Attention and Perceptual Learning

Perceptual learning as a phenomenon. We are all of us experts. Each one of us is able to effortlessly distinguish between a large number of people that we encounter in our everyday lives. When considered as stimuli in the abstract this is a difficult discrimination task as the stimuli are all rather similar to one another, yet we can identify an individual in a moment with an accuracy that would put a supercomputer to shame. We will argue that this particular expertise comes about because we are very experienced with the class of stimuli in question (in this case faces), and that it is simply a more commonplace example of a type of expertise that we can study in twitchers (experts in the identification of birds), field botanists (experts in plant identification in a particular habitat) and dog show judges (experts in fine discriminations betweens dogs drawn from the same class such as gundogs; as reported in Diamond and Carey, 1986). We can also manipulate experience with stimuli in the laboratory and show that an enhanced ability to discriminate between stimuli is a consequence of the right kind of experience with the right kind of stimuli (McLaren, Leevers and Mackintosh 1994; McLaren, 1997;Lavis and Mitchell, 2006; Mundy, Honey and Dwyer, 2007). This phenomenon goes by the name of perceptual learning, and many peoples first reaction to it is to make the claim that experience helps one learn what aspects of the stimulus to attend to. Perceptual learning, in its broadest sense, can be understood to be any enhancement of learning to a stimulus as a consequence of experience with that stimulus. Thus the proposal would be that perceptual learning is a consequence of the allocation of attention based on past experience with a particular stimulus type. This assertion is a natural (and plausible) response to the data, but is it the correct line to take? The question addressed in this chapter is whether perceptual learning phenomena can be explained by appealing to mechanisms that control attention (we will conclude that in some cases they can) or whether there is a need to appeal to other mechanisms as well (we hope to show that there is!). Starting with some examples of what might be considered perceptual learning phenomena taken from our early work, we go on to discuss them in the context of current research on perceptual learning, with an emphasis on analysing the possible role that attention might play in producing these effects. We will focus on perceptual learning in humans, but will allow ourselves to appeal to evidence from infra-human studies where appropriate.