Abstracts

Rationale: In patients with intractable epilepsy, accurate pre-surgical mapping of eloquent cortex is often needed before surgical resection can be performed. Current methods of functional brain mapping, including functional magnetic resonance imaging (fMRI) and the gold-standard electrocortical stimulation mapping (ESM), require a high level of the patient's attention and active cooperation or a significant portion of time to generate surgically useful maps. In this study, we employed an innovative and less time- and cooperation- demanding modality, real-time functional mapping (RTFM), to evaluate its sensitivity and specificity compared to ESM and fMRI for eloquent language localization. Methods: 21 patients with intractable epilepsy, median age 18.5 (SD 10.61 years), 10m/11f completed localization paradigms for expressive and receptive language function with each mapping modality, fMRI, ESM, and RTFM. Analysis consisted of determining sites of functional significance and then calculating specificity and sensitivity of RTFM in comparison to ESM and fMRI. Additional RTFM analysis included the use of a "next-neighbor" approach. Results: Our preliminary comparison of RTFM with ESM showed a 25% sensitivity and 91% specificity in determining language-specific cortex. Similarly, the preliminary comparison of RTFM with fMRI showed a 29% sensitivity and 92% specificity. A "next-neighbor" approach improved the sensitivity at the cost of a reduced specificity, resulting in 58% sensitivity and 75% specificity in the RTFM to ESM comparison, and 77% sensitivity and 75% specificity in the RTFM to fMRI comparison. Importantly, RTFM results during different language activation tasks demonstrated variable sensitivity and specificity values when compared with both ESM and fMRI. RTFM sensitivity obtained with a verb generation task was the closest to ESM-observed results. Conclusions: Although RTFM is not as sensitive as ESM, it does have 91% concordance with ESM in determining areas of language-unrelated activity. This, paired with the use of the "next-neighbor" approach, which provides greater sensitivity (58%), allows RTFM to be used for fast and effective detection of eloquent language areas. Therefore, we believe that RTFM could be most effectively utilized for determining which cortical regions should be focused on during ESM. Specifically, this may allow ESM to be done more efficiently, requiring less time commitment from the patient. The use of RTFM in evaluating patients can provide clinicians with additional information in directing the course for functional mapping procedures and surgical planning.