Could training executive functions have unintended negative consequences?
A recent study that trained working memory found decreased performance in an untrained task
The importance of executive functions in subjects across the curriculum has led many researchers to consider how training executive functions might improve performance in these subjects. We already know that it is possible to train executive functions, such as working memory or inhibitory control, but the key question is whether this improvement transfers to academic subjects that require these skills. So far, there is little research indicating successful transfer, but the field is moving towards an approach of training the executive function within the academic subject of interest: in contrast to simply training the executive function. This is more likely to be fruitful because it requires more explicit use of the skill in a new setting. Part of the training in this case might be simply raising awareness that a particular skill is useful within that subject.
A less discussed issue in training research is the possible unintended consequences of training. A recent paper by Matzen and colleagues found that performance on a recognition memory task decreased following working memory training in adults. Here, the training was not within a subject domain, but was a typical adaptive working memory programme, aiming to improve both verbal and spatial working memory. The authors hypothesised that there would be near-transfer to other working memory tasks, and far-transfer to a recognition memory task. In fact, while performance on the baseline working memory tasks increased, the training led to no near-transfer and lower performance on the far-transfer recognition memory task.
For me, the most interesting aspect of this paper is that participants were asked about their memory strategies. Analysis of this data suggests that participants who had received the working memory training were using less effective strategies: they were using the strategies learned during training but these were not effective in the recognition memory task. Ironically then, participants did show far-transfer to a new task, but it did not have the anticipated positive effect. The authors suggest that future studies include a larger battery of tasks to examine whether decreased performance is simply a quirk of specific measures or a genuinely concerning effect.
I think this finding further highlights the need for training to be explicit and occur within the subject domain. Explaining clearly to participants why a skill is being trained, and why it is useful within the context might help to guard against implicit unwanted transfer to other tasks. In my own research, I am considering an inhibitory control training programme within the context of science and maths. Explaining the mechanism through which inhibitory control impacts science and maths reasoning, alongside the subject-embedded inhibitory control training programme, might increase awareness of participants’ strategy use and allow selection of the appropriate strategy for the task at hand. Considering the limited transfer effects in the literature, this may be more beneficial than training inhibitory control in isolation and looking for transfer to science and maths.
Nonetheless, I will bear in mind the possibility that any training could have unintended negative consequences, and consider what these might be and how they could be measured and combatted. If we continue to find decreased performance in untrained areas, this will raise important questions about why and when we should implement training programmes, weighing up the benefits in one domain against the detriment to others.
The full reference for the Matzen paper:
Matzen, L. E., Trumbo, M. C., Haass, M. J., Hunter, M. A., Silva, A., Stevens-Adams, S. M., Buning, M. F., & O’Rourke, P. (2016). Practice makes imperfect: Working memory training can harm recognition memory performance. Memory & Cognition, 44, 1168. doi:10.3758/s13421-016-0629-4
This blog post first appeared on my personal website.