Abstracts

This is a Late Breaking abstract

Rationale: Epilepsy neuroimaging using cryogenically cooled magnetometers (cryo-MEGs) is a relatively recent addition to the neuroimaging toolkit but has proven itself extensively in the previous years as an effective addition for treatment planning. This is due largely to cryo-MEG's advantage of using magnetic signals rather than electrical discharge which is subject to smearing due to the inconsistent conductivity properties of the head. Optically pumped magnetometers (OPM-MEGs) improve upon this concept by removing the extensive initial cost of setting up a liquid helium cooling system and provide portability. By combining these novel more portable systems with mainstream data processing pipelines via Brainstorm this study hopes to get one step closer to adding OPM-MEG systems to the epilepsy imaging toolkit.

Methods: Five healthy normal volunteers were asked to undergo OPM-MEG scanning in an actively magnetically shielded room. After sensor calibration volunteers were asked to undergo 4 tasks each that were 60 seconds in duration. The first task was a 60 second baseline where participants were asked to relax with their eyes open. The second task was a basic finger-tapping task with the volunteer using their right index finger to tap for 60 seconds to stimulate the left motor area. The third task was a 60 second visual checkerboard task consisting of alternating checkerboard images alternating at 10Hz to stimulate the occipital area. The final task was a language task where volunteers were presented a noun on screen and must think of a verb associated with that noun. This entire process was repeated 3 times for each volunteer scan. Data were then imported into the Brainstorm software package inside MATLAB and a channel file was created based on the already known locations of the sensors on the helmet relative to participants' heads. Pre-processing consisted of a high-pass filter at 0.3 Hz and a notch filer at 60 HZ harmonics. Following these steps data was converted into the source space using Brainstorm's dSPM minimum norm imaging method. Source space images were converted to power using pWelch methodology and spectrum normalized to total power. Final results were computed by projecting the power data onto the ICBM12 template and averaging the group.

Results: After source imaging and averaging across the group (n=5) all task epochs show power changes in expected areas and frequency bands. Finger tap epoch data (Figure 1A) reveals beta frequency band (13-30 Hz) power changes in the left hemisphere motor area, visual task data (Figure 1B) shows gamma frequency band (31-80 Hz) activation in the occipital regions, and language task data (Figure 1C) shows prominent activations near the wernicke's and broca's areas consistent with previous language task data.

Conclusions: OPM-MEG imaging systems provided expected results from 3 separate tasks using Brainstorm successfully with with collected OPM-MEG data. Further testing of this data collection and processing methodology using the popular software brainstorm could yield an easy to follow OPM-MEG pipeline for the epilepsy research community to use in the clinical domain.

Funding: Supported by the G.R. White Trust and Effie and Wofford Cain Endowment