The Role of Attention in Animal Learning

Virtually all animals exhibit the basic and important ability to learn that one event signals another. For example, people can learn that big black clouds signal an imminent downpour, and animals can learn about different signals for food and danger. Interestingly, however, simply pairing two events does not always result in learning that one of these events is a signal for the other: learning about the relationship between two events can be hindered - and sometimes prevented altogether - by the mere presence of another environmental event that is more informative. The purpose of the research I propose to conduct is to understand why this happens: why does learning sometimes fail? One reason why an animal may fail to learn about an event is because it has learned to shift its attention away from it. This project will investigate this possibility. One series of experiments will explore whether stimuli which have failed to be learned about consequently command less attention. Another series will explore whether learning about events that occur at the same time result in shifts in attention between events. These experiments will use rats which, for many years, have proved to be an excellent laboratory subject for the study of learning, and which have provided the best data to guide our development of theories of learning. This research is important because learning plays a central role in many of our explanations for how the human brain and mind works. If it is found that learning contributes to shifts in attention then the research will make a significant contribution to our understanding of normal psychological functioning. This research will also be of interest because mental disorders such as schizophrenia are thought to be the consequence of a disorder of attention. As we refine our theoretical understanding of how attention works, then we can better understand, and treat, those affected when it goes wrong. This research will cost £434,562 and will be conducted at the University of Nottingham School of Psychology.

Studies of conditioning in animals have revealed that learning about a target stimulus is hindered if it is conditioned in conjunction with another stimulus (overshadowing), or if the accompanying stimulus has a history of conditioning itself (blocking). Using an appetitive Pavlovian conditioning procedure with rats, the first series of experiments will examine whether blocking and overshadowing of the target stimulus are a consequence of it being a relatively poor predictor of the outcome, and thus loosing attention (Mackintosh, 1975). If they are then a subsequent test discrimination which is based upon blocked or overshadowed stimuli should be especially difficult to solve, as these stimuli should be ignored. Using a flavour aversion procedure, the second series of experiments will explore the contribution of stimulus predictability to variations in attention in within-event learning. As a consequence of exposure to two similar events (AX and BX) the unique features of these events (e.g. A) become more salient than the common element (X). Experiments will investigate whether the increase in attention to A is a consequence of it being a good predictor of the stimulus that accompanies it (i.e. always X), and the loss in attention to X a result of it being a poor predictor of the stimulus that accompanies it (sometimes A, sometimes B). The final series of experiments will examine whether stimulus selection effects, such as overshadowing and blocking, are present in within-event learning. Some researchers (e.g. Rescorla & Durlach, 1981) have proposed that within-event learning is fundamentally different to standard between-event learning. There is therefore a need to establish whether within-event learning possesses the same properties as between-event learning.