An exciting new paper by Lisa Knoll, Delia Fuhrmann, and colleagues from UCL, has just been published. The aim of the study was to look for sensitive periods in adolescence for training cognitive skills related to mathematics, a departure from much of the literature which focuses on early years training.
A total of 633 participants aged 11 to 33 years took part in one of three training groups, each of which provided 20 days of online training of a specific skill. The numerosity discrimination group were trained on the ability to quickly compare the numbers of dots in two sets. The relational reasoning group were trained on the ability to find patterns in relationships between shapes. Since brain areas involved in these tasks continue developing throughout adolescence, and performance in these tasks improves throughout adolescence, they might be good targets for intervention, so as to enhance the development of these skills. They are also known to be involved in mathematics ability, making the findings of potential relevance to education. The final training group was trained on the ability to process changes in faces. The cognitive and neural mechanisms of face processing are different to those involved in the other skills trained here, plus face processing is not involved in mathematics. Therefore this is the control group, and no transfer effects were expected between face processing and the other abilities.
There is some debate about what constitutes a good control group. Sometimes we want to see if the training is better than "business as usual", in which case we'd want a control group where no training took place. On the other hand, we want to make sure any training effects seen are specific to the task rather than due to simply being involved in a training programme. An active control group like the one in this study, means that any differences between the training groups and control group cannot be down to participation in an online training study, which might be novel and exciting for adolescents.
The researchers were looking for three main things: overall effects of training for the different groups, age effects of training where different aged participants might respond differently to training, and transfer effects where untrained tasks might show effects of training too. Transfer effects are notoriously difficult to find, and it is often only closely related tasks that show any transfer. In order to measure all of these effects, a pre-test, post-test (after training and 3-7 weeks after pre-test), and follow up test (3-9 months after post-test) was given to each participant.
The training itself consisted of 20 days of online training, where each day no more than 12 minutes was spent training. For the adolescents, this took place during normal classroom time.
The first finding of interest was that performance improved on the trained task following the training, and the extent of improvement differed across groups. The numerosity discrimination group saw improvements at post-test that were not sustained at follow up, and the gain seemed to be driven by the adult participants. The relational reasoning group showed improvements compared to pre-test at both post-test and follow up. Further, the improvements in this group were greater than in the other groups. Those who received face perception training also improved at post-test, but not at follow up, and when controlling for confounds this effect was no longer statistically significant.
To examine age effects, the groups were split into younger adolescents (11-13 years), mid-adolescents (13-15 years), older adolescents (15-18 years) and adults (18-33 years). The analysis of age group effects showed that it was only older adolescents and adults who improved in numerosity discrimination at post-test, although this was no longer statistically significant when confounds were included in the analysis. The relational reasoning training elicited improvements in all age groups at post-test and follow up, and these effects were stronger for older adolescents and adults than younger and mid-adolescents. Age group did not moderate the effect of face perception training.
Finally, there was no evidence of any transfer effects. This is unsurprising given the literature. Here, the authors suggest that investigating a broader range of tasks, both similar and dissimilar, might show some degree of transfer. If we are considering the relevance to education, transfer is of critical importance. Since we know that numerosity discrimination is related to mathematics, it would be really interesting to know whether the training did have an impact of mathematics ability. Previous research suggests that this is unlikely. However, this doesn't mean that cognitive training is necessarily a fruitless task. Research in this field is moving in the direction of creating training programmes within the subject domain. Rather than simply training the underlying cognitive skill, it might be crucial to show learners how the skill, say relational reasoning, is important for mathematics, and to practise it within the context. One theory is that we do not see transfer of improved skills because although students have the skill, they just aren't aware they should be using it in this setting.
This research shows us that we can train the cognitive skill, with improvements, and here the biggest gains are in relational reasoning. The next step is to integrate this into a programme targeting the particular subject where we want to see improved performance. A particularly exciting aspect of this paper is that it is shows a sensitive period for improvement in these skills during late adolescence, a developmental period that is rarely considered in similar studies. This means that adolescence is an important time for education, and late adolescence in particular might be a good time for such targeted interventions.
The full text of the original research paper can be viewed freely here.
Full reference: Knoll, L. J., Fuhrmann, D., Sakhardande, A. L., Stamp, F., Speekenbrink, M., & Blakemore, S.-J. (2016). A Window of Opportunity for Cognitive Training in Adolescence. Psychological Science, 27, 1620-1631.