Abstracts

RATIONALE: Objective: To compare language mapping using functional MRI (fMRI) and electrical cortical stimulation (ESM) in patients undergoing surgery for uncontrolled epilepsy
Rationale. FMRI is a non-invasive technique that has been shown to lateralize language as accurately as the intracarotid amytal procedure (IAP), but there are few data on its value for intrahemispheric localization. Previous O-15 PET activation studies (Bookheimer et al Neurology 1997) had shown a good correlation with ESM.
METHODS: We performed blood oxygenation level contrast fMRI, and ESM, in two patients with uncontrolled epilepsy. Both had left mesial temporal sclerosis on structural MR, and left speech dominance on IAP. MR was conducted on a conventional 1.5 Tesla scanner using a standard echoplanar sequence. The functional study employed a block design composed of six epoch cycles; each cycle consisted of an experimental task which alternated with a visual control task: Eight functional images were collected during each epoch of 32 seconds. The stimuli were presented through a MacIntosh computer to a rear projection screen positioned at the end of the scanner bed. For ESM, one patient had intraoperative mapping, and one extra operative subdural grids. Stimuli were presented on a laptop computer. Tasks for both fMRI and ESM included object naming and reading prose.
RESULTS: Patient 1 (intraoperative cortical mapping) had complete speech arrest when inferior frontal (Broca) region was stimulated and impaired naming or reading when superior temporal gyrus was stimulated. No effect was found when middle and inferior temporal gyri, the anterior portion of the superior temporal gyrus, and the inferior portion of pre- and post-central sulcus were stimulated. FMRI showed left inferior-mid frontal and mid-superior temporal activation on reading and naming tasks.
Patient 2 (chronic subdural grid recording) had complete speech arrest on superior temporal and fusiform gyrus stimulation. No effect occurred in the inferior portion of the lateral frontal region as well as the inferior portion of pre- and post-central sulcus. FMRI showed left inferior-mid frontal and mid-superior temporal activation.
In each case, resection including anterior temporal cortex, hippocampus and amygdala was performed without any postoperative language impairment.
CONCLUSIONS: We found incomplete overlap between fMRI and ESM language localization in two patients. ESM identifies regions crucial for specific functions, and fMRI, regions participating in them. Thus, absolute concordance should not be expected. fMRI may prove eventually to be a more conservative technique.
[Supported by: NINDS Intramural Program]