Mechanisms of Human Perceptual Learning: Interference, Consolidation, and Transfer
Human perceptual learning is the process of improving in basic sensory discriminations. This process may involve visual, auditory, tactile, and olfactory systems and forms essential foundations of human cognitive abilities. To understand neural correlates underlying this process and to improve learning in applied sensory domains, this dissertation discusses three aspects of mechanisms in human visual perceptual learning - interference, consolidation, and transfer. A number of studies have reported that perceptual learning is highly specific to trained stimuli features. These findings are often taken as evidence that the learning takes place at primary sensory areas. Of note, a study has shown that interference of the visual hyperacuity task occurs when two similar stimuli are learned sequentially. However, this was under the debate that some researchers' findings failed to support this effect and claimed that retrograde interference doesn't exist. The first chapter addresses this controversy by replicating both patterns of results, and demonstrates that retrograde interference in perceptual learning does occur when subjects' eye-movements are tightly controlled in a peripheral visual task. The existence of retrograde interference suggests that visual perceptual learning requires a period of stabilization before being interfered with by a second stimulus. Therefore, the second chapter investigates memory consolidation, a period that stabilizes memory traces after initial acquisition. This chapter aims to investigate the effects of caffeine and nicotine on memory consolidation of implicit and explicit learning and to discuss the role of a neurotransmitter, acetylcholine (ACh), in consolidation. This dissertation then moves on to the third chapter, which discusses the mechanism of retinotopic specificity in perceptual learning. Whether the perceptual gain results from synaptic changes in early sensory cortices, or whether changes in higher brain areas should also be taken into account still remains the central debate in this field. Notably, a novel double-training paradigm has revealed diminished spatial specificity when multiple stimuli were trained at different locations. To resolve this controversy, the third chapter discusses the roles of stimulus representation and training stimuli's precision in learning effects under double training, and finds that learning specificity depends highly upon particularities of the training procedure.