Reliability of language mapping with magnetic source imaging in epilepsy surgery candidates.

The external validity of a noninvasive language mapping protocol with magnetoencephalography (MEG) has been established through direct comparisons with invasive functional mapping techniques. This study examines the test-retest and interrater reliability of this protocol under realistic testing conditions in 21 epilepsy surgery candidates. Brain activation maps were obtained in the context of an auditory word recognition task and represented by temporally contiguous dipolar activity sources. Both the duration and strength of the associated magnetic flux were used as measures of the magnitude of regional brain activity. Hemispheric asymmetry indices based on these measures showed good interrater reliability and intraparticipant reproducibility. Similar findings were obtained with respect to the location of the geometric center of receptive language-specific cortex (Wernicke's) area in the dominant hemisphere. The results further support the adequacy of this MEG-based brain mapping protocol as a noninvasive tool for receptive language localization in epilepsy surgery candidates.