Training Executive Functions: Lessons Learned from Prefrontal Cortex Physiology
Final Report Abstract
Executive functions (EF( comprise several cognitive functions like shifting, inhibition and updating thereby enabling adaptive behavior in a constantly changing environment. EF are critical to our physical and mental well-being and as such at the core of psychological functioning in general. The starting point for our research was the increasing evidence that tremendous research efforts with computerized EF training had brought up rather sobering results. In fact, recent studies showed rather limited progress with respect to the generalizability and sustainability of training gains. In this research project, we therefore aimed to gain a better understanding of the mechanisms underlying successful EF training by increasing positive transfer and minimizing negative transfer. Our approach was guided by two principles: (1) We aimed to avoid negative transfer, and therefore first had to show that negative transfer in fact follows from the (widely used) repetitive invariant training regimes. (2) We aimed to show that training abstract representations should benefit wider transfer to tasks that share the same abstract task structure. In a first study, using short term task switching training, we were able to show the negative transfer as a consequence of repetitive training with invariant tasks as compared to a condition with constantly changing tasks. Moreover, we found evidence that training variability and higher task demands, while slowing the learning curve, results in better transfer performance. In a next step, we tried to gain insight regarding the near-lack of far transfer following WM training using an individual difference approach. We specifically engaged the hypothesis that due to the repetitive nature of computerized training, the training tasks which have originally related to fluid intelligence, become gradually less and less related to it. To test this hypothesis, we re-analyzed results from two relatively large-scale studies, with data relevant to the hypothesis. Our core results refute our original hypothesis and instead show that WM-training tasks maintain and even increase their individual-differences correlations with fluid intelligence in the course of training. These findings may suggest that WM training to improve Gf may rest on an unrealistic assumption of shared processes between Gf and WM that our analysis did not support. Finally, to investigate the second principle, namely to show that it is possible to train abstract control structures, we looked into the potential transfer of WM gating policies to unrelated tasks. More precisely, we examined whether a short training phase that either emphasizes the selective input or output of information from WM would have a corresponding impact on an unrelated task that profits more or less from input vs. output gating. Results look promising in that they showed better cued task switching performance after the training of selective input gating. In a further study using rapid instructed task learning, we found further evidence for the idea that abstract learning is possible. And again, the effect was stronger when task demands and task variability increased. Two contrasting hypotheses had been discussed in the literature. On the one hand, WM training was initially believed to have very broad long-lasting effects. On the other hand, recent meta analyses consistently showed lack of far-transfer, coupled with limited and short-lived intermediate transfer. Taken together, our research points to the conclusion that is somewhere in between these two extremes: It seems as if conditions which promote abstract coding result in learning of abstract task structures which support intermediate transfer. We believe that adopting this realistic stance can lead to the design of effective interventions, but such interventions should be tailored to the desired specific end result rather than being broad as originally hoped for.
Publications
- (2018). When less is more: Costs and Benefits of Varied vs. Fixed Content and Structure in Short Term Task Switching Training. The 4th Conference on Cognition Research of the Israeli Society for Cognitive Psychology (ISCOP), Akko, Israel
Sabah, K., Dolk, T., Meiran, N., & Dreisbach, G.
- (2019). Mechanisms of working-memory training: Insights from individual differences. Intelligence, 73, 78-87
Meiran, N., Dreisbach, G., & von Bastian, C. C.
(See online at https://doi.org/10.1016/j.intell.2019.01.010) - (2019). When less is more: : Costs and Benefits of Varied vs. Fixed Content and Structure in Short Term Task Switching Training. Psychological Research, 83(7),1531-1542
Sabah, K., Dolk, T., Meiran, N. & Dreisbach, G.
(See online at https://doi.org/10.1007/s00426-018-1006-7) - (2020). Enhancing task-demands disrupts learning but enhances transfer gains in short-term task switching training. The 7th Israeli Conference on Cognition Research of the Israeli Scocity of Cognitive Psychology (ISCOP), Akko, Israel
Sabah, K., Meiran, N., & Dreisbach, G.
- Enhancing task-demands disrupts learning but enhances transfer gains in short-term task switching training. Psychological Research (2020). April 18
Sabah, K., Dolk, T., Meiran, N. & Dreisbach, G.
(See online at https://doi.org/10.1007/s00426-020-01335-y)
