Study: Computerized Cognitive Training Improved Inhibitory Control in Children With Epilepsy

01What the study measured

The single-arm study followed a pre-test–intervention–post-test design without a control group. Researchers enrolled 26 participants (12 female) with a mean age of 12.11 years (SD 2.96), ranging from 7 to 17 years and spanning a wide variety of epileptic syndromes, including two cases of Dravet Syndrome and two of Rolandic epilepsy. A power analysis using G*Power indicated that 25 participants were needed to reach roughly 0.81 power for a medium effect (f² = 0.35).

Each child trained at home, supervised by a caregiver, for 16 weeks. CogniFit’s personalized program delivered short sessions of 12–15 minutes, with participants asked to complete at least 3 sessions per week and the difficulty adjusted to each child’s performance. Adherence was strong: participants completed an average of 40 sessions (40.23, SD 17.82), for a mean of 356 minutes of total training time (SD 221).

Cognitive control was measured before and after with a child-friendly version of the anti-saccade test, drawn from the DIGICOG executive-function battery and built around fish hidden inside four bubbles. Across 64 experimental trials, each child had to ignore a flashing star and respond only to a target fish, a deliberate test of the ability to suppress an automatic, reflexive glance in favor of a controlled response.

02What changed after 16 weeks

After 4.76% of outlier responses were discarded, a linear mixed model showed that participants responded significantly faster at post-test than at pre-test, 1,735 versus 1,861 ms, a 126 ms improvement, F(1, 1365) = 13.12, p < 0.001. The same speed-up appeared on the separate filler trials, where post-test responses were 143 ms faster (1,756 vs 1,899 ms).^[1]

The test itself behaved as intended. The harder anti-saccade condition produced both slower responses and higher error rates than the easier pro-saccade condition (16.8% vs 9.4% errors; χ²(1) = 24.46, p < 0.001), confirming the task discriminated the intended inhibitory-control demand. Across all conditions, 19 of the 26 children (73%) improved their overall performance after training, and no participants or their guardians reported any unintended effects from the cognitive training.^[1]

One result deserves a careful read. While reaction times fell, the analysis did not find a statistically significant change in error rates from pre-test to post-test (χ² < 1.5, p > 0.22). In other words, the measured benefit showed up as faster responding rather than as fewer mistakes, a meaningful distinction the authors report transparently.^[1]

03What this study does not show

The authors are explicit about the limits. With no control group and widely varying epilepsy types, they write that “it is difficult to attribute changes to the cognitive training intervention alone.” Pharmacological information was incomplete: 14 of the 26 guardians chose not to disclose which antiepileptic drugs their child was taking, or whether the child was on monotherapy or polytherapy, a relevant gap, because some treatments can themselves alter executive function.

Because the design captured a single pre–post change in one cohort, it cannot establish whether the gains persist over time or transfer beyond the trained task, and the 26-participant sample is small. The authors call for studies using “stratified epilepsy types and randomized controlled trials” to determine the effectiveness and generalizability of cognitive training in this population more conclusively. This was, by design, a feasibility study, an early signal, not a definitive efficacy verdict.

04Where this fits in the evidence

The finding extends a small but growing body of work. The authors point to controlled trials of executive-function training in attention-deficit/hyperactivity disorder, and to earlier computerized programs in epilepsy, including a randomized study of attention rehabilitation in focal seizures and memory-focused training delivered to adults after temporal lobe surgery. They note, however, that “the literature on cognitive training in the context of pediatric epilepsy remains insufficient,” positioning this report as one of the few feasibility tests in minors with the condition.

The study was led by José Luis Tapia of the Centro de Investigación Nebrija en Cognición (CINC) at Universidad Nebrija in Madrid, with Luis Miguel Aras of the Servicio Navarro de Salud in Pamplona and Jon Andoni Duñabeitia of UiT The Arctic University of Norway and Universidad Nebrija. It was published open-access (CC BY) in Children (2024, volume 11, article 484; DOI 10.3390/children11040484). The research was sponsored by ApoyoDravet with the help of a grant from Jazz Pharmaceuticals, Inc.; the authors declared no conflicts of interest, and state that the funder has no financial interest in the study.