“Learning to learn” as a new path for learning generalization in working memory: the case of action video game play

Abstract

It has recently been hypothesized that the broad perceptual and cognitive enhancements induced by action video game (AVG) play relies partially on the ability to quickly learn a new task (i.e., ‘learning-to-learn’; Bavelier et al., 2012). Here, we report convergent results supporting this hypothesis from three studies conducted at several distinct sites. In the first study, we recruited avid action video game players (AVGP, n=9) and non-video game players (NVGP, n=20) to perform a dual N-back working memory learning task. During the task, participants simultaneously performed an adaptive N-back task in both visual and auditory domains. We measured the N-back level reached in each of ten learning sessions. We found that AVGPs showed a higher learning rate than NVGPs. In the second study, we trained two groups of non-video gamers on action games (n = 14) or life/business simulation games (n = 11), respectively, for forty hours. Learning speed in the same N-back learning task was estimated after game training. We found that the action group learned faster than the control group, demonstrating a causal link between AVG experience and enhanced learning. In the third study, we replicated the same training experiment as above in 52 additional participants with low-to-moderate game experience, this time ensuring that experimenters were blind to group assignment. Again, faster learning was seen in the AVG trained group, particularly in participants with low initial game experience (the effect was in the same direction but non-significant in participants with more moderate previous game experience). In both Studies 2 and 3, we also found that individual differences in attentional control measured by a multiple-object-tracking task predicted participant’s learning rate in the N-back task. Taken together, these convergent results provide evidence for the “learning-to-learn” mechanism as a new framework to conceptualize the broad learning generalization induced by AVG play.