Enhanced visibility graph for EEG classification.

Area of Science:

• Neuroscience
• Signal Processing
• Machine Learning

Background:

• Electroencephalography (EEG) is crucial for understanding brain activity in cognitive states and neurological disorders.
• Existing EEG analysis methods may not fully capture complex signal dynamics.
• There is a need for advanced techniques to improve EEG classification accuracy.

Purpose of the Study:

• To propose and evaluate an end-to-end framework for EEG classification.
• To integrate power spectral density (PSD) and visibility graph (VG) features with deep learning (DL) models.
• To compare the performance of different DL architectures for EEG data analysis.

Main Methods:

• Developed an integrated framework combining PSD and VG feature extraction.
• Applied four deep learning architectures: MLP, LSTM, InceptionTime, and ChronoNet.
• Evaluated the framework on multiple EEG datasets under various experimental conditions.

Main Results:

• The proposed framework demonstrated high accuracy in classifying EEG data.
• Visibility graph (VG) features showed particular efficacy in improving classification performance.
• The study provided insights into the strengths and limitations of different feature extraction and DL methods.

Conclusions:

• The integrated framework offers a holistic approach to EEG signal analysis.
• VG features combined with DL show significant promise for advancing EEG-based systems.
• This work contributes to more accurate and reliable neuroscience and clinical applications.