Fig. 2 portrays blood–oxygen-level-dependent (BOLD) activation patterns before training for the letter-memory criterion task as well as for the 3-back and Stroop transfer tasks. Several observations from this figure should be highlighted. First, the young exhibited overlapping fronto-parietal activation for all three tasks prior to training, indicating a partly shared executive network. Further, there was robust striatal activity in the young for the letter-memory and 3-back tasks. The old also exhibited fronto-parietal activation during letter memory before training; however,unlike the young there was no pre-training striatal engagement forletter memory and 3-back in the old. The most striking training-related increase in BOLD activity was that the young showed increased striatal activity (with a peak in left caudate) during letter memory post training, and this change overlapped with corresponding striatal activity increases during 3-back (Fig. 3). Further, these activity changes fell within the same region as was activated for these tasks before training. No corresponding pattern was seen in neocortex. Here, the dominating pattern was reducedBOLD activity post training, likely reflecting decreased demands on executive control processes across the course of the intervention.