An Automated Algorithm for the Identification of Somatosensory Cortex Using Magnetoencephalography

Area of Science:

• Neurosurgery
• Biomedical Engineering
• Neuroimaging

Background:

• Accurate localization of eloquent cortex is vital for neurosurgical procedures like epilepsy and tumor resections.
• Current non-invasive methods using magnetoencephalography (MEG) rely on equivalent current dipoles, which involve subjective parameters.
• There is a need for automated, objective methods for precise cortical area localization.

Purpose of the Study:

• To develop an automated algorithm for identifying activated cortical areas during somatosensory tasks using MEG data.
• To utilize singular value decomposition (SVD) for outlining task-related cortical regions.
• To validate the algorithm's proof of concept in subjects with epilepsy.

Main Methods:

• Development of an automated algorithm based on singular value decomposition (SVD).
• Application of the algorithm to magnetoencephalography (MEG) recordings during a somatosensory task.
• Evaluation using data from 10 epilepsy patients, comparing results to the expected somatosensory cortex location.

Main Results:

• The automated algorithm demonstrated a statistically significant overlap with the expected somatosensory cortex in 6 out of 10 epilepsy subjects.
• The algorithm successfully identified the dominant cortical area and boundaries involved in task-related responses.
• Proof of concept was established, indicating the algorithm's potential for clinical application.

Conclusions:

• The developed automated algorithm shows feasibility for non-invasive localization of task-related cortical areas.
• The algorithm offers an objective alternative to subjective parameters in current MEG source localization methods.
• Further testing in a larger cohort is warranted to confirm the algorithm's efficacy and clinical relevance.