Phase-amplitude coupling within MEG data can identify eloquent cortex
Srijita Das1, Kevin Tyner1, Stephen V Gliske1
1Department of Neurosurgery, University of Nebraska Medical Center, 988437 Nebraska Medical Center, Omaha, NE 68198-7400, United States of America.
Journal of Neural Engineering
|May 2, 2025
Summary
We developed a new algorithm using phase-amplitude coupling (PAC) to objectively map the brain's somatosensory cortex in epilepsy patients, improving pre-surgical planning.
Area of Science:
- Neuroscience
- Biomedical Engineering
- Medical Imaging
Background:
- Accurate identification of eloquent cortices is crucial for minimizing post-surgical deficits in epilepsy and tumor patients.
- Current pre-surgical mapping methods are subjective and require significant expertise, highlighting the need for objective approaches.
- Phase-amplitude coupling (PAC) shows promise as a biomarker for functional brain mapping due to its role in task-induced brain activity.
Purpose of the Study:
- To develop and validate a novel phase-amplitude coupling (PAC)-based algorithm for non-invasive identification of the somatosensory eloquent cortex.
- To utilize magnetoencephalography (MEG) data from epilepsy patients for objective pre-surgical functional mapping.
- To establish PAC as a reliable biomarker for identifying task-specific brain activity.
Main Methods:
- Analysis of somatosensory and spontaneous magnetoencephalography (MEG) recordings from 30 epilepsy patients.
- Calculation of phase-amplitude coupling (PAC) on source-reconstructed data (5-12 Hz and 30-300 Hz) followed by rank-2 tensor decomposition.
- Application of density-based clustering and a patient-specific support vector machine (SVM) classifier to identify active brain regions based on PAC values.
Main Results:
- The PAC algorithm successfully identified five of six expected somatosensory brain regions during stimulation (p=1.08×10-8).
- A linear mixed-effects model confirmed statistically significant task-specific PAC in anatomically relevant regions (p < 0.01).
- The patient-specific SVM classifier achieved high specificity (99.46%) and precision (66.9%) in identifying active brain regions.
Conclusions:
- The developed PAC-based algorithm reliably and objectively identifies somatosensory cortex activation in epilepsy patients at both individual and population levels.
- This study demonstrates the feasibility of using PAC as a non-invasive marker for functional brain mapping.
- Future research will explore the application of this PAC algorithm for mapping other eloquent cortices, such as language, motor, and auditory areas.
