Quantifying instability in neurological disorders EEG based on phase space DTM function.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Classifying neurological disorders using electroencephalography (EEG) is vital.
• Current methods often rely on frequency domain features and functional brain networks.
• EEG signal instability differences are increasingly recognized in neurological disorders.

Purpose of the Study:

• To propose a novel feature descriptor for characterizing EEG instability.
• To evaluate the effectiveness of this descriptor in classifying neurological disorders.
• To demonstrate the robustness of the proposed method against noisy signals.

Main Methods:

• Phase Space Reconstruction (PSR) to form signal point clouds.
• Construction of a pseudo-metric space and calculation of pseudo-distances.
• Generation of Distance to Measure (DTM) functions and Multivariate Kernel Density Estimation (MKDE) for feature extraction.

Main Results:

• The proposed DTM-based feature descriptor effectively characterizes EEG instability.
• Achieved high classification accuracies for epilepsy (98.00%), Alzheimer's (96.25%), and Parkinson's disease (96.71%, 95.34%).
• Outperformed existing nonlinear descriptors and showed robustness with noisy EEG signals.

Conclusions:

• The DTM function is a promising feature descriptor for EEG instability.
• This method offers a robust and accurate approach for neurological disorder classification.
• The findings highlight the potential of phase space analysis in neurophysiological research.