Altarelli, I. 1 , Reeder, P. 2 , Newport, E. 3 , Aslin, R. 4 & Bavelier, D. 1, 4
1 Brain and Learning Lab, FPSE, University of Geneva, Switzerland
2 Department of Psychological Sciences, Gustavus Adolphus College, USA
3 Department of Neurology, Georgetown University Medical Center, USA
4 Department of Brain and Cognitive Sciences, University of Rochester, USA
Playing fast-paced, action-packed video games, like first person shooter games, has been shown to lead to benefits in various aspects of perception, attention and cognition. It also confers advantages in motor and in perceptual learning. The main hypothesis to date is that action gamers have better "learning to learn" abilities, compared to controls. They would benefit from an increased ability to determine task relevant information, while ignoring sources of noise or distraction.
In the present study, we tested whether these enhanced learning abilities generalize to statistical learning. A group of action video game players (AVGP - at least 5 hours/week of action video game play in the past year) and a group of non gamers (NVGP - little to no experience with action games) listened to auditory strings drawn from a miniature artificial language. The language was composed of nonsense words linked by grammatical dependencies of the form (Q)AXB(R), where each letter represents a grammatical category, and Q and R serve as optional flanker categories.
As predicted, when compared to NVGP, we found AVGP to be faster in acquiring the rules of our artificial language.