The role of task-irrelevant information in perceptual learning

In human perception, visual perceptual learning is a well-known effect, showing that the adult neural system can achieve long-term enhanced performance on a visual task as a result of visual experience (Fahle & Poggio, 2002). The mechanisms underlying visual perceptual learning were debated for decades. Task-relevance, attention and awareness were thought necessary for perceptual learning (Shiu & Pashler, 1992; Ahissar & Hochstein, 1993; Schoups, Vogel, Qian, and Orban, 2001), but this view has been challenged by the discovery of task-irrelevant perceptual learning (TIPL), which occurs for task-irrelevant, unattended and even sub-threshold stimuli (Watanabe, Nanez & Sasaki, 2001). TIPL is a slow phenomenon, because thousands of training trials are necessary in order to observe perceptual learning for task-irrelevant stimuli. However, a fast form of TIPL (fast-TIPL) has been recently studied in the context of perceptual memories, accounting evidence of a learning mechanism similar to TIPL, in which task-irrelevant stimuli are better learned when presented at behaviourally relevant points in time (Lin, Pype, Murray & Boynton, 2010). In the present dissertation, the role of task-irrelevant stimuli in visual perceptual learning is examined. The first line of experiments aimed to deepen the understanding about the mechanisms underlying visual perceptual learning by investigating whether perceptual learning can be produced by the mere exposure to a task-irrelevant, sub-threshold feature, even when, during training, participants attend and perform a task on another feature of a homologous stimulus to that used during test stages. Additionally, the task-specificity of TIPL was examined. The results provided further evidence about TIPL by corroborating the hypothesis that TIPL can occur even when the training stimuli are homologous to those in pre- and post-test. A further interesting finding was that the visual perceptual learning yielded by the task-irrelevant and sub-threshold feature is task-specific, because it occurs only in the task for which participants received a specific training, and is not transferred to another task performed on the same stimulus. Second, it has been investigated whether and how the modulation of the primary task difficulty level affects TIPL, by using a fast-TIPL paradigm which allows to study the phenomenon of TIPL within a single experimental session. In a dual-task condition, the amount of attention towards task-irrelevant stimuli which is needed for fast-TIPL to be observed was investigated by modulating the attentional and cognitive load of the primary task. The results showed a massive dual-task interference between the processing of primary task stimuli and the processing and encoding of task-irrelevant stimuli: the increase in the attentional and cognitive load required by the primary task determined a complete depletion of attentional resources such as no other resources remained available to process the task-irrelevant stimuli.