Aging is associated with a decline in a wide range of cognitive functions and working memory (WM) deterioration is considered a main factor contributing to this. Therefore, any attempt to counteract WM decline seems to have a potential benefit for older adults. However, determination of whether such methods like WM trainings are effective is a subject of a serious debate in the literature. Despite a substantial number of training studies and several meta-analyses, there is no agreement on the matter of their effectiveness. The other important and still not fully explored issue is the impact of the preexisting level of intellectual functioning on the training's outcome. In our study we investigated the impact of WM training on variety of cognitive tasks performance among older adults and the impact of the initial WM capacity (WMC) on the training efficiency. 85 healthy older adults (55-81 years of age; 55 female, 30 males) received 5 weeks of training on adaptive dual N-back task (experimental group) or memory quiz (active controls). Cognitive performance was assessed before and after intervention with measures of WM, memory updating, inhibition, attention shifting, short-term memory (STM) and reasoning. We found post-intervention group independent improvements across all cognitive tests except for inhibition and STM. With multi-level analysis individual learning curves were modeled, which enabled examining of the intra-individual change in training and inter-individual differences in intra-individual changes. We observed a systematic and positive, but relatively small, learning trend with time. Moderator analyses with demographic characteristics as moderators showed no additional effects on learning curves. Only initial WMC level was a significant moderator of training effectiveness. Older adults with initially lower WMC improved less and reached lower levels of performance, compared to the group with higher WMC. Overall, our findings are in accordance with the research suggesting that post-training gains are within reach of older adults. Our data provide evidence supporting the presence of transfer after N-back training in older adults. More importantly, our findings suggest that it is more important to take into account an initial WMC level, rather than demographic characteristics when evaluating WM training in older adults.
Keywords: cognitive training; dual N-back; older adults; working memory capacity; working memory training.