Training Executive Functions: Lessons Learned from Prefrontal Cortex Physiology

执行功能（EFs）使我们能够根据快速变化的约束条件灵活地调整我们的思想和行动，以优化目标的实现。EF依赖于包括前额叶皮层（PFC）在内的大脑区域网络，并与复杂问题的解决、认知发展和与年龄相关的认知衰退密切相关。此外，大多数精神病理问题都伴随着EF的损伤。因此，人们越来越有兴趣寻找改善EF的方法，并特别注意电脑化训练。然而，迄今为止的结果是模棱两可的。一些评论得出了乐观的结论。然而，元分析评论表明，虽然有可靠的实践效应（在训练任务中看到的改善）和相对适度的近迁移效应（结构相似任务的改善），远迁移效应（可能利用训练EF的结构不同任务的改善）主要是微小的。我们认为，目前大多数计算机化EF训练计划的设计并没有很好地建立在PFC神经生理学的现有理论基础上。在这个研究项目中，我们将测试两个相关的假设，关于为什么到目前为止计算机化训练的成功被证明是非常有限的：(1)从PFC生理学来看，除非在训练期间学到的抽象任务结构反映在迁移任务中，否则远迁移首先是不可预期的。(2)目前ef训练的设计方式导致了负迁移效应（由于练习导致的性能下降），这掩盖了（已经相当小的）有益迁移效应。这些负迁移效应的发生是因为(a)（指导性任务规则）的应用增加了屏蔽，(b)因为训练的重复性（不断重复少量的训练任务）使EF的参与最小化。

Executive functions (EFs) enable us to flexibly adjust our thoughts and actions according to rapidly changing constraints to optimize goal attainment. EF rely on a network of brain areas including the prefrontal cortex (PFC) and are heavily involved in complex problem solving, cognitive development, and aging-related cognitive decline. Moreover, most psychopathological problems go along with impairments in EF. Consequently, there has been growing interest in finding ways to improve EF, with special attention given to computerized training. However, the results so far are equivocal. Some reviews have reached optimistic conclusions. Yet, meta-analytic reviews indicate that while there are reliable practice effects (improvement seen in the training task) and relatively modest near transfer effects (improvement in structurally similar tasks), far transfer effects (improvements in structurally dissimilar tasks that presumably tap the trained EF) are predominantly tiny. We argue that the design of most of the current computerized EF training programs is not well grounded in extant theories of PFC neurophysiology. In this research project we will test two related hypotheses concerning why the success of computerized training had so far proven very modest: (1) From PFC physiology, far transfer is not to be expected in the first place unless the abstract task structure that is learned during training is mirrored in the transfer task. (2) How EF-training is currently designed results in negative transfer effects (performance deterioration due to practice) which mask the (already quite small) beneficial transfer effects. These negative transfer effects occur because (a) the application of (instructed task rules) increases shielding and (b) because the repetitive nature of the training (constant repetition of a small number of training tasks) minimizes the involvement of EF.