Abstract
Action video game play benefits performance in an array of sensory, perceptual, and attentional tasks that go beyond the specifics of game play. Rather than reflecting an enhancement in myriad independent cognitive abilities, we have recently suggested that this may reflect a singular improvement in the ability to perform statistical inference on sensory data (Green, Pouget, Bavelier, 2010). Using a perceptual decision-making task and a classical diffusion-to-bound model, action video game play was found to result in an increased rate of information accumulation and a decrease in decision bound, resulting in more correct decisions per unit of time and reflecting better statistical inference. Here, we ask how performance differences may emerge between action gamers and non-action gamers. A general change in sensitivity could be at play, predicting performance differences on the very first trial. Alternatively, an improved ability to learn task-relevant statistics may also account for the data. This latter view predicts comparable performance between groups initially, with group differences only emerging as participants gain experience with the task. To test these hypotheses, participants underwent four days of a dot-motion perceptual decision-making task. Consistent with the learning to learn hypothesis, a hierarchical Bayesian analysis of the behavioral data using an extended diffusion-to-bound model demonstrated that while parameter values were indistinguishable early in the experiment, differences in learning rate emerged quickly, with gamers clearly outpacing non-gamers. Combined with the fact that gamers also continued to show significant learning for a greater proportion of the total experiment (i.e. their performance plateaued later), this resulted in significantly more total learning in gamers. In sum, the true advantage conferred by action video game experience may be the ability to efficiently learn the statistics of novel task environments, an account that would naturally explain the breadth of tasks shown to benefit from such training.
Meeting abstract presented at VSS 2012