Machine Learning Methods for Diagnosing Autism Spectrum Disorder and Attention- Deficit/Hyperactivity Disorder Using Functional and Structural MRI: A Survey

Machine Learning (ML) techniques especially Deep Learning (DL) models are advanced tools that have achieved remarkable success in many scientific fields such as computer vision, speech recognition, natural language processing, medical imaging, and computational neuroscience. Over the past decade, using conventional ML and DL tools using brain imaging data such as Magnetic Resonance Imaging (MRI) and functional MRI (fMRI) has become very popular. These techniques provide a way to characterize informative patterns from the structure and function of the brain and distinguish them from non-discriminative and noisy information in order to diagnose brain disorders. In this study, we focus on two major brain disorder among children, Attention Deficit Hyperactivity Disorder (ADHD) and Autism Spectrum Disorder (ASD). Considering the importance of early and accurate diagnosis of these disorders in children's lives, we conduct a thorough investigation of available ML and DL methods for diagnosing ADHD and ASD using brain fMRI and MRI data. We discuss the advantages and pitfalls of available methods and address the possible directions for future developments in this field.