Authors :
Presenting Author: Freya Prentice, BA, MSc – University College London
Maria Eriksson, PhD – University College London
Patricia Martin-SanFilippo, DClinPsy – Great Ormond Street Hospital for Children
Sara Shavel-Jessop, DClinPsy – Great Ormond Street Hospital for Children
Frederique Liegeois, PhD – University College London
Torsten Baldeweg, MD – University College London
Rationale:
Evidence for material-specific memory deficits in children with temporal lobe epilepsy (TLE) is mixed (Kahana Levy et al., 2021), perhaps reflecting reduced lateralization of children’s mesial memory networks. In adults with left and right TLE, language impairments have been shown to affect verbal memory performance and after controlling for language, material specific patterns can disappear (Hermann et al., 1992). The aim of this study was to (1) examine material-specificity in children with TLE; and (2) investigate the contributions of hippocampal integrity and language impairments to verbal associative memory.
Methods:
We included 70 children assessed at a pediatric hospital (London, UK) between 2002-2022 as part of the epilepsy surgery program with (1) a clinical diagnosis of TLE; and (2) preoperative assessments on the Children’s Memory Scale (CMS) and the Clinical Evaluation of Language Functions (CELF). Children were classified as having a language deficit or not (CELF receptive language composite ≤ 85). Segmentation of hippocampi was completed using SynthSeg on the preoperative T1-weighted scans, and errors were manually corrected using ITK-snap. Hippocampal volumes were corrected for total intracranial volume. We examined material-specificity using a 2 (side: left vs right) x 2 (receptive language: impaired vs non-impaired) x 2 (material: verbal vs visual) mixed design ANOVA. To identify predictors of verbal memory we used a multiple linear regression with clinical variables and hippocampal volumes as predictors and the Delayed Word Pairs subtests of the CMS as the outcome measure.
Results:
We found an interaction between material and language impairment (F(1,66)=15.99, p< .001; Figure 1); with an effect of language impairment on verbal (p< .001) but not visual memory (p=.906). Post-hoc pairwise comparisons demonstrated lower verbal than visual memory in those with a language impairment regardless of the side of epilepsy (all p< .05). Our regression model explained 40% of the variance in Word Pair performance (