Autism spectrum disorder (ASD) is a heterogeneous neurodevelopmental disorder characterized by impaired communication and limited social interactions. The shortcomings of current clinical approaches, which are based exclusively on behavioral observations of symptomology, and the poor understanding of the neurological mechanisms underlying ASD necessitate the identification of new biomarkers. These biomarkers can aid in studying brain development and functioning and can lead to accurate and early detection of ASD.
In this project, we aim to develop novel deep learning models to detect complex biomarkers that can classify ASD brains from neurotypical (NT) brains. One way to study these biomarkers is through the use of neuroimaging technologies, such as fMRI. fMRI data measures the blood-oxygen-level-dependent (BOLD) signal of each small voxel (i.e., “volume element”) at a given time point. Therefore, the data consists of a time series of each voxel representing its activity over time. Our goal is to create deep learning algorithms capable of handling this complex brain neuroimaging data to accurately classify ASD brains from NT brains.
Methods
Method development for this work is ongoing.