Schizophrenia Diagnosis via FFT and Wavelet Convolutional Neural Networks utilizing EEG signals

Abstract BackgroundSchizophrenia is a chronic mental illness in which a person's perception of reality is distorted. Early diagnosis can help to manage symptoms and increase long-term treatment. The electroencephalogram (EEG) is now used to diagnose certain mental disorders. MethodIn this paper, we developed an artificial intelligence methodology built on deep convolutional neural networks with specialized layers. In the first phase, we used the Gramian Angular Field (GAF) including two methods: Gramian Angular Summation Field (GASF) and Gramian Angular Difference Field (GADF) to represent the EEG signals as various types of images. Then, well-known CNN architectures includes of Transformer CNN-LSTM and two new custom architectures which utilizing two-dimensional Fast Fourier transform layers (CNN-FFT) and wavelet transform layers (CNN-Wavelet) are preformed to extract useful information from the data. These layers allow automated feature extraction from EEG representation in the time and frequency domains. ResultsCNN-FFT and Transformer models derive most valuable features from signals based on the findings. CNN-FFT obtained the highest accuracy of 99.04 percent. Transformer, which has a 98.32 percent accuracy rate, also performs admirably. ConclusionThis experiment outperformed other previous studies. Consequently, the strategy can aid medical practitioners in automated detection and early treatment of schizophrenia. The code of the work is also publicly available on github: https://github.com/i1idan/schizophreniadiagnosis-eeg-signals