Braingame Brian: A Randomized Controlled Trial for an Executive Functioning Training for Children with ASD

Background: There is an urgent need for effective interventions for children with autism spectrum disorders (ASDs). Current interventions focus mainly on teaching social or communicative skills, and appear to be relatively unsuccessful. Few studies focused directly on fundamental abilities such as executive functioning (EF). Children with ASD are known to experience difficulties in EF. Hence, training EFs seems promising, especially since EF interventions show positive effects in disorders highly comorbid with ASD such as ADHD.

Objectives: Two EF interventions - a working memory (WM) training, and a cognitive flexibility training - are studied in a large randomized controlled trial of children with ASD. The objective is to improve the trained EF (near transfer), and to obtain generalization of improvement to other EFs (far transfer), and to EFs in daily life (far transfer).

Methods: Children with ASD (n=102, 8-12 years, IQ>80) are randomly assigned to one of three interventions; a WM-, cognitive flexibility-, or non-EF training (active control condition) build into a computer game (Braingame Brian). The training consists of 25 sessions (40 minutes each), performed within six weeks. Each session contains both WM and cognitive flexibility training tasks. The task to be trained (e.g., WM in the WM training) increases in difficulty adaptive to performance, whereas the other task remains at a low level. To examine efficacy of the training, WM (Corsi), cognitive flexibility (switch task), and everyday EF (BRIEF) are measured pre-training, post-training, and 6-week-follow-up.

Results: Currently, data of 76 children are complete. In January 2014, data of all children will be complete. Preliminary analyses reveal that 1) Corsi performance of all children improved during the training, and remained stable at follow up. More importantly, children who received WM training improved more than children who received flexibility training, and marginally more than children who received non-EF training. 2) On the switch task all children decreased in error switch costs (difference between errors on switch and  repeat trials), but increased in reaction time (RT) switch cost (difference between RT on switch and repeat trials) after the training, but overall RT decreased. Surprisingly, this improvement was manifested between post-training and follow-up. Switch task performance did not differ between the interventions. 3) All children improved on the WM, flexibility and total scale of the BRIEF, but there were no differences between the interventions. The dropout rate was 25%.

Conclusions: The WM training seems to induce near transfer; children who received WM training improved most in WM. However, the WM training does not seem to induce far transfer, i.e. both cognitive flexibility and daily life EF did not improve more than in children who received flexibility or non-EF training. The flexibility training induced neither near, nor far transfer. Children who received flexibility training did not improve more in flexibility, WM, or daily life EF compared to children who received WM or non-EF training. Since there are large individual differences within ASD, we will also apply multilevel techniques in the final analyses to find possible predictors of training outcome and compliance.