Autism Spectrum Disorder Classification in Children based on Structural MRI Features Extracted using Contrastive Variational Autoencoder

Ruimin Ma,Ruitao Xie,Yanlin Wang,Jintao Meng,Yanjie Wei,Wenhui Xi,Yi Pan

2023-07-03

Abstract:Autism spectrum disorder (ASD) is a highly disabling mental disease that brings significant impairments of social interaction ability to the patients, making early screening and intervention of ASD critical. With the development of the machine learning and neuroimaging technology, extensive research has been conducted on machine classification of ASD based on structural MRI (s-MRI). However, most studies involve with datasets where participants' age are above 5. Few studies conduct machine classification of ASD for participants below 5-year-old, but, with mediocre predictive accuracy. In this paper, we push the boundary of predictive accuracy (above 0.97) of machine classification of ASD in children (age range: 0.92-4.83 years), based on s-MRI features extracted using contrastive variational autoencoder (CVAE). 78 s-MRI, collected from Shenzhen Children's Hospital, are used for training CVAE, which consists of both ASD-specific feature channel and common shared feature channel. The ASD participants represented by ASD-specific features can be easily discriminated from TC participants represented by the common shared features, leading to high classification accuracy. In case of degraded predictive accuracy when data size is extremely small, a transfer learning strategy is proposed here as a potential solution. Finally, we conduct neuroanatomical interpretation based on the correlation between s-MRI features extracted from CVAE and surface area of different cortical regions, which discloses potential biomarkers that could help target treatments of ASD in the future.

Computer Vision and Pattern Recognition

What problem does this paper attempt to address?

The main objective of this paper is to utilize Contrastive Variational Autoencoder (CVAE) to extract Autism Spectrum Disorder (ASD)-specific features from structural Magnetic Resonance Imaging (s-MRI) data and classify ASD in children aged between 0.92 and 4.83 years based on these features. Specifically, the paper addresses the following key issues:

1. **Improving early diagnosis accuracy**: Existing studies mainly focus on ASD classification in children over 5 years old, while research on children under 5 is scarce and the prediction accuracy is not high. This paper achieves a classification accuracy higher than 0.97 in this age group by using CVAE to extract features from s-MRI data.

2. **Feature extraction and classification**: The paper proposes a CVAE-based method to extract ASD-specific features as well as common shared features. The ASD-specific features can effectively distinguish ASD participants from Typical Control (TC) participants, thereby achieving high-precision ASD classification.

3. **Performance improvement with small sample sizes**: When the dataset is very small, classification performance may decline. To address this issue, the paper introduces a transfer learning strategy, leveraging knowledge from the ABIDE-I dataset to improve classification performance.

4. **Neuroanatomical interpretation**: The paper also conducts neuroanatomical interpretation analysis, exploring the correlation between s-MRI features and the surface area of different cortical regions to identify potential biomarkers that may aid in future clinical treatment of ASD.

In summary, this paper proposes a new feature extraction method and validates its effectiveness on real data, providing important reference for machine classification of early ASD. Additionally, the application of transfer learning strategies further enhances classification performance in small sample sizes, offering strong support for early screening and intervention of ASD in the future.