Abstracts

Rationale: Electrocorticography provides an excellent recording covering a broad range of frequencies of cortical activity but requires surgical placement of electrodes. Changes in high-gamma band activity have been used to map motor areas with high spatial and temporal precision (Crone et al 1998). Advances in EEG technology have allowed for the accurate capture of this activity (Ball et al 2008, Darvas et al 2010). Pediatric patients have anatomical differences from adults that can affect the scalp-based measurement of electrocerebral activity. Surgical treatment is an appealing intervention but is often not considered due to the need for invasive neurodiagnostic procedures and associated risks. Surface EEG of the high-gamma band power has the potential to decrease dependence on invasive ECoG recording for functional mapping. Methods: Nine children aged 7 to 16 with a prior diagnosis of epilepsy underwent placement of electrodes according to the 10-10 system with 8 additional electrodes placed over the bilateral sensorimotor areas. Data acquisition was performed using Natus Neuroworks equipment. Participants performed repetitive index finger tapping while fixating on a remote object. Each subject with acceptable datasets performed between 55 to113 repetitions, with trial duration of 3 to 12 minutes. EMG was placed on the extensor indicis and recorded simultaneously with EEG. Data analysis was performed in Matlab (2017a, Mathworks, Natick, MA), using the EEGLab package (14.1, Delorme & Makeig 2004) and custom scripts. EMG signal was modulated by a bandpass filter, rectification, then low pass filter. The onset of movement was determined when the processed EMG signal crossed a signal-specific threshold amplitude. The resultant epochs were reviewed, some of which were rejected after visual evaluation for excessive noise or artifact. Time-frequency analysis was performed at each electrode overlying the contralateral motor area. Event-related spectral perturbations (ERSP) were computed for each trial, and the optimal signal in the high-gamma-band range (70 to 100 Hz) was selected for power analysis. Signal-to-noise ratio was calculated for each trial. Results: Of the original 9 patient datasets, 7 were excluded due to patient’s inability to complete the task (n=2), ongoing interictal rhythmic activity (n=1), or poor signal quality due to excessive noise or artifact (n=4). The remaining two patients completed both right and left motor tasks. ERSP with motor tasks were calculated at all electrode locations (e.g. Figure 1, which demonstrates activity at C4). Contralateral central, centro-parietal motor and supplementary motor areas demonstrated a decrease in beta band (13 to 30 Hz) and increase in high-gamma band (70 to 100 Hz) associated with the movement. The signal-to-noise ratio for high-gamma band power changes ranged from 1.2 to 1.5.Figure 1: A: Surface EEG ERSP at C4 during a left-sided motor task from a single subject. Dashed blue and red lines represent high-gamma and beta band frequencies evaluated. Black line represents mean EMG activity (amplitude not to scale). B: Frequency-specific ERSP at beta and high-gamma band frequencies. Black line represents normalized mean EMG activity. Conclusions: High-Gamma band power changes with motor activity were detected through the use of surface EEG in a cohort of pediatric patients. Methods and results highlight some areas for improvement in data acquisition and signal processing to advance this technology as an alternative method for detecting motor-related activity. This study demonstrates the feasibility of high-frequency scalp EEG for noninvasive functional mapping in children. Funding: This project relied on institutional seed funds.