EEG-Based Classical Framework for Autism Spectrum Disorder Diagnosis

The objective of signal processing using transformation techniques, is to accurately detect the frequency and extract small variations within the signals. Previous studies have explored various algorithms aiming to extracted the signals and classify Autism Spectral Disorder (ASD) based on Electroencephalography (EEG) signals, offering valuable insights. However, there is still scope for improvement in handling highly non-stationary signals across ASD diagnosis. The article propose a model that named as Convolution neural network (CNN-ASD) and the Continuous wavelet transform (CWT) to properly take features from EEG signals. This approach uses an adaptive filter bank to control dynamic data in the frequency domain to accurately distinguish both stages. Our experiments demonstrate that, the proposed model outperforms BiLSTM with AM and ResNet50, and it achieves notable classification accuracy for normal and ASD classes. This method enhances the performance of medical devices and benefits for their use in healthcare.