Assessment of Anesthetic Depth Through EEG Mode Decomposition Using Singular Spectrum Analysis

Area of Science:

• Neuroscience
• Anesthesiology
• Signal Processing

Background:

• Electroencephalography (EEG) is crucial for monitoring anesthesia depth.
• Conventional Fourier analysis struggles with non-stationary EEG dynamics during anesthesia.
• Singular Spectrum Analysis (SSA) offers a novel approach to analyze complex EEG signals.

Purpose of the Study:

• To investigate the utility of SSA combined with Hilbert transform for extracting EEG features during sevoflurane anesthesia.
• To assess the ability of SSA-derived features to correlate with anesthetic depth.
• To compare SSA with conventional spectral analysis for EEG monitoring.

Main Methods:

• Frontal EEG data from 10 patients under sevoflurane anesthesia were analyzed.
• EEG signals were decomposed into intrinsic mode functions (IMFs) using SSA.
• Hilbert spectral analysis was applied to IMFs for instantaneous frequency and amplitude extraction.

Main Results:

• SSA captured phase-dependent EEG changes, including alpha spindle activity and high-frequency components.
• IMF features showed strong correlations with the Bispectral Index (BIS) (R² = 0.88, MAE < 4).
• SSA demonstrated superior temporal resolution and stable feature extraction compared to conventional methods.

Conclusions:

• SSA combined with Hilbert analysis provides a robust framework for quantitative EEG analysis during general anesthesia.
• This approach may enhance real-time, individualized assessments of anesthetic depth.
• The findings suggest potential for improved patient safety and anesthetic management.