Autism spectrum disorder (ASD) is characterized by qualitative impairment in social reciprocity, and by repetitive, restricted, and stereotyped behaviors/interests.

ASD is recognized to occur in more than 1% of children. Despite continuing research advances, their pace and clinical impact have not kept up with the urgency to identify ways of determining the diagnosis at earlier ages, selecting optimal treatments, and predicting outcomes. For the most part this is due to the complexity and heterogeneity of ASD.

PROBLEM STATEMENT : Autism Classification on rs-FMRI

Resting state fMRI (rsfMRI or R-fMRI) is a method of functional magnetic resonance imaging (fMRI) that is used in brain mapping to evaluate regional interactions that occur in a resting or task-negative state, when an explicit task is not being performed. A number of resting-state conditions are identified in the brain, one of which is the default mode network.

The interim project presentation is as below:

The presentation slides can be found here.

Implementation Strategies

• Classification on components obtained via GIFT ICA ~ 3D CNN/AutoEncoder
• Averaging out 4D Data in time and perform classification to exploit the usefulness of spatial information for classification ~ 3D CNN/AutoEncoder
• Classification on MRI Anatomical 3D Data after performing registration
• *New Approach: AutoEncoder + GRU/LSTM : Encoding the 4D rs-fMRI into a single-vector representation through which we learn the temporal information via GRU/LSTM

The final project presentation, as below, discusses all the implementations, performances obtained and justifications :

The final presentation slides can be found here. The activation outputs can be found here.

Copyright © black_white | HTML5 | CSS