Abstracts

Rationale: Executive function impairments are common co-morbidities in children with epilepsy. We assess the effectiveness of a computerized cognitive training program (Cogmed) in children with epilepsy. We expand our initial results (presented at AES 2012) by investigating whether improvements remain at 3-months post-intervention, generalize to other skills beyond working memory, and can be predicted by pre-intervention factors. Methods: Twenty-five children (16 male; ages 8-15) with well-controlled seizures were randomized to a Hi or Lo Dose condition (15 Hi Dose); evaluations were at pre-, post-, and 3-month follow up (3mo.) (n=17) of the intervention. Inclusion criteria included FSIQ>80 and elevated scores (T>60) by parent report on the Working Memory Index of the Behavior Rating Inventory of Executive Function. Effectiveness of the program was evaluated by change on attention, working memory, academic fluency skills; and parent report of child behavior. Measures included the Automated Working Memory Assessment (AWMA) composite scores for Verbal Short Term Memory (VSTM), Verbal Working Memory (VWM), Spatial Short Term Memory (SSTM), and Spatial Working Memory (SWM); Woodcock-Johnson III Math and Reading Fluency; and the Child Behavior Checklist (CBCL). Repeated measures analysis of variance (ANOVA), logistic regression, and reliable change indices (RCI) were conducted. Results: There was a main effect of Time (p<.01); however, outcomes differed by dose, skill, and time point. Post-hoc analyses revealed short term memory skills improved pre to post (p<.05) and declined at 3mo. but did not fully return to baseline. There was a trend for SWM improvement from pre to 3mo. (p=.08). Math fluency skills also improved post-intervention and continued to increase at 3mo. (p<.05). RCI at 90%CI revealed that 80% of the Hi Dose and 40% of the Lo Dose improved on SSTM between pre and post; 33% and 11% improved on VSTM pre-post; and 30% and 25% improved on math fluency between pre and 3mo. Parent report of greater ADHD symptoms prior to intervention predicted smaller SSTM improvements post-intervention (R²=.391, p<05). Conclusions: Cognitive training showed reliable change for short-term memory skills and generalized to math fluency. RCI revealed a dose effect as more children in the Hi Dose showed clinically meaningful improvement. At three month follow up, gains in short-term memory skills were not maintained, yet spatial working memory and math fluency skills were substantially improved since baseline. Children too inattentive or hyperactive to engage in the Cogmed training may not benefit from the program. This is a promising non-pharmacological intervention for the cognitive comorbidities associated with childhood epilepsy.