DarkASDNet: Classification of ASD on Functional MRI Using Deep Neural Network

Abstract

Non-invasive whole-brain scans aid the diagnosis of neuropsychiatric disorder diseases such as autism, dementia, and brain cancer. The assessable analysis for autism spectrum disorders (ASD) is rationally challenging due to the limitations of publicly available datasets. For diagnostic or prognostic tools, functional Magnetic Resonance Imaging (fMRI) exposed affirmation to the biomarkers in neuroimaging research because of fMRI pickup inherent connectivity between the brain and regions. There are profound studies in ASD with introducing machine learning or deep learning methods that have manifested advanced steps for ASD predictions based on fMRI data. However, utmost antecedent models have an inadequacy in their capacity to manipulate performance metrics such as accuracy, precision, recall, and F1-score. To overcome these problems, we proposed an avant-garde DarkASDNet, which has the competence to extract features from a lower level to a higher level and bring out promising results. In this work, we considered 3D fMRI data to predict binary classification between ASD and typical control (TC). Firstly, we pre-processed the 3D fMRI data by adopting proper slice time correction and normalization. Then, we introduced a novel DarkASDNet which surpassed the benchmark accuracy for the classification of ASD. Our model's outcomes unveil that our proposed method established state-of-the-art accuracy of 94.70% to classify ASD vs. TC in ABIDE-I, NYU dataset. Finally, we contemplated our model by performing evaluation metrics including precision, recall, F1-score, ROC curve, and AUC score, and legitimize by distinguishing with recent literature descriptions to vindicate our outcomes. The proposed DarkASDNet architecture provides a novel benchmark approach for ASD classification using fMRI processed data.

Construction of an immune-related ceRNA network to screen for potential diagnostic markers for autism spectrum disorder

“…Analyses of ROC curves revealed that the diagnostic model based on a combination of these five diagnostic genes had an AUC of 0.923. This value is higher than the AUC (0.910) of the three-methylation-markers diagnostic model identified using the same algorithm ( Zhang et al, 2022 ), the AUC (0.910) of the five-gut-bacterial-markers diagnostic model identified by a metagenomic analysis combined with a random forest algorithm ( Wan et al, 2022 ), and the AUC (0.860) of the autism-risk-index diagnostic method based on eye-tracking measures ( Frazier et al, 2018 ), but slightly lower than the AUC (0.947) of the DarkASDNet diagnostic model based on 3D-fMRI ( Ahammed et al, 2021 ). Finally, we confirmed the expression of these five genes in collected serum samples and the validation dataset.…”

Progress in magnetic resonance imaging of autism model mice brain

“…The pathogenic factors of autism in human patients are not yet clear, the severity of core symptoms differs, the rate of missed diagnosis and misdiagnosis in behavioral diagnosis is high (Fusar‐Poli et al, 2022), and there is a lack of an efficient and accurate classification method. fMRI has been applied to the clinical diagnosis of patients with autism, and it has been suggested that using imaging features may be a new way to classify autism subtypes (Ahammed et al, 2021).…”

“…In addition, the 26 autism model mice were clustered based on differences in brain volume and anatomical phenotypes, to obtain three groups of mice: (1) Those with Nrxn1α , En2 , Shank3 , and Fmr1 mutations; (2) the AndR , BTBR, Gtf2i dp/dp , Itgβ3 , 15q11‐13, Slc6A4 Ki (129), and Nl3 Ki models; and (3) the 16p11, BALB/C, Cntnap2 −/− , Gtf2i +/− , Mecp2 , Slc6A4 Ki (B6), Slc6A4 knockout, and XO models. White matter dysplasia has been widely confirmed in patients with autism, and the severity of white matter dysplasia is related to the severity of autism (Ahammed et al, 2021). From the perspective of white matter, the model mice in the first group have increased white matter volume, the second group shows reduced white matter volume, while the white matter volume of the third group shows no significant change.…”

Feature Signature Discovery for Autism Detection: An Automated Machine Learning Based Feature Ranking Framework

“…Several studies were undertaken on classifying ASD from images through traditional machine learning models, hybrid models that infuse feature selection with classifcation, deep learning models, and AutoML models [20,[23][24][25][26][27]. A concise review of the recent research on ASD classifcation is depicted in Table 1.…”