Abstracts

Rationale: Previous work (Berl et al. 2005, 2014) demonstrates that atypical language dominance, particularly bilateral language, occurs in frontal areas. Epilepsy populations differ significantly from TD populations, displaying a 25-30% incidence of atypical language dominance. We challenge the assumption that atypical language is language reorganization but rather may reflect working memory demands associated with the task that is more demanding for patients. We test this by examining the functional differences induced by left-hemisphere focus epilepsy on language dominance under conditions in which verbal working memory demands increase. Methods: Nineteen pediatric patients with left hemisphere  epilepsy (LHE), mean age: 11.6 years, and 19 pediatric typically developing (TD) controls, mean age: 11.8 years, completed neuropsychological testing and a panel of EPI BOLD 3T block design language tasks. Patients had no MRI abnormalities in fronto-temporal areas. Two language measures, the CELF-IV Core Language Composite (CLC) and the Expressive One Word Picture Vocabulary Task (EOWVT), and two working memory/executive functioning (EF) measures,  the WISC-IV Digits Span Backwards Scaled Score (DSB), and parent rating on the Inattention scale measure from the Dupaul ADHD Parent Rating Scale (ADHD-I) were used to correlate with task activation.The fMRI auditory language tasks were designed to manipulate and increase working memory and language demands. A well-known language task was used; auditory category decision task (AUDCAT), which is an adapted verbal fluency paradigm. A classic working memory paradigm was used; Nback with both a 2-back (2B) compared to a 0-back condition using single words as targets. Working memory and language demands were increased simultaneously with a SemBack task that used a traditional Nback design but the respondent had to decide if the word before was the same semantic category as just heard. Imaging processing and statistical analyses were conducted in SPM12 and VBM12 Toolbox. A general linear model was used to examine the group, neuropsychological performance, and interactions on task activation at p=.001 (uncorrected, k>20 voxels). Results: The LHE patients were typically functioning such that they did not differ from TD on neuropsychological functioning, both groups fell in the average range overall. Both groups also displayed the expected activation of left lateralized language for AUDCAT and bilateral frontal-parietal for working memory tasks (2Back and SemBack), and two-sample t-tests revealed minimal  group differences in activation. However, there were significant interactions of group when examining the association between neuropsychological performance and activation. Patients with worse cognitive skills (vocabulary or EF) recruited the right hemisphere, particularly frontal areas (Right IFG/MFG) across tasks (Figure 1: 2Back; Figure 2: SemBack). This was not the case for controls. Conclusions: In a left-hemisphere epilepsy group that performed similarly to TD controls on neuropsychological performance and fMRI task activation, there is greater right hemisphere activation with increased verbal working memory demands for those with worse working neuropsychological skills. This suggests that bilateral language activation observed in patients is not necessarily language reorganization per se but may represent executive functioning demands. Perhaps right hemisphere activation is compensation for verbal working memory weaknesses and performance is equalized by increased activation in the other hemisphere. Funding: K23 from NINDS 5K23NS065121-03