Schrödinger filtering: a precise EEG despiking technique for EEG-fMRI gradient artifact

Area of Science:

• Neuroimaging
• Signal Processing
• Biophysics

Background:

• Gradient artifacts in simultaneous EEG-fMRI contaminate neural signals.
• Spike artifacts overlap with EEG frequencies, confounding analysis.
• Artifacts distort functional connectivity measures.

Purpose of the Study:

• Introduce Schrödinger filtering for EEG-fMRI spike artifact removal.
• Evaluate its efficacy in spike removal and signal preservation.
• Compare its performance against existing despiking techniques.

Main Methods:

• Decomposition of EEG data using the Schrödinger equation.
• Identification and isolation of spike components.
• Automatic subtraction of artifactual components.

Main Results:

• Schrödinger filtering achieved high spike removal and signal preservation.
• Evoked neural activity remained unaffected.
• Reduced spurious correlations in spontaneous EEG activity.
• Outperformed median filtering, amplitude thresholding, and wavelet denoising.

Conclusions:

• Schrödinger filtering is a robust method for EEG-fMRI artifact removal.
• It enhances data quality for both frequency-based and connectivity analyses.
• Recommended for future EEG-fMRI studies and other spike artifact applications.