Abstracts

Rationale: Inter-hemispheric differences in both function and structure of the frontal lobe have been correlated with language laterality, but it is not clear whether there is a structural connectivity correlate for hemispheric language dominance. To answer this question, diffusion tensor imaging (DTI) (a measure of anatomical connectivity), and fMRI activity in Broca’s Area were compared with results of the Wada procedure for determining language laterality in patients prior to surgical treatment for epilepsy. Methods: Diffusion-weighted and functional echo-planar images were obtained from 15 patients with language Wada results (12 left-, 3 right-hemisphere dominant). Broca’s area was defined as BA44 and BA45 from a probabilistic cytoarchitectonic template mapped to single-subject imaging space (Eickhoff et al., 2005). Two separate laterality indices [(L-R)/(L+R)] were computed in each participant using: 1) DTI: the number of high anisotropy arcuate fasciculus fibers connecting Broca’s Area to parietotemporal regions, and 2) fMRI: the number of suprathreshold active voxels in Broca’s Area during a battery of language production tasks obtained during fMRI. Results: The results of the invasive Wada test were used as the gold standard. DTI laterality indices correctly classified 12 of 15 patients. Those misclassified by DTI included 2 left and 1 right hemisphere Wada patients. fMRI laterality indices correctly classified 14 of 15 patients according to their Wada results. The 1 patient misclassified by fMRI had left hemisphere language according to the Wada procedure. A total of 12 patients had structural and functional laterality indices that were both in agreement with the Wada results (Figure 1). Conclusions: We compared the accuracy of fMRI and DTI of Broca’s Area for determining language laterality. While fMRI has been shown to be highly correlated with the Wada procedure (Binder et al 1996), the present results suggest that DTI may be a reliable alternative when fMRI is not available. These data represent the first attempts to lateralize language function based on average anisotropy of fiber pathways; we anticipate that with more refined techniques of fiber selection, the classification of hemispheric dominance may be more accurate, and hold promise for replacing the Wada test.