Previous training studies (e.g., Green and Bavelier, 2003 Nature 423 534–537) link action video games (AVG) to improvements in visual cognition. However, specific components of AVGs that result in these improvements remain undetermined. One possibility is that only near-transfer occurs because game-play demands closely resemble the cognitive behaviors tested. Alternatively, far-transfer may occur to cognitive behaviors not specifically trained. To test between these possibilities, 4 groups of participants played 4 different AVG for 20 hours. Games differed in speed, number of simultaneous items to-be-tracked, and attentional switching demands. Group 1's AVG was similar to games in previous training studies with high speed and attentional demands, and multiple concurrent enemies (Modern Combat). Group 2's AVG had lower object-tracking demands due to sequential instead of concurrent enemies (Metal Gear Solid Touch). Group 3's AVG had slower speed by delaying counterattacks of enemies (Arcade Super Sniper). A fourth group played a hunting AVG that included target searching with no counterattacking enemies, and thus no speed, switching, or multiple object-tracking demands (Deer Hunter). Three tests of visual cognition—attentional blink, change detection, and visual search—were administered before and after training. Attentional blink was reduced only in groups 1 and 2 whose training contained high-speed and attentional switch demands. Only Group 1, with training of attending to multiple simultaneous attackers, improved change detection in multiple objects. All games contained visual search, and thus all groups improved in visual search after training. Results support the near-transfer proposal for visual cognition improvement.
