Schizophrenia Detection in Adolescents from EEG Signals using Symmetrically weighted Local Binary Patterns

Area of Science:

• Neuroscience
• Biomedical Engineering
• Machine Learning

Background:

• Schizophrenia is a complex mental disorder.
• Early detection of schizophrenia in adolescents is crucial for timely intervention.
• Electroencephalogram (EEG) signals offer a non-invasive window into brain activity.

Purpose of the Study:

• To develop and validate an automated approach for detecting schizophrenia in adolescents using EEG signals.
• To evaluate the efficacy of Symmetrically Weighted Local Binary Patterns (SLBP) for schizophrenia detection.
• To compare the proposed method's performance against existing approaches.

Main Methods:

• Extraction of SLBP-based histogram features from adolescent EEG channels.
• Application of a correlation-based feature selection algorithm to reduce feature vector length.
• Classification of EEG signals into schizophrenia and healthy groups using a LogitBoost classifier.

Main Results:

• The SLBP approach effectively characterized changes in EEG signals associated with schizophrenia.
• The automated system achieved a classification accuracy of 91.66% in distinguishing schizophrenia from healthy EEG signals.
• The proposed method outperformed recently reported approaches for schizophrenia detection.

Conclusions:

• SLBP-based feature extraction is a promising technique for automated schizophrenia detection in adolescents.
• The developed automated approach demonstrates high accuracy and potential for clinical application.
• This method offers a valuable tool for improving early diagnosis and management of schizophrenia.