Perceptual Learning: Effects and Mechanisms

Experts in a variety of fields (e.g., professional wine-tasters, experienced radiographers) show the ability to make fine discriminations that nonexperts find impossible. Skills of this sort are acquired as a result of experience, partly by way of explicit training (as when an expert instructs his apprentice), but also simply as a consequence of appropriate exposure to the items to be discriminated / the opportunity to compare two wines, or two images appears to be critical here. This perceptual learning effect can be demonstrated in the laboratory -- naive individuals, presented with complex visual displays that are initially indistinguishable, become able to tell them apart after repeated exposure when this is arranged in a way that allows comparison between different displays. Our proposal makes use of this laboratory procedure to answer two basic questions about perceptual learning. First, experience with the stimulus allows it to be perceived differently. But what is the nature of the change in the perceived stimulus? We plan to do experiments that will test the suggestion that, for the experienced observer, the distinctive features of the stimulus come to stand out and attract attention, whereas other aspects of the stimulus display lose effectiveness. Second, we want to know about the mechanisms that are responsible for this form of learning. What learning processes are responsible for changing the perceptual effectiveness of a stimulus? Previous work has shown that perceptual learning effects obtained in studies of animal learning can be generated by the operation of very simple learning processes (specifically, those known as habituation and association formation). Our experiments will test the application of these concepts to the human case.

It has previously been demonstrated that preexposure to alternating presentations of two complex figures (checkerboard patterns with an extensive common background on which unique features are superimposed) will facilitate subsequent discrimination between them. This procedure constitutes an analytically tractable instance of the phenomenon of perceptual learning. We propose to conduct experiments that will use variations of this basic paradigm to determine the nature of the perceptual changes responsible for the effect and of the learning processes that underlie them. One set of experiments is designed to confirm that such preexposure allows an increase in the effective salience of the unique features of the stimuli and to explore the possibility that common features suffer a loss of effective salience. A second set will explore the nature of the learning mechanisms responsible for changes in stimulus effectiveness. Related work with animal subjects indicates a role for relatively simple learning processes (association formation and habituation) in producing such changes, and our experiments will investigate their applicability to the human case.