Recent discoveries indicating that the brain retains its ability to adapt and change throughout life have sparked interest in cognitive training (CT) as a possible means to postpone the development of dementia. Despite this, most research has focused on confirming the efficacy of training outcomes, with few studies examining the correlation between performance and results across various stages of training. In particular, the relationship between initial performance and the extent of improvement, the rate of learning, and the asymptotic performance level throughout the learning curve remains ambiguous. In this study, older adults underwent ten days of selective attention training using an adaptive algorithm, which enabled a detailed analysis of the learning curve's progression. Cognitive abilities were assessed before and after CT using the Mini-mental State Examination (MMSE) and the Montreal Cognitive Assessment (MoCA). The findings indicated that: (1) Initial performance is positively correlated with Learning amount and asymptotic performance level, and negatively correlated with learning speed; (2) Age is negatively correlated with learning speed, while it is positively correlated with the other three parameters. (3) Higher pre-training MMSE scores predicted higher post-training MMSE scores but less improvement; (4) Higher pre-training MoCA scores predicted higher post-training MoCA scores and less improvement; (5) The parameters of the learning curve did not correlate with performance on the MMSE or MoCA. These results indicate that: (1)Selective attention training using adaptive algorithms is an effective tool for cognitive intervention; (2) Older individuals with poor baseline cognitive abilities require more diversified cognitive training; (3) The study supports the compensation hypothesis.
Keywords: Adaptive algorithm; CT; Cognitive ability; Compensation assumption; Learning curve.
© 2025. The Author(s).