NORMA eResearch @NCI Library

Enhancing the accuracy of Autism detection using fMRI images with Graph Autoencoder and Graph Neural Networks

Rai, Sumit (2022) Enhancing the accuracy of Autism detection using fMRI images with Graph Autoencoder and Graph Neural Networks. Masters thesis, Dublin, National College of Ireland.

[thumbnail of Master of Science]
PDF (Master of Science)
Download (10MB) | Preview
[thumbnail of Configuration manual]
PDF (Configuration manual)
Download (4MB) | Preview


Autism Spectral Disorder (ASD) is a neurological developmental disorder. It affects the brain and the symptoms usually manifest as challenges in social skills, restrictive interests and repetitive behaviour. The traditional methods for ASD detection have been based on behavioral observations which are neither efficient nor accurate. Recently, resting-state Functional Magnetic Resonance Imaging (rs-FMRI) has been used to understand the mechanisms of brain disorders such as ASD. In this paper, a graph neural network with graph auto-encoder is proposed and implemented to enhance the accuracy of the ASD detection using rs-FMRI. The correlation between blood oxygen level-dependent (BOLD) signals in the region of interest (ROI) in the brain is used to create functional connectivity matrix. Then, the graph auto-encoder is used for feature representation and a graph neural network is used for the classification task. Both the networks (GAE & GNN) are trained together to tune the latent representation by graph auto-encoder for classifying the subjects as Autistic or non-Autistic. An accuracy of around 55% is achieved on ABIDE preprocessed dataset, which is less than ideal for medical applications. The results show that the proposed framework should be improved further to achieve an improved classification accuracy over other research which are usually above 78% on ABIDE-I preprocessed dataset.

Item Type: Thesis (Masters)
Subjects: Q Science > QA Mathematics > Electronic computers. Computer science
T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science
R Medicine > RC Internal medicine > RC0321 Neuroscience. Biological psychiatry. Neuropsychiatry
T Technology > Biomedical engineering
Divisions: School of Computing > Master of Science in Data Analytics
Depositing User: Tamara Malone
Date Deposited: 01 Mar 2023 11:53
Last Modified: 01 Mar 2023 11:54

Actions (login required)

View Item View Item