Abstracts

Rationale: Recently the importance of the mesial temporal structures in naming has been reappraised since the patients with language-dominant left temporal lobe epilepsy (TLE) showed a decline in naming when these structures were resected with sparing the lateral temporal naming sites defined by electrical stimulation (ES) (Hamberger et al., 2010). We hypothesized that resection of the basal temporal language area (BTLA), originally coined by L ders et al. (1986), and/or its connection is relevant to naming decline. We aimed to delineate its anatomo-functional property, connectivity and impact on postoperative language outcome.Methods: We studied 12 TLE patients who underwent invasive presurgical evaluation with subdural grid implantation over the language-dominant, left lateral and basal temporal areas (8 with with hippocampal sclerosis). Cortical language mapping was performed with 50 Hz ES with 5-6 language/semantic tasks. In 8 patients, cortico-cortical connectivity from BTLA was evaluated by cortico-cortical evoked potentials (CCEPs) (Matsumoto et al., 2004). Precise electrode location was identified by MRI taken after implantation in individual brains, then co-registered into the MNI standard space for group analysis. 10 patients underwent subtemporal amygadalo-hippocampectomy (AH) leaving the white matter pathways for the basal temporal areas intact (Miyamoto et al., 2004). 5 patients also had tailored basal temporal corticectomy.Results: 2-11 BTLA electrodes (median 4) were identified per patient (mean 41 mm from the temporal pole). Positive rate of language impairment was the highest in the fusiform gyrus (FG, 75%), followed by the inferior temporal gyrus (ITG, 45%) and the parahippocampal gyrus (PHG, 33%). In language mapping, impairment was seen among many tasks (both expressive and receptive) in visual and auditory modalities: the highest in picture naming (90%), followed by word-picture matching task (83%), paragraph reading and Kanji word reading (~60%), and verbal command and Kana word reading (30-40%). Core BTLA ( 4 task impairment) clustered in the anterior FG & ITG (mean MNI coordinate = -43, -15, -41). CCEPs revealed BTLA connectivity to the ipsilateral anterior temporal language area (anterior STG & MTG), posterior language area (posterior STG & MTG), and ventrolateral prefrontal area (mainly Brodmann s area 45). Regarding language outcome with subtemporal AH, 1 patient with tailored corticectomy of 1 core BTLA electrode (anterior FG) showed moderate anomia and a difficulty in reading and writing Kanji word. 3 patients with tailored corticectomy of non-core BTLA (1-2 electrodes, anterior PHG) did not show any evident anomia. In the remaining 6 patients, all BTLA electrodes were spared and no subjective language deficits occurred.Conclusions: Anatomical location (consistent with activation loci of semantic judgment tasks), patterns of functional impairment (higher positive rate in semantic tasks) and connectivity (a part of the ventral language pathway), and language outcome after core BTLA resection indicate a major role of BTLA in the semantic aspect of language function.