Abstracts

Rationale: The role of early brain lesions in the development of distributed language networks is debated. Studying patients with short durations of epilepsy and structural lesions may provide insight into the relationship between epileptogenesis and underlying pathologic substrate in language network perturbation. Methods: 21 patients with left hemisphere epileptic foci and structural lesions (11 males, 10 females; mean age 13.9 yr, range 4.3-42 yr; mean age seizure onset 12.7 yr, range 2-41 yr; mean duration 1.2 yr, range 0-2.7 yr; 17 right-handed, 4 left-handed/ambidextrous; 15 developmental tumors, 3 glioblastoma multiforme, 2 focal cortical dysplasia, 1 angioma bleed). Subjects performed an auditory word definition language paradigm (reverse speech control) using 3T BOLD EPI fMRI. Data were analyzed using SPM2 (individual maps at p<.001, uncorrected, >20 voxels per cluster). ROI laterality indices [LI=(∑Lvox -∑Rvox) / (∑Lvox+∑Rvox)] for inferior frontal gyrus (IFG), middle frontal gyrus (MFG), Wernicke’s area (WA) were assessed; LI<0.2 in IFG and WA, or LI<-0.2 in either IFG or WA defined atypical language. Results: 6 of 21 (29%) subjects had atypical language, reflecting rates higher than the expected adult rate of ~5% in the normal population (χ2=24.56, p<.001). For all subjects, mean IFG LI=0.30±0.63, mean MFG LI=0.30±0.43, and mean WA LI=0.53±0.43. Incidence of atypical language was not correlated with age (p>.05). Rate of atypical language may differ by seizure duration: 1 of 9 (11%) had atypical language with <6 months duration (mean subject age 11.8 yr; mean duration 0.2 yr), and 5 of 12 (42%) with 6 months to 2.7 years duration (mean subject age 15.5 yr; mean duration 1.9 yr); however, base rate of atypical language is small and was not significant (χ2=2.35, p=0.13). Conclusions: Subjects with short duration epilepsy have comparable language dominance profiles to chronic populations. Atypical dominance found in one-fourth of patients likely reflects a preexisting structural abnormality. However, given small base rates, we cannot entirely exclude the possibility that the epileptogenic process may have an effect on language representation within a short time frame.