Classification of BOLD FMRI Signals using Wavelet Transform and Transfer Learning for Detection of Autism Spectrum Disorder

The World Health Organization (WHO) has reported a continuous rise in the prevalence of autism worldwide, in which 1 in 160 children in the world has ASD. The problem in ASD treatment has no definite cure, and one possible option is to control the disorder's progress. Several attempts to use resting-state functional magnetic resonance imaging (fMRI) as an assisting tool combined with a classifier have been reported. Still, researchers barely reach an accuracy of 70% for replicated models with independent datasets. Most of the ASD studies have used functional connectivity and structural measurements and ignored the temporal dynamics features of fMRI data analysis. The purpose of this study is to present several deep learning models to diagnose ASD based on temporal dynamic features of fMRI data and improve the classification results on a sample of data. The sample size is 82 subjects (41 ASD and 41 normal cases) collected from three different sites of Autism Brain Imaging Data Exchange (ABIDE). The default mode network (DMN) regions are selected for blood-oxygen-level-dependent (BOLD) signals extraction. The extracted BOLD signals' time-frequency components are converted to scalogram images and used as input for pre-trained convolutional neural networks for feature extraction such as Googlenet, DenseNet201, Resnetl8, and Resnet101. The extracted features are trained using two classifiers: support vector machine (SVM) and K-nearest neighbors (KNN). Finally, the performance of each model is evaluated based on accuracy, sensitivity, and specificity metrics. The best results obtained from the KNN classifier with DenseNet201 as a pre-trained model are accuracy 85.9%, sensitivity 793%, specificity 92.6%. Compared with previous studies, it is concluded that the proposed model can be considered as an efficient tool for the diagnosis of ASD. From another perspective, the proposed method can be applied to analyzing rs-fMRI data related to brain disorders.

Automated summary

The prevalence of autism is on the rise globally, posing a challenge in treatment without a definite cure. Research often overlooks the temporal dynamics of fMRI data in ASD studies. This study presents deep learning models based on temporal dynamic features to enhance ASD diagnosis accuracy.