Object: In this study the authors evaluated the sensitivity and selectivity of a noninvasive language mapping procedure based on magnetoencephalography (MEG), for determining hemispheric dominance for language functions.
Methods: Magnetic activation profiles of the brain were obtained from 100 surgical candidates (age range 8-56 years) with medically intractable seizure disorder by using a whole-head MEG system within the context of a word recognition task. The degree of language-specific activity was indexed according to the number of consecutive sources (modeled as single, moving current dipoles) in perisylvian brain areas. Only activity sources that were observed with a high degree of spatial and temporal overlap in two split-half data sets were used to compute the MEG laterality index. Independently, all patients underwent Wada testing for the determination of hemispheric dominance for language. Independent clinical judgments based on MEG and Wada data showed a high degree of concordance (87%). Magnetoencephalography laterality judgments had an overall sensitivity of 98%, but a lower selectivity of 83%, which was due to the fact that MEG detected more activity in the nondominant hemisphere than was predicted based on the Wada test. A number of objective criteria were derived based on this large patient series to ensure data quality and bolster the clinical usefulness of MEG for language mapping.
Conclusions: Although the availability of MEG is still limited across epilepsy surgery centers, this study method may be substituted for the Wada procedure in assessing hemispheric dominance for language in select cases.