Executive control training does not generalize, even when associated with plastic changes in domain-general prefrontal areas.

How executive function training paradigms can be effectively designed to promote a transfer of the effects of interventions to untrained tasks remains unclear. Here, we tested the hypothesis that training with a complex task involving motor, perceptual and task-set control components would result in more transfer than training with a simple motor control task, because the Complex training would lead to more involvement-and in turn modification-of domain-general executive control networks. We compared performance and electrophysiological activity before and after 10 days of executive control training with the complex (n = 18) versus the simple task (n = 17). We further assessed the effect of the two training regimens on untrained executive tasks involving or not one of the trained control components. A passive control group (n = 19) was used to assess retest effects. Both training groups improved at the trained task but exhibited different plastic changes within left-lateralized and medial frontal areas at 200-250 ms post-stimulus onset. However, contrary to our hypotheses, they showed equivalent improvement to the passive group to the transfer tasks. Our collective results reveal that the effect of training with a task involving multiple executive control components is highly specific to the trained task, even when the training modifies the functional networks underlying the trained executive components. Our findings corroborate current evidence that general cognitive enhancement cannot be achieved with training, even when the interventions modify domain-general brain areas.