Source Extent Estimation in OPM-MEG: A Two-Stage Champagne Approach.

Area of Science:

• Neuroscience
• Biophysics
• Medical Imaging

Background:

• Accurate estimation of brain source extent is crucial for preoperative functional localization in epilepsy using magnetoencephalography (MEG).
• Conventional source imaging methods often yield inaccurate source extent estimations, hindering clinical applications.
• Existing techniques struggle to precisely delineate the spatial extent of neural activity.

Purpose of the Study:

• To introduce a novel two-stage method, TS-Champagne (two-stage Champagne approach), for enhanced source extent estimation in MEG.
• To improve the accuracy and robustness of source localization compared to conventional methods.
• To validate the performance of TS-Champagne using numerical simulations and experimental MEG data.

Main Methods:

• TS-Champagne employs a two-stage process: initial estimation using Champagne-NL (Champagne algorithm with noise learning), followed by incorporating spatial priors derived from the initial estimate.
• Spatial basis functions, representing potential source centers and their neighbors, are constructed and used as priors in the second stage.
• Numerical simulations and experiments with an optically pumped magnetometer (OPM)-MEG system were conducted for performance evaluation.

Main Results:

• TS-Champagne demonstrated superior robustness across various simulation conditions, including different source extents, numbers of sources, signal-to-noise ratios, and source correlations.
• The method outperformed Champagne-NL and other benchmark techniques in accuracy and reliability.
• Experimental validation showed spatially and temporally consistent source activity reconstruction, aligning with established neurophysiological findings.

Conclusions:

• TS-Champagne offers a significant advancement in accurately estimating brain source extent using MEG.
• The method's robustness and accuracy make it a promising tool for preoperative functional localization in epilepsy.
• TS-Champagne shows high feasibility for practical applications in clinical neuroimaging.