Free Communications: Neuroimages MEG and EEG

Abstract

Abstract Functional Magnetic Resonance Imaging- and Magnetoencephalography-based Brain Mapping for Identifying Language Functions and Its Validation by Electrocortical Stimulation. 1 Kyousuke Kamada, 1 Kensuke Kawai, 2 Yoshitaka Masutani, 2 Shigeki Aoki, and 1 Takaaki Kirino ( 1 Department of Neurosurgery, the University of Tokyo, Tokyo, Japan 2 Radiology, the University of Tokyo, Tokyo, Japan ). Purpose: Functional magnetic resonance imaging (fMRI) detects changes in cerebral blood flow and has been applied to the identification of the dominant hemisphere in language functions. Most of the fMRI-language studies disclose frontal activation by various lexico-semantic tasks. Magnetoencephalography (MEG) directly detects neuronal activity and records intracellular electric current flow in the brain. It localizes semantic responses peaking at approximately 400 msec after word presentation (late responses) in the temporoparietal regions, and the MEG responses are suspected to be strongly related to the receptive language function. We have, therefore, utilized these two modalities to independently visualize the expressive- and receptive-language functions. The successful mapping results by noninvasive techniques were verified by electrocortical stimulation. Methods: We investigated two patients with intractable epilepsy; a right-handed patient with left frontal epilepsy (Patient 1) and a left-handed patient with bilateral temporal lobe epilepsy (Patient 2). Neurological examinations disclosed no language deficits in both patients. They were asked to generate verbs related to acoustically presented nouns (verb generation) and to categorize visually presented words as abstract or concrete (A/C categorization) for fMRI and MEG investigations, respectively. The averaged magnetic signals were digitally filtered between 0.1 and 30 Hz. Significant MEG deflections were visually identified on the basis of the root mean squared fields of more than 10 sensors in the fronto-temporal (FT) or temporo-occipital (TO) regions. Locations and dipole moments of equivalent current dipoles were calculated every 2 msec from 250 to 600 msec after the stimulus was started, using the single equivalent dipole model. After subdural electrodes were placed over the hemisphere, the results of fMRI and MEG, and postoperative CT scan demonstrating the electrode positions were co-registered and superimposed on the anatomical MRI by maximizing the mutual information of all data sets with Affine transformation. The noninvasive functional mapping results were validated by 60Hz bipolar stimulation with the subdural electrodes. During the stimulation, patients were instructed to keep reading words, naming pictures, or generating verbs. Results: FMRI demonstrated activations in the inferior (IFG) and middle frontal gyri (MFG) of the suspected dominant hemisphere. While MEG disclosed highly concentrated semantic dipoles in the superior temporal (STG), supramarginal (SmG) and fusiform gyri (FuG) of the dominant hemisphere, FMRI and MEG clearly indicated language dominancy on the left hemisphere in Patient 1 and on the right in Patient 2. During locus validations, the electrocortical stimulation to the IFG and STG consistently induced speech arrest and sensory aphasia including dyslexia, respectively. In addition, the stimulation to the right FuG in Patient 2 revealed pure dyslexia with no suppression in the picture-naming task. Since the cortical stimulation to the MFG caused no neurological symptom, the MFG activation on fMRI did not correspond to the results of the stimulation. These findings suggest that the MFG might play roles in speech association or working memory and might be outside the critical language area compared to the other activated areas on fMRI and MEG. Conclusion: Noninvasive functional mapping techniques are powerful tools to identify language dominancy. IFG activation on fMRI, as well as STG, SmG, and FuG dipole clusters on MEG, are highly reliable for the detection of the critical language areas. MFG activation on fMRI, however, needs more careful validation and interpretation. It is necessary to modify and improve the lexico-semantic tasks of functional brain mapping for appropriate visualization of the critical language areas.