UC Santa Barbara

Brain-training games are largely ineffective at improving player performance on cognitive measures, whereas action video games are largely effective at doing so, despite not being designed for that purpose. This dissertation draws upon the schema induction literature to propose that greater variety during the training phase is responsible for this discrepancy. Three studies training three different cognitive domains (mental rotation, working memory, and visual attention) were conducted comparing a condition that trained with relatively low variety of stimuli or environments to a condition that trained with relatively high variety. In Experiment I a group that trained mental rotation using three categories of shapes was compared to three groups that trained mental rotation with just one of the shape types on near, medium, and far transfer measures of mental rotation and spatial visualization. No difference was found between the low variety groups and the high variety groups on any of the cognitive posttest measures. In Experiment II a group that trained with an n-back using just letter stimuli was compared to a group that trained with an n-back using 6 types of stimuli and an inactive control group on near, medium and far transfer measures of working memory and fluid intelligence. The high variety group did not improve relative to the low variety group, and in some cases the low variety group improved more than the high variety group. In Experiment III participants with minimal prior experience with video games were assigned to play Call of Duty in a high variety condition, a low variety condition, or an inactive control condition. These groups were compared on their improvement on three measures of visual attention skill. In no case did the high variety group improve more than the low variety group. When the active groups were combined they did not improve more than the inactive control group. Across all three studies no benefit was found for the high variety condition relative to the low variety condition on related transfer measures of the respective cognitive skills. It is concluded that variety is not a useful addition to training interventions if the goal is to increase low-road transfer of

cognitive skills.