34. Structural connectivity of eloquent speech areas defined by direct cortical stimulation mapping

Abstract

Background and aims The anatomical borders of eloquent speech areas, e.g. Broca's area, have been vaguely defined and exhibit high inter-subject variability (Amunts and Zilles, 2012), mainly due to inconsistent morphology of the inferior frontal gyrus (IFG) (Ebeling et al., 1989) and small-scale incongruity of functional segregation with respect to anatomical landmarks (Garrett et al., 2012; Quinones-Hinojosa et al., 2003). However, knowing the exact boundaries of the eloquent cortex became clinically important in preventing postoperative deficits after brain surgery (Garrett et al., 2012; Hollon et al., 2015). The awake electrical stimulation mapping (ESM) has become the preferred method to delineate the speech cortex, as the current non-invasive alternatives lack specificity and spatial resolution (Garrett et al., 2012; Hollon et al., 2015). Since the structural connectivity-based parcellation using diffusion tensor imaging (DTI) has been shown to closely relate to the cytoarchitectonics in the IFG (Klein et al., 2007), we tested whether the location of eloquent cortex could be better predicted by preoperative DTI rather than by its relation to gyrification. Methods Seventeen patients (8 males, median age 33, range 18–63) suffering from primary brain tumor underwent pre-operative 3T MRI scanning using a high resolution T1 MPRAGE and a single shell diffusion weighted imaging (DWI) sequence (60 directions, b =1000 s/mm 2 , resolution 2×2×2mm), and intraoperative direct cortical electrical stimulation with a bipolar Ojemann electrode during awake language mapping. After standard preprocessing, probabilistic fiber tracking (PROBTRACKX in FSL (Behrens et al., 2007)) was carried out using 5mm spheres as cortico-subcortical seeds placed at the intraoperatively defined positive stimulation sites. The resulting cortical projections were then overlaid onto the individual cortex surface (Freesurfer) and compared visually and using binary logical operations. Results A variety of stimulation-positive functional sites was detected, producing anomia, counting, speech and motor arrest. Anomia sites were found both in the pars triangularis and pars opercularis of the inferior frontal gyrus (also in individual patients). Even opercular anomia sites followed the connectivity pattern of triangular anomia sites and had stronger connections to the rostral parts of the mesial prefrontal cortex and pre-supplementary motor area (pre-SMA) than to the precentral gyrus and SMA proper, as opposed to opercular speech or counting arrest sites which also connected more strongly to more posterior parts of the brain, mostly posterior temporal cortex. Thus, even adjacent sites on the same gyrus that differed in function showed dissociable structural connectivity. Conclusions Our results suggest that structure–function relationships are more evident between function and structural connectivity than between function and gyrification. Acknowledgement This study was supported by the German Research Foundation, the European Union, and the German Academic Exchange Service (DAAD).