This prospective study compared the topography of high-gamma modulation (HGM) during a story-listening task requiring negligible patient cooperation, with the conventional electrical stimulation mapping (ESM) using a picture-naming task, for presurgical language localization in pediatric drug-resistant epilepsy.


Patients undergoing extraoperative monitoring with subdural electrodes were included. Electrocorticographic signals were recorded during quiet baseline and a story-listening task. The likelihood of 70- to 150-Hz power modulation during the listening task relative to the baseline was estimated for each electrode and plotted on a cortical surface model. Sensitivity, specificity, accuracy, and diagnostic odds ratio (DOR) were estimated compared to ESM, using a meta-analytic framework.


Nineteen patients (10 with left hemisphere electrodes) aged 4-19 years were analyzed. HGM during story listening was observed in bilateral posterior superior temporal, angular, supramarginal, and inferior frontal gyri, along with anatomically defined language association areas. Compared to either cognitive or both cognitive and orofacial sensorimotor interference with naming during ESM, left hemisphere HGM showed high specificity (0.82-0.84), good accuracy (0.66-0.70), and DOR of 2.23 and 3.24, respectively. HGM was a better classifier of ESM language sites in the left temporoparietal cortex compared to the frontal lobe. Incorporating visual naming with the story-listening task substantially improved the accuracy (0.80) and DOR (13.61) of HGM mapping, while the high specificity (0.85) was retained. In the right hemisphere, no ESM sites for aphasia were seen, and the results of HGM and ESM comparisons were not significant.


HGM associated with story listening is a specific determinant of left hemisphere ESM language sites. It can be used for presurgical language mapping in children who cannot cooperate with conventional language tasks requiring active engagement. Incorporation of additional language tasks, if feasible, can further improve the diagnostic accuracy of language localization with HGM.


Leave a comment.

Your email address will not be published. Required fields are marked*

Andoird App