The role of comparison in perceptual learning

Abstract

This thesis focuses on the analysis of perceptual learning from an associative framework. For this purpose we used a variety of procedures in both human and non-human animals with the aim of finding a common ground between species. Chapter II is dedicated to experiments using visual stimuli in human participants. We demonstrated that additional exposure to the unique elements of a checkerboard only improves discrimination when it points to their location within the stimuli. Thus, the memory representation of the unique elements is not relevant to explain perceptual learning under these conditions, but rather the task can be solved focusing only on their location. We also demonstrated that explicit instructions to look for differences are needed to obtain perceptual learning, and that alternative instructions that require similar focus on the stimuli do not improve discrimination. These results suggest that perceptual learning with visual stimuli in humans is not mediated by salience modulation of the unique elements caused by mere exposure, but instead depends on a location bias and instruction-driven self-reinforcement. In Chapter III we adapted the procedure used with humans to animal subjects. Hence, we obtained perceptual learning using a procedure with short inter-stimulus intervals, in contrast to the usual procedure with intervals of several hours. We achieved this by controlling the influence of the excitatory associations between the stimuli on the test. Furthermore, we demonstrated that adding a distractor in the middle of the exposed stimuli abolished perceptual learning, thus replicating a similar result with humans. Our results highlight the possibility that comparison might be a relevant mechanism to explain both human and animal perceptual learning, and that there is no need to postulate separate mechanisms for different species. Finally, in Chapter IV we replicated the standard animal perceptual learning procedure using a flavour preference conditioning paradigm. We posit that perceptual learning might be involved in human feeding behaviour, thus having several applications such as the development of effective interventions to promote healthy eating or the prevention of intake habits that can lead to obesity.