EEG-based schizophrenia detection: integrating discrete wavelet transform and deep learning.

Area of Science:

• Neuroscience
• Computational Psychiatry
• Signal Processing

Background:

• Schizophrenia (SZ) is a widespread psychological disorder impacting millions globally.
• Current diagnosis relies on manual assessment by medical practitioners.
• Machine learning and deep learning on electroencephalogram (EEG) signals offer potential for diagnostic support.

Purpose of the Study:

• To develop a highly accurate, automated system for detecting Schizophrenia (SZ) using EEG signals.
• To leverage deep learning and wavelet transform for improved SZ detection.
• To provide a reliable tool to assist medical professionals in SZ diagnosis.

Main Methods:

• Utilized Discrete Wavelet Transform (DWT) to decompose EEG signals into sub-bands and extract features.
• Explored various DWT settings, identifying Daubechies (db4) wavelet with 7-level decomposition as optimal.
• Employed a multilayer perceptron neural network (MLP) for classification between SZ patients and healthy controls (HC).

Main Results:

• The proposed system achieved high classification accuracies: 99.61% on Dataset-1 (DS1) and 99.12% on Dataset-2 (DS2).
• The method demonstrated superior performance compared to existing state-of-the-art techniques for SZ detection.
• Validated on two public datasets comprising a total of 109 records (46 HC and 63 SZ).

Conclusions:

• The developed deep learning and wavelet transform-based system provides a reliable and accurate method for SZ detection using EEG signals.
• This automated approach shows significant potential to support physicians in the clinical diagnosis of Schizophrenia.
• The findings highlight the efficacy of advanced signal processing and machine learning in psychiatric disorder detection.