing the hypothesis that learning of multi-feature patterns results in processing that is non-independent and parallel, as would be expected if the stimuli were learned as integral objects. Stimuli contained contrast-defined features, extracted from the 1865 drawing of the Cheshire Cat in Alice’s Adventures in Wonderland. Participants began by performing a detection task implemented as a double-factorial paradigm (DFP), in which stimuli contained 0, 1, or 2 targets, at two levels of contrast, and were instructed to give a positive response if they judged the stimulus to have either 1 or 2 targets. They then practiced with all possible stimuli for 10-15 days, using an adaptive staircase procedure to track thresholds. Finally, they again performed the DFP detection task with stimuli presented at threshold and the two initial levels of contrast. EEG data were collected during both pre- and post-practice performance of the CID. Analysis of the RT data, at both the level of the mean and the distribution, suggested that the stimuli were initially processed serially, with practice resulting in a transition to dependent parallel processing and modulations of amplitudes of early ERP features, pre- and post-stimulus spectral power, and reductions in post-stimulus g-band power, selective to frontal electrodes. These results are consistent with the idea that perceptual learning of multi-feature patterns produces shifts in fundamental characteristics of processing consistent with the learning of an integral object.
Meeting abstract presented at VSS 2015