Abstracts

Rationale: Matsumoto et al. (Brain 2004) previously reported the results of cortico-cortical evoked potentials (CCEPs) showing reciprocal connections between Broca’s and Wernicke’s areas, as well as connectivity with basal temporal cortex, supporting the view that the human language system consists of a broadly distributed network. We further investigated the functional connectivity of the language system by stimulating multiple sites within perisylvian areas and anterior and basal temporal areas. Methods: Four patients with implanted subdural electrodes were studied. Electrocortical stimulation mapping (ESM) was performed for clinical purposes to identify areas with putative language function. CCEPs were elicited by single pulse stimuli between pairs of electrodes in broadly defined anterior and posterior language areas, mouth motor area, and/or anterior temporal area in sites with and without ESM language disruption. The resulting evoked potentials were averaged across 100 pulses. Results: Stimulation at sites within 1-2 cm of each other elicited CCEPs with strikingly different anatomic distributions (4 of 4 patients), particularly with stimulation in the anterior temporal lobe. In three patients, for example, stimulation in posterior superior temporal gyrus, near primary auditory cortex, elicited more local responses, whereas stimulation of nearby sites in mid- and posterior temporal cortex elicited more widespread responses, suggesting greater divergence in downstream auditory processing pathways. Similar to previous studies, CCEPs in Wernicke’s area were obtained by stimulating in Broca’s area (2 of 3 patients), and reciprocal CCEPs were obtained in Broca’s area by stimulating in Wernicke’s (1 of 3 patients). CCEPs were also elicited in mouth motor cortex by stimulating in Broca’s area (2 of 3 patients) and in Wernicke’s area (1 of 3 patients). In one of them, the CCEPs between mouth motor cortex and Broca’s area were reciprocal. Two patients had electrode coverage of fusiform gyrus. In one of them, reciprocal connections between Broca’s area and fusiform gyrus were observed, while in the other, reciprocal CCEPs were elicited between fusiform gyrus and inferior temporal gyrus, and a non-reciprocal CCEP was elicited from superior temporal gyrus to fusiform gyrus. In one patient, stimulation at atypical ESM+ language sites in anterior temporal lobe elicited responses in ESM+ typical perisylvian areas. Conclusions: The anatomical distribution of CCEPs appears to be remarkably dependent on variations in the site of stimulation within subregions of language cortex. These results suggest that although the cortical networks supporting human language function are widely distributed, their functional connectivity is not random and potentially can be segregated into different pathways using CCEPs. Future studies with more comprehensive CCEP mapping, and correlation with surgical outcomes and other anatomic and electrophysiologic measures of connectivity, are needed to verify these conclusions and to explore the potential clinical utility of CCEPs in functional mapping.