Action video game play facilitates “learning to learn”

Previous work has demonstrated that action video game training produces enhancements in a wide range of cognitive abilities. Here we evaluate a possible mechanism by which such breadth of enhancement could be attained: that action game training enhances learning rates in new tasks (i.e., “learning to learn”). In an initial controlled intervention study, we show that individuals who were trained on action video games subsequently exhibited faster learning in the two cognitive domains that we tested, perception and working memory, as compared to individuals who trained on non-action games. We further confirmed the causal effect of action video game play on learning ability in a pre-registered follow-up study that included a larger number of participants, blinding, and measurements of participant expectations. Together, this work highlights enhanced learning speed for novel tasks as a mechanism through which action video game interventions may broadly improve task performance in the cognitive domain.

The paper by Zhang et al. examined in two studies (study 2 is also pre-registered) whether action video game facilitates "learning to learn". The design chosen by the authors is exquisite. The intervention design in both studies proves the causal impact of action video game on "learning to learn" across cognitive domains-The sample size used by the authors is valid and so is the statistical analysis supported by Bayes factors. A big issue in the field is controlling for expectations (of the experimenter and of the participants). The authors crucially controlled for experimenter blinding, participant, blinding to conditions other than their own, and assessments of participant expectations. The data presented in the two studies are consistent with the "learning to learn" hypothesis. It is difficult to find any criticisms (methodological or theoretical) to this paper and I hope that the authors will keep producing such important contributions in the field.
Reviewer #2 (Remarks to the Author): Green and Bavelier initially reported "learning to learn" defined as enhancement of learning rates in new tasks, as a result of action game training in Nature in 2003. Ever since the initial finding, research on video game play has made tremendous influences on various fields of research including learning and memory, and education. However, recently the possibility of contamination of some artifacts in "learning to learn" resulting from video game training have been discussed. The current paper conducted experiments in which these artifacts were very carefully controlled. The results still showed clear and significant "learning to learn". The manuscript was well written. The results were robust. Above all, showing that "learning to learn" occurs by video game play training without the discussed artifacts is extremely important in the research on video game play. I do not have any particular concern. I strongly recommend this paper to be accepted for publication in Communications Biology.

Signed by Takeo Watanabe
Reviewer #3 (Remarks to the Author): In this manuscript, Zhang et al describe a study aimed at investigating how active videogames impact the learning abilities in other non-related tasks (perceptual learning and working memory). To test this hypothesis, the authors performed two intervention studies in which separate groups of adult volunteers played either action video games or "control" video games for 45h. Perceptual learning (orientation) and working memory (n-back) were measured to assess the participants' learning abilities before and after video games play. The main result of the study is that, compared to the control group, the action video game play group exhibited a faster learning rate both for perceptual learning and the working memory task despite initial similar learning abilities and despite reaching a comparabble final performance. From this result, the authors conclude that action video game play enhances the ability to "learn to learn", facilitating the ability to learn new tasks.
The experimental paradigm is sound and well controlled, I also would like to congratulate the authors for the effort in performing a second study using the pre-registration and including a large sample size. The statistical analyses are appropriate and the manuscript is overall well written. I only have a few minor points concerning the paper.
(1) I think that the figures should include single subjects's data, so that the reader could appreciate the interindividual variabiliy of the sample. For example, the bar plot of the learning rate in the two conditions might be either replaced or accompanied by a scatter plot.
(2) the authors controlled for several different factors that could affect the outcome of the study, including expectations, attentional abilities, motivation... however, two basic factors that might be potentially relevant have not been considered. I wonder whether the observed effect of action video games might be different across genders and with age.
1 Dear Editorial team, 1 2 We are really happy to hear about the acceptance in principle and thankful to the reviewers 3 for their enthusiastic response and positive feedback. Since the Reviewers 1&2 had no 4 comment to be addressed, we addressed Reviewer 3's comments below, one by one, our 5 responses appearing in blue. Changes in the text of the manuscript was highlighted in red. In this manuscript, Zhang et al describe a study aimed at investigating how active videogames 21 impact the learning abilities in other non-related tasks (perceptual learning and working 22 memory). To test this hypothesis, the authors performed two intervention studies in which 23 separate groups of adult volunteers played either action video games or "control" video games 24 for 45h. Perceptual learning (orientation) and working memory (n-back) were measured to 25 assess the participants' learning abilities before and after video games play. 26 The main result of the study is that, compared to the control group, the action video game play 27 group exhibited a faster learning rate both for perceptual learning and the working memory 28 task despite initial similar learning abilities and despite reaching a comparable final 29 performance.

30
From this result, the authors conclude that action video game play enhances the ability to 31 "learn to learn", facilitating the ability to learn new tasks.

33
The experimental paradigm is sound and well controlled, I also would like to congratulate the 34 authors for the effort in performing a second study using the pre-registration and including a 35 large sample size. The statistical analyses are appropriate and the manuscript is overall well 36 written. I only have a few minor points concerning the paper. 37 38 (1) I think that the figures should include single subjects's data, so that the reader could 39 appreciate the interindividual variabiliy of the sample. For example, the bar plot of the 40 learning rate in the two conditions might be either replaced or accompanied by a scatter plot. 41 We agree and added individual data points for all bar plots. 42 43 (2) the authors controlled for several different factors that could affect the outcome of the 44 study, including expectations, attentional abilities, motivation... however, two basic factors 45 that might be potentially relevant have not been considered. I wonder whether the observed 46 effect of action video games might be different across genders and with age. 47 This is a good point. We addressed this issue in details and added the analysis to the 48 supplements.