Abstracts

Rationale: TSC is a multisystem genetic disorder. Brain abnormalities include cortical tubers, which can be epileptogenic. Nearly 90% of patients with TSC have epilepsy and 2/3 of them are refractory to pharmacotherapy. Epilepsy surgery can be considered and requires a presurgical workup that may involve assessment of language lateralization. Noninvasive techniques, such as functional MRI (fMRI) and magnetoencephalography (MEG) are now used to assess language lateralization in epileptic patients. This is the first study to investigate language lateralization in a group of TSC patients using MEG and fMRI. Methods: Fifteen patients with TSC performed a lexico-semantic decision task during MEG and fMRI recordings. MEG data was acquired from a 306-channel whole-head MEG system. Standard BOLD fMRI and anatomical MRI scans were performed on a 1.5T scanner. MEG data was co-registered with structural MRI using fiducial points from a 3D digitizer. The lexico-semantic task consisted of a sequential visual presentation of 160 words. Patients were asked to respond manually by pressing the right button on a response pad if the presented word was concrete (ex: apple) and the left button if the word was abstract (ex: freedom). A digital band pass filter of 0.1-40Hz was applied off-line. Data were averaged across all 160 trials without distinction between concrete and abstract words. Minimum-norm estimates (MNE) were computed using MNE software (version 2.5, Hamalainen, 2006) allowing identification of cerebral generators of language evoked fields (EF) in each patient. Laterality indices (LI) were computed by comparing MNE amplitude between 250 and 550 ms in the inferior frontal, middle and superior temporal, inferior parietal and supramarginal gyri from both hemispheres. LI > 0.10 was interpreted as left, LI < -0.10 as right, and LI between -0.10 and 0.10 as bilateral language dominance. MEG results were compared to fMRI data acquired while the patients performed the same task. fMRI data were analyzed using FSFAST software. Results: MNE showed language cerebral activations starting in the middle and superior temporal gyri traveling progressively to the inferior frontal gyrus in all patients. Inferoparietal and supramarginal activations were also measured in most patients, although these activations were weaker compared to frontotemporal responses. A perfect concordance between MEG and fMRI LI was obtained for all patients. As also found with fMRI, fourteen patients had left language lateralization on MEG results (LI between 0.11 and 0.44) and one patient showed a bilateral language dominance (LI = 0.05). Conclusions: This is the first neuroimaging study to investigate language lateralization in a group of patients with TSC. A perfect concordance between MEG and fMRI results for language lateralization was obtained in all 15 patients using a lexico-semantic task. Compared to fMRI, MEG provides temporal information that helps better understand cerebral language activations. MEG may constitute an alternate means of safely investigating language lateralization in TSC patients.