Explainable and Scalable Machine-Learning Algorithms for Detection of Autism Spectrum Disorder using fMRI Data

Diagnosing Autism Spectrum Disorder (ASD) is a challenging problem, and is based purely on behavioral descriptions of symptomology (DSM-5/ICD-10), and requires informants to observe children with disorder across different settings (e.g. home, school). Numerous limitations (e.g., informant discrepancies, lack of adherence to assessment guidelines, informant biases) to current diagnostic practices have the potential to result in over-, under-, or misdiagnosis of the disorder. Advances in neuroimaging technologies are providing a critical step towards a more objective assessment of the disorder. Prior research provides strong evidence that structural and functional magnetic resonance imaging (MRI) data collected from individuals with ASD exhibit distinguishing characteristics that differ in local and global spatial, and temporal neural-patterns of the brain. Our proposed deep-learning model ASD-DiagNet exhibits consistently high accuracy for classification of ASD brain scans from neurotypical scans. We have for the first time integrated traditional machine-learning and deep-learning techniques that allows us to isolate ASD biomarkers from MRI data sets. Our method, called Auto-ASD-Network, uses a combination of deep-learning and Support Vector Machines (SVM) to classify ASD scans from neurotypical scans. Such interpretable models would help explain the decisions made by deep-learning techniques leading to knowledge discovery for neuroscientists, and transparent analysis for clinicians.

ASD-DiagNet: A hybrid learning approach for detection of Autism Spectrum Disorder using fMRI data

Mental disorders such as Autism Spectrum Disorders (ASD) are heterogeneous disorders that are notoriously difficult to diagnose, especially in children. The current psychiatric diagnostic process is based purely on the behavioural observation of symptomology (DSM-5/ICD-10) and may be prone to over-prescribing of drugs due to misdiagnosis. In order to move the field towards more quantitative fashion, we need advanced and scalable machine learning infrastructure that will allow us to identify reliable biomarkers of mental health disorders. In this paper, we propose a framework called ASD-DiagNet for classifying subjects with ASD from healthy subjects by using only fMRI data. We designed and implemented a joint learning procedure using an autoencoder and a single layer perceptron which results in improved quality of extracted features and optimized parameters for the model. Further, we designed and implemented a data augmentation strategy, based on linear interpolation on available feature vectors, that allows us to produce synthetic datasets needed for training of machine learning models. The proposed approach is evaluated on a public dataset provided by Autism Brain Imaging Data Exchange including 1035 subjects coming from 17 different brain imaging centers. Our machine learning model outperforms other state of the art methods from 13 imaging centers with increase in classification accuracy up to 20% with maximum accuracy of 80%. The machine learning technique presented in this paper, in addition to yielding better quality, gives enormous advantages in terms of execution time (40 minutes vs. 6 hours on other methods). The implemented code is available as GPL license on GitHub portal of our lab (https://github.com/pcdslab/ASD-DiagNet).