An Automatic Framework for Detecting Autism Spectrum Disorder From EEG Signals Using TFD

Autism Spectrum Disorder (ASD) is a intricate neuro developmental disorder with many neurological problems. Social interaction and communication issues, repetitive behaviours, and limited interests are its main symptoms. Manual ASD diagnosis testing is prone to human error, time-consuming, and difficult owing to contamination from a number of factors. Electroencephalogram (EEG) signal are extensively utilised to identify ASD as they represent brain abnormalities. This study employed a novel method that included pre-processing, segmentation, Time-frequency distribution (TFD) of various algorithms such as short time fourier transformation (STFT), continuous wavelet transformation (CWT), and smoothed pseudo-Wigner-Ville distribution (SPWVD), which produced corresponding spectrograms, scalograms, and SPWVD-TFD. These TFD are introduced into the DenseNet-121 and ResNet-101 pre-trained (ImageNet data-set) models, and then subsequently fed into the proposed ASD-Net. Deep learning networks (DLM) models were utilised to identify ASD and Normal subject using these TFD images. We acquired a 97.35% mean accuracy utilising the SPWVD-based TFD and ASD-Net model. In compared to the benchmark DenseNet-121 and ResNet-101, the developed convolution neural network (CNN) model with five convolution layers not only needs less learnable parameters but is also computationally efficient and quick.