Abstracts

Rationale: Impaired cognition (1) and language ability (2-4) predict behavior problems in children. This study examined if the number of psychiatric diagnoses (multiple, single, and none) is related to measures of intelligence and language in children with localization related epilepsy with complex partial seizures (CPS) and childhood absence epilepsy (CAE). Poor working memory, a consistent finding in pediatric epilepsy (5), plays a role in intelligence (6), language (7), and behavior (2) in healthy children (HC). This study, therefore, also investigated if working memory mediates the hypothesized association among the comorbidities. Methods: 105 CPS and 79 CAE patients, aged 6- 15 years, were administered The Schedule for Affective Disorders and Schizophrenia for School-Age Children, Present and Lifetime Version (KSADS), Wechsler Intelligence Scale for Children, Digits Backward (DB), and Test of Language Development-2. Subjects were classified as having multiple, single, or no KSADS diagnoses. The relationship of Verbal IQ (VIQ), Performance IQ (PIQ), and Speech Language Quotients (SLQ) of epilepsy groups with number of KSADS diagnoses were examined using MANCOVAs. To determine if the epilepsy groups differed in these relationships, the interaction term epilepsy group x KSADS diagnosis was examined. Since these interaction terms were not significant, the epilepsy groups were combined for the remaining analyses. Significance of the main effect of KSADS diagnoses was followed up with post hoc ANCOVAs of VIQ/PIQ/SLQ scores with the number of KSADS diagnoses groups. Mediation by DB on the effect of KSADS diagnoses on IQ/SLQ was also examined. Results: VIQ (F (2,181) 5.09, p = .007), PIQ (F (2,181) 3.07, p = .05), and SLQ scores (F (2,163) 4.94, p = .02) were significantly associated with number of KSADS diagnoses. The children with multiple diagnoses had significantly lower IQ (p = .01 - .006) and SLQ scores (p =.06) than those with no diagnosis. Other than VIQ, the multiple diagnosis group had significantly lower PIQ (p = .01 -.006) and SLQ scores (p = .006) than the single diagnosis group. The IQ/SLQ scores were not significantly different between the single and no diagnosis groups. The effect of KSADS diagnostic group on VIQ//PIQ/SLQ was less when DB was included as a predictor in the model (F .71- 2.49, p .18 - .09) compared to the model without DB (F 3.07- 5.19, p .05 - .007). Hence, DB mediated the effect of VIQ, PIQ, and SLQ scores on the number of K-SADS diagnoses.Conclusions: The mediating role of working memory in the association among severe psychopathology, low IQ, and impaired language skill suggests the need for treatment studies of cognitive remediation focused on working memory in pediatric CPS and CAE.