Description

This study addresses the fundamental issue of specificity and generality of training in the context of Perceptual Learning (PL). PL broadly encompasses the set of mechanisms through which experience with the environment gives rise to changes in perceptual processing. Careful research in this domain can greatly enhance basic understanding of the perceptual systems and the plasticity of these systems. Furthermore, translational approaches underpinned by the basic science of PL are becoming increasingly prominent. This includes a host of emerging translational approaches for the rehabilitation of both perceptual deficits and for cognitive training, which are believed to share cortical plasticity mechanisms. However, while existing research provides evidence that PL approaches can improve perceptual skills, the ability to develop effective interventions is limited by a lack of understanding of the behavioral outcomes associated with different PL approaches. One major obstacle to successful translation of PL is that the field to-date has been strongly driven by "novel" and "provocative" findings demonstrated via small N studies with very few projects digging deep to achieve robust and reliable results. In turn, not surprisingly, the field of PL, like many others in psychology, has suffered from numerous replication challenges. Here the investigators address these limitations by comparing a large number of different training tasks using common outcome measures and in a large subject population. Each training tasks involves a different "critical feature" for learning proposed by one or more research groups. However, these training tasks have never been directly compared or contrasted. Robust and reliable results will be achieved by training a large sample of participants on PL tasks and assess the outcomes via a common set of measures. The investigators will also collect a broad assessment of individual differences, which will provide a unique dataset that can resolve controversies in the literature and lead to new understandings. The proposed analytical approach tests several key hypotheses in the field, explores the extent to which different training approaches lead to systematically different profiles of learning, and examines how these can differ based upon the individuals being trained.

The present study investigates the mechanisms of Perceptual Learning (PL), with a focus on training task characteristics that induce generalizable enhancements in visual performance (i.e., that produce benefits on tasks beyond just the trained task, as this is critical for training to have translational value). To achieve this, the investigators target PL of spatial vision in human subjects, which is the most common target of PL in basic science research.

The investigators run a large number of participants across 12 separate conditions thought to mediate effects of PL (e.g., training with flanking stimuli, use of noise, manipulating difficulty during training, multisensory facilitation with sound, training with a diversity of stimuli, and cueing attention during training) using common outcome measures, and analyze both the effect of training condition and individual differences that impact learning. A novel crossover design is used to train participants in the younger cohort sequentially on two tasks and examine learning and generalization to determine which methods involve redundant or independent processes. It is noted that the first phase of training is the primary clinical trial and that the crossover is to address a mechanistic question regarding interactions of different training types. It is hypothesized that training with mostly difficult, precise stimuli will give rise to less generalization than training with easy, imprecise stimuli after sequential double training.

The investigators also specifically examine biological variables, such as sex and age, in addition to personality traits, sleep habits, motivation, and individual differences in baseline performance to gain a more complete understanding of how these factors may moderate outcomes of PL (including generalization to more real-world contexts, such as reading). It is hypothesized that a number of individual differences, both cognitive and personality based, will predict PL outcomes.

Participants in the younger cohort will be randomized to 1 of 12 arms, with each arm consisting of two conditions (of twelve possible conditions) that participants undergo sequentially. Participants in the older adult cohort will be randomized to 1 of 5 arms, with each arm consisting of a single condition.