Analysis of spike waves in epilepsy using Hilbert-Huang transform

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Epilepsy detection relies on identifying characteristic spike waves in electroencephalogram (EEG) signals.
• Traditional EEG analysis methods may not fully capture the complex time-frequency dynamics of these spike waves.
• Advanced signal processing techniques are needed for more accurate and sensitive epilepsy diagnosis.

Purpose of the Study:

• To investigate the time-frequency characteristics of epilepsy spike waves using the Hilbert-Huang transform (HHT).
• To differentiate between spike waves and non-spike waves based on their HHT decomposition.
• To assess the potential of HHT analysis for improving epilepsy symptom detection.

Main Methods:

• Applied the Hilbert-Huang transform (HHT) to decompose electroencephalogram (EEG) signals into intrinsic mode functions (IMFs).
• Analyzed the time-frequency characteristics, including instantaneous, marginal, and Hilbert energy spectra, of spike and non-spike waves.
• Calculated Pearson correlation coefficients and energy-IMF distributions for comparative analysis.

Main Results:

• Significant energy contributions (>10%) were observed in specific IMFs (IMF1, IMF2, residual) for Spike III and referred waves.
• Energy ratios for IMF1, IMF2, IMF3, and residual functions of Spike I and Spike II waves exceeded 10% of their total energy.
• Weighted average frequencies and magnitudes differed between spike and non-spike waves across various IMFs, indicating distinct signal patterns.

Conclusions:

• The Hilbert-Huang transform effectively reveals unique time-frequency signatures of epilepsy spike waves.
• Distinct energy distribution patterns within IMFs can differentiate epileptic activity from normal EEG signals.
• HHT analysis shows promise as a valuable tool for enhancing the accuracy and sensitivity of epilepsy detection.