Autism detection based on MRI images using Deep Learning

Abstract

Autism spectrum disorder (ASD) is a neuro dysfunction or neurodevelopmental disorder. This causes a patient to have trouble with social interaction which causes social instability. It also causes speech problems or difficulty with any sort of verbal communication as well as nonverbal communication. The biggest issue with Autism is that it is difficult to diagnose it at an early level. The difficulty in diagnosing is due to the lack of a particular medical test for it. Researchers have yet to discover a bio marker or specific gene that can detect autism. Doctors still use outdated methods to identify autism nowadays. Doctors frequently keep track of a patient’s behavior since childhood. To address this issue and diagnose autism, artificial intelligence will be used in our research to develop an ASD diagnosis method. Our research will employ neuroimages. Functional MRI and Structural MRI images will be used to train our neural network model. ABIDE, a versatile dataset was used to initialize this research. This includes struc tural MRI and fMRI data from young and old ASD patients as well as healthy individuals. After examining the MRI pictures, a method was developed to pick out particular layers from those images. The dataset was then constructed using images from ABIDE for our models to train and test without performing any pre-processing. A variety of cutting-edge deep learning architectures were chosen to train using our created dataset. Novel architectures were used to attain an accuracy of 80% to practically 86%. Custom block was used later in the research to expand the dataset and achieve more accuracy. Finally, based on our findings, a model will be found that can more accurately identify autism from MRI pictures.