Abstracts

Rationale: Currently, electrical stimulation mapping (ESM) is the gold standard for identifying eloquent cortex which should be spared resection during epilepsy surgery. However, there are many limitations to this technique, particularly the risk of after discharges and seizures. ESM can be time-consuming, requires excellent patient cooperation, and is not always well tolerated by patients. Here, we apply the technique of electrocorticography (ECoG) to analyze cortical neurophysiology during a battery of five language tasks that mirrors our ESM paradigm. We hypothesize that cortical language sites identified through task-evoked high gamma will inform localization of sites identified by ESM. Methods: We developed a battery of five language tasks that mirror our ESM paradigm. Specifically, our language tasks involve visual naming, word reading, auditory repetition, auditory naming, and auditory comprehension. We used a set of fifty stimuli taken from the revised Snodgrass and Vanderwart object pictoral set (Rossion and Pourtois; 2004) and the stimuli are matched across the five tasks. Our auditory naming and comprehension stimuli were taken from the Hamberger sets when matched stimuli were available (Hamberger and Seidel; 2003 and Hamberger, Friedman, and Rosen; 1996). During our analysis we focused on changes in high-gamma activity (70 – 150 Hz), as high-gamma activation has been previously shown to robustly track single trial cortical activity and correlates with neural population firing rates and fMRI BOLD responses. Our battery was administered to patients who also underwent ESM for the purposes of their clinical care to identify eloquent cortex. Our ESM paradigm uses similar language tasks of continuous speech, visual naming, auditory naming, and auditory comprehension. Functional language maps created using ESM and high-gamma ECoG were compared. Results: Of the five language tasks used for our high-gamma ECoG based functional language maps, the tasks of visual naming, auditory naming, and auditory comprehension were most comparable to the ESM functional language map. In addition there were more electrodes that had significantly greater high-gamma activity during the language tasks than electrodes that were identified by ESM. Conclusions: High-gamma ECoG is a clinically useful tool to complement stimulation based language mapping. By analyzing high-gamma ECoG during a battery of five language tasks we can further improve the clinical utility of high-gamma ECoG to identify eloquent language cortex. Our future directions include using machine learning to optimize our high-gamma ECoG analysis to best predict ESM maps and to compare both our high-gamma ECoG and ESM maps to language outcomes post resection. Funding: None