Abstracts

Rationale: While fMRI is a widely-accepted method for pre-neurosurgical localization of eloquent cortical functions, neuronavigated transcranial magnetic stimulation (nTMS) has emerged as an important tool for noninvasive pre-surgical functional language mapping. Specific to pediatric epilepsy surgery, language maps resultant from fMRI protocols and the language maps resultant from nTMS language mapping can yield discrepant results. This may in part relate to distinct physiologies measured by these techniques: Whereas fMRI measures the neurovascular response in the form of the BOLD signal, nTMS language mapping is reliant on the induction of a virtual lesion by repetitive cortical electrical stimulation. Accordingly, we test whether and how often nTMS and fMRI provide discrepant expressive language mapping results in children undergoing evaluation for resective epilepsy surgery. Methods: Patients (age 9-22 years) with intractable focal epilepsy, underwent both fMRI and nTMS for language mapping as part of the pre-surgical work-up (n=38). We compare lateralization of expressive language (left frontal, right frontal, or bilateral), detected by the two measures. In cases where nTMS and fMRI results were discordant, we then compared these results to invasive mapping either by electrical cortical stimulation (ECS) or intracarotid amobarbital injection (the Wada test). Results: Successful expressive language mapping by nTMS and fMRI was achieved in 30 patients, where 16/30 (53%) of cases yielded results discordant between the two modalities. fMRI was performed without sedation in all instances where the fMRI/nTMS results were concordant. Of those with discordant fMRI/nTMS results, three underwent passive fMRI language mapping with sedation. In 25% of patients with discordant results (n=4), fMRI revealed strictly unilateral expressive language (strictly left n=1, strictly right n=3), while nTMS showed bilateral expressive language. In 56% of patients (n=9), fMRI indicated bilateral cortical activation during expressive language, while nTMS revealed unilateral expressive language function confined to the left hemisphere (nTMS left, n=9). Notably, in the remaining patients with discrepant results (n=3; two with focal cortical dysplasia and one with Rasmussen's Encephalitis), unilateral findings from the two modalities were such that fMRI showed right hemispheric expressive language, and nTMS showed contrasting left hemispheric lateralization (nTMS left). In all three, seizures originated from the left hemisphere. Also in all three, Wada and ECS confirmed language function in the left hemisphere.  Conclusions: In children and adolescents with focal epilepsy, presurgical language mapping via fMRI and nTMS may yield discordant results between the two modalities. Particularly in instances when fMRI indicates right frontal expressive language localization, nTMS may more accurately localize expressive language as indicated by invasive pre-resection testing.  Funding: No funding