Related Experiment Video
Updated: Jul 20, 2026
08:08
Using Wavelet Entropy to Demonstrate how Mindfulness Practice Increases Coordination between Irregular Cerebral and Cardiac Activities
Published on: May 10, 2017
14.7K
Wavelet Transform, Reconstructed Phase Space, and Deep Learning Neural Networks for EEG-Based Schizophrenia
Amjed Al Fahoum1, Ala'a Zyout1
1Biomedical Systems and Informatics Engineering Department, Yarmouk University, Irbid 21163, Jordan.
International Journal of Neural Systems
|July 16, 2024
Summary
This study introduces an expert system using electroencephalogram (EEG) signals for early schizophrenia diagnosis. The system accurately differentiates schizophrenia patients from healthy individuals, offering a reliable diagnostic tool.
Area of Science:
- Neuroscience
- Medical Imaging
- Artificial Intelligence
Background:
- Electroencephalogram (EEG) testing is a cost-effective, safe, and reliable method for diagnosing neurological disorders.
- Traditional EEG analysis methods (time, frequency, time-frequency domains) struggle to detect subtle differences in nonstationary signals between individuals with and without schizophrenia.
- Early diagnosis of schizophrenia is crucial for effective treatment and improved patient outcomes.
Purpose of the Study:
- To develop an innovative expert system for the early diagnosis of schizophrenia using exclusively EEG signals.
- To enhance the detection of subtle, distinguishing features in nonstationary EEG signals that are not apparent in conventional analysis domains.
- To improve the accuracy and reliability of schizophrenia diagnosis through advanced signal processing and machine learning techniques.
Main Methods:
- Utilized reconstructed phase space (RPS) and continuous wavelet transform to maximize signal differences.
- Employed a deep learning network for enhanced image classification of EEG features.
- Validated algorithm efficacy using distinct dataset splits (70% training, 15% validation, 15% testing), 5-fold cross-validation, and leave-one-out classification, each iterated 100 times for robustness.
Main Results:
- The expert system demonstrated remarkable accuracy, precision, sensitivity, F1 score, Matthews correlation coefficient, and Kappa.
- Consistent performance was observed across all evaluation strategies, confirming the algorithm's robustness and reliability.
- The system successfully and autonomously differentiated individuals diagnosed with schizophrenia from healthy controls.
Conclusions:
- The proposed expert system, leveraging EEG signals with advanced signal processing and deep learning, provides an accurate and reliable method for early schizophrenia diagnosis.
- The innovative approach effectively identifies subtle EEG signal characteristics, overcoming limitations of traditional analysis methods.
- This system holds significant potential as a non-invasive, cost-effective tool for clinical psychiatric and neurological diagnostics.
