Abstracts

Rationale: Determination of language laterality is a crucial step for pre-surgical planning in refractory epilepsy. Wada test is still considered the gold-standard technique for determining the dominant language hemisphere. Its use has been declining, with centers preferring non-invasive methods such as Functional Magnetic Resonance (fMRI) and Magnetoencephalography (MEG). Reports have shown relatively high indices of laterality concordance between MEG and WADA test, but MEG remains underutilized. We evaluated the concordance in language laterality between MEG and fMRI in patients with refractory epilepsy undergoing work-up in our NAEC level 4 epilepsy center.

Methods: Subjects included had a MEG scan and an fMRI as part of their epilepsy surgical evaluation. MEG scan was performed using a 306-channel MEGIN Triux-Neo MEG system. We obtained receptive language with auditory stimuli. All data were analyzed with MEGIN DANA software and mapped using equivalent-current dipole (ECD) fitting (figure 1). We defined the MEG laterality index (LI) as: Left Dominant (LD), Right Dominant (RD), or Bilateral Dominance (BD). We compared it with qualitative reports of their respective fMRI. Our fMRI procedure includes a battery of language tasks that includes both receptive and expressive language tasks. fMRI laterality was considered bilateral when language had no clear hemispheric dominance. We classified laterality level of concordance as: Concordant (C), Discordant (D), or Partially Concordant (PC) when one of the methods evidenced bilateral dominance and the other unilateral dominance.

Results: Forty-one adult patients with refractory focal epilepsy were studied (Table 1). Language testing generated significant bilateral ECD in 44% of subjects. The laterality dominance percentages were LD in 81% (MEG) and 83% (fMRI); RD in 17% (MEG) and 12% (fMRI); BD in 2% (MEG) and 5% (fMRI). The methods were C in 35 subjects (85.4%), PC in 3 subjects (7.3%), and D in 3 subjects (7.3%). We did not find any statistically significant difference between the different concordance groups and clinical variables.

Conclusions: We found 85.4% complete concordance between both methods. Our concordance agrees with previous reports comparing MEG with fMRI or WADA, reinforcing its potential in non-invasive language lateralization. We found 7.3% for D and the same for the PC group. Besides the fact that fMRI and MEG use distinct approaches, our MEG scans utilize one paradigm, while fMRI combines different paradigms for data analysis and results. This difference could account for some discrepancies when comparing both methods. Overall, our findings have shown reliable concordance between MEG and fMRI, reinforcing the importance of MEG in presurgical planning for epilepsy surgery. Its role may be as a complementary or alternative tool for establishing language dominance. Standardization of acquisition protocols and analysis are imperative. The routine combination of fMRI and MEG may ultimately lead to a significant improvement in its accuracy.

Funding: Please list any funding that was received in support of this abstract.: No financial support was received for this abstract.