Longitudinal Bayesian analysis of lobe‐specific cognitive outcomes after pediatric epilepsy surgery

OBJECTIVE: To examine the impact of the operated brain region and clinical-surgical variables on intellectual performance in children and adolescents undergoing neurosurgery for pharmacoresistant epilepsy. METHODS: Fifty-one participants were classified according to surgical site: temporal lobe (TL; n = 23), frontal lobe (FL; n = 9), posterior cortex (PC; n = 9), and hemispherotomy (H; n = 10). Neuropsychological assessment was performed using age-appropriate versions of the Wechsler Intelligence Scales (Wechsler Intelligence Scale for Children/Wechsler Adult Intelligence Scale [WISC/WAIS]). Bayesian linear mixed models compared intelligence quotients and scale indices across three time points: preoperative (T0), 6-month postoperative (T1), and 24-month postoperative (T2). For each outcome, two models were fitted: a simple model and a multiple model adjusted for seizure frequency and epilepsy duration. Estimates were derived from posterior distributions with 95% credibility intervals. Analyses were conducted in R using MCMC sampling, with convergence assessed by trace plots and the Gelman-Rubin diagnostic. A change of ≥10 points was considered a clinically relevant improvement, whereas a decrease of ≥5 points indicated a relevant decline in longitudinal within-group comparisons. Results reflect group-level trajectories rather than individual outcomes. RESULTS: At T1, the H group showed declines in Full-Scale Intelligence Quotient (FSIQ), Verbal Comprehension Index (VCI), Perceptual Organization Index (POI), Working Memory Index (WMI), and Processing Speed Index (PSI). At T2, the FSIQ, POI, and WMI returned to near baseline, whereas VCI and PSI remained reduced. The TL group showed declines in VCI, WMI, and PSI at T1, with a persistent reduction in VCI at T2. FL and PC groups maintained intelligence indices close to baseline at both postoperative assessments. SIGNIFICANCE: Epilepsy surgery preserved intellectual functioning, particularly in FL and PC procedures. However, persistent VCI decline in H and TL groups suggests effects on verbal abilities associated with temporal lobe involvement. Sustained PSI reduction after hemispherotomy may indicate decreased cognitive efficiency and slower information processing.