Fig. 4

Results of the PC-based cognitive test applied in the PFIFF2 study for the cognitive training group (unpublished data). Participants performed a memory-based task-switching test used in the previous PFIFF study (see the caption of Fig. 3). The test was applied before training (t1, solid black), after 3 months of training (t2, dotted black) and 3 months after the training was finished (t3, solid grey). Reaction times did not change significantly across the three time points. Error rates before training (t1) were similar to the presented in Fig. 3. However, a clear decrease of error rates occurred after the training (t2) which persisted 3 months later (t3). Bottom left: Target-locked ERPs at the electrode FCz. The frontocentral negativity, the N2 associated with response selection process increased from t1 to t2 and remained stable from t2 to t3, suggesting more efficient decision after training. Bottom left: Response-locked ERPs in erroneous responses. The error negativity (Ne) was initially weak and similar to the observed in Fig. 3 but increased considerably after the training (t2) and remained stable 3 months later. No effects were found for the waiting control groups between t1 and t2 (not depicted in the picture)