Schizophrenia detection using Entropy Difference-based Electroencephalogram Channel Selection Approach.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Machine Learning

Background:

• Schizophrenia diagnosis relies on clinical assessments, often lacking objective biomarkers.
• Electroencephalogram (EEG) signals offer a potential source of objective data for neurological disorder detection.
• Current EEG-based schizophrenia detection methods face challenges in channel selection and computational complexity.

Purpose of the Study:

• To develop a novel and efficient algorithm for identifying schizophrenia using electroencephalogram (EEG) data.
• To introduce an entropy difference (ED)-based channel selection method to pinpoint the most informative EEG channels for schizophrenia detection.
• To enhance the accuracy and reduce the computational load of schizophrenia classification from EEG signals.

Main Methods:

• An entropy difference (ED)-based algorithm was employed for selecting the most significant EEG channels.
• Discrete Wavelet Transform (DWT) was used to decompose selected EEG signals into subbands.
• Symmetrically-weighted local binary patterns were extracted for feature extraction from subband variations.
• Support Vector Machine (SVM) was utilized for classifying individuals with and without schizophrenia.

Main Results:

• The proposed ED-based channel selection algorithm successfully identified discriminative EEG channels for schizophrenia detection.
• The method achieved a remarkable 100% classification accuracy using features from a single selected channel.
• The ED-based channel selection approach demonstrated superior performance compared to existing entropy-based methods.

Conclusions:

• The novel ED-based channel selection algorithm offers a highly accurate and computationally efficient method for schizophrenia detection using EEG.
• This approach significantly advances the potential of EEG as an objective biomarker for schizophrenia diagnosis.
• The findings suggest a promising direction for developing advanced neuroimaging-based diagnostic tools for mental health disorders.