Abstracts

Rationale: BECTS is a common childhood epilepsy syndrome traditionally assumed to have a benign course. Recent studies have shown that some cognitive functions, especially language, may be impaired (Northcott, et al. 2007). Centrotemporal spikes (CTS) in BECTS originate from perisylvian areas near regions involved in language function; CTS may disrupt language functions. We previously found significant differences in neural circuitry for language between BECTS patients and age-matched controls and noted lower levels of language and cognitive skill in BECTS suggestive of BECTS-related remodeling of language networks that may underlie observed lower language skills. However, quantitative measures of CTS have not previously been related to neuroimaging data to assess their specific effects on the neural circuitry for language. In this study, we related counts of CTS to fMRI activation during 3 language tasks.Methods: Participants were 15 BECTS patients (ages 6-11, 3 females). Scanning was performed at 3T with 3 fMRI language tasks: a semantic decision vs. tone decision task shown to activate left-lateralized frontal/temporal language regions (Binder, et al. 1997), a story processing vs. tone listening task that engages bilateral temporal regions (Schmithorst, et al. 2006) and a prosody discrimination vs. semantic judgments task that typically activates right frontal regions (Gandour, et al. 2003). CTS were counted by visual analysis and noted as originating from left or right centrotemporal regions in each patient s clinical EEG recording most recent to the fMRI procedure (duration range 25-47 minutes). Frequency of CTS ranged from 0.97-39.5/min (mean 9.0/min). For each fMRI task, a general linear model and paired t-test was implemented to generate activation maps for each task for each patient. Then, two group composite maps were generated for each task to identify regions where activation was correlated with counts of left and right CTS, respectively. These maps were thresholded at z>6, cluster size 30 voxels, resulting in p>0.05 corrected significance. Results: There was no clear pattern of correlations between CTS counts in the dominant hemisphere for each task. CTS in the non-dominant hemisphere for each task systematically influenced patterns of activation. For semantic decision, there was a significant negative correlation between count of right-sided CTS and activation in right-hemisphere homologues of inferior frontal and superior temporal language regions (Fig.1A). For story processing, there was a significant positive correlation between count of right-sided CTS and activation in left middle temporal cortex (Fig.1B). For the prosody task, there was a significant negative correlation between the count of left-sided CTS and activation in left middle frontal cortex (Fig.1C).Conclusions: Our results suggest that while CTS may not directly influence the dominant language circuitry, they systematically reduce BECTS patients ability to use the contralateral language homologues often engaged during challenging language tasks. This may underlie some of the subtle language problems observed in BECTS.