Visual memory after epilepsy surgery in children: A standardized regression-based analysis of group and individual outcomes

- We assess visual memory before and up to 2 years after childhood epilepsy surgery.
- The prospective longitudinal controlled study applies a standardized test battery.
- Analysis of variance replicates previous results: no change or small improvements.
- SRB analysis unmasks negative changes by correcting for practice effects.
- Postsurgical visual memory performance is disappointing in 50-70% of patients.

Visual memory is vulnerable to epilepsy surgery in adults, but studies in children suggest no change or small improvements. We investigated visual memory after epilepsy surgery, both group-wise and in individual children, using two techniques to assess change: 1) repeated measures analysis of variance (ANOVA) and 2) an empirically based technique for detecting cognitive change [standardized regression-based (SRB) analysis]. A prospective cohort consisting of 21 children completed comprehensive assessments of memory both before surgery (T0) and 6 (T1), 12 (T2), and 24 months (T3) after surgery. For each patient, two age- and gender-matched controls were assessed with the same tests at the same intervals. Repeated measures ANOVA replicated the results of previous studies reporting no change or minor improvements after surgery. However, group analysis of SRB results eliminated virtually all improvements, indicating that the ANOVA results were confounded by practice effects. Standardized regression-based group results showed that in fact patients scored lower after surgery than would be predicted based on their presurgical performance. Analysis of individual SRB results showed that per visual memory measure, an average of 18% of patients obtained a significantly negative SRB score, whereas, on average, only 2% obtained a significantly positive SRB score. At T3, the number of significantly negative SRB scores outweighed the number of significantly positive SRB scores in 62% of patients. There were no clear associations of clinical variables (including side and site of surgery and postsurgical seizure freedom) with memory outcome.The present analysis revealed that given their individual presurgical functioning, many children obtained disappointing results on some visual memory tests after epilepsy surgery. Comparison of the SRB analysis with ANOVA results emphasizes the importance of empirically based techniques for detecting cognitive effects of epilepsy surgery in childhood.