Magnetoencephalography (MEG) Data Processing in Epilepsy Patients with Implanted Responsive Neurostimulation (RNS)...

Area of Science:

• Neuroscience
• Medical Imaging
• Signal Processing

Background:

• Drug-resistant epilepsy (DRE) often necessitates surgical or neuromodulatory interventions like responsive neurostimulation (RNS).
• Neuroimaging, particularly magnetoencephalography (MEG), is crucial for epilepsy assessment, but RNS implants distort MEG signals.
• Existing methods struggle to mitigate signal distortions caused by RNS devices, complicating post-implantation neuroimaging.

Purpose of the Study:

• To introduce and evaluate an independent component analysis (ICA)-based approach for enhancing MEG signal quality in patients with RNS implants.
• To improve the diagnostic accuracy and clinical utility of MEG for epilepsy assessment in patients who have undergone RNS implantation.

Main Methods:

• Utilized an automated ICA-based approach with MNE-Python for MEG data preprocessing.
• Applied temporal signal space separation (tSSS) alongside ICA for noise removal.
• Analyzed power spectral density (PSD) and signal-to-noise ratio (SNR), and performed single equivalent current dipole (SECD) modeling for MEG dipole analysis.

Main Results:

• The ICA-based noise removal preprocessing significantly improved the signal-to-noise ratio (SNR) in MEG data from RNS patients.
• Qualitative assessments indicated enhanced signal readability and improved MEG dipole analysis.
• The method demonstrated a marked enhancement in overall MEG data quality for patients with RNS implants.

Conclusions:

• ICA-based processing is a clinically relevant and effective method for improving MEG data quality in epilepsy patients with RNS devices.
• This technique facilitates more accurate neurophysiological assessment and potentially better surgical planning in challenging patient populations.
• Enhanced MEG signal quality through ICA can lead to more informed clinical decisions for patients with drug-resistant epilepsy and RNS implants.