Multimodal Autism Spectrum Disorder Diagnosis Method Based on DeepGCN
Multimodal data play an important role in the diagnosis of brain diseases. This study constructs a whole-brain functional connectivity network based on functional MRI data, uses non-imaging data with demographic information to complement the classification task for diagnosing subjects, and proposes...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2023-01-01
|
| Series: | IEEE Transactions on Neural Systems and Rehabilitation Engineering |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10247643/ |
| _version_ | 1797679565876756480 |
|---|---|
| author | Mingzhi Wang Jifeng Guo Yongjie Wang Ming Yu Jingtan Guo |
| author_facet | Mingzhi Wang Jifeng Guo Yongjie Wang Ming Yu Jingtan Guo |
| author_sort | Mingzhi Wang |
| collection | DOAJ |
| description | Multimodal data play an important role in the diagnosis of brain diseases. This study constructs a whole-brain functional connectivity network based on functional MRI data, uses non-imaging data with demographic information to complement the classification task for diagnosing subjects, and proposes a multimodal and across-site WL-DeepGCN-based method for classification to diagnose autism spectrum disorder (ASD). This method is used to resolve the existing problem that deep learning ASD identification cannot efficiently utilize multimodal data. In the WL-DeepGCN, a weight-learning network is used to represent the similarity of non-imaging data in the latent space, introducing a new approach for constructing population graph edge weights, and we find that it is beneficial and robust to define pairwise associations in the latent space rather than the input space. We propose a graph convolutional neural network residual connectivity approach to reduce the information loss due to convolution operations by introducing residual units to avoid gradient disappearance and gradient explosion. Furthermore, an EdgeDrop strategy makes the node connections sparser by randomly dropping edges in the raw graph, and its introduction can alleviate the overfitting and oversmoothing problems in the DeepGCN training process. We compare the WL-DeepGCN model with competitive models based on the same topics and nested 10-fold cross-validation show that our method achieves 77.27% accuracy and 0.83 AUC for ASD identification, bringing substantial performance gains. |
| first_indexed | 2024-03-11T23:16:38Z |
| format | Article |
| id | doaj.art-7f52079288154160a82f466dd1fe919d |
| institution | Directory Open Access Journal |
| issn | 1558-0210 |
| language | English |
| last_indexed | 2024-03-11T23:16:38Z |
| publishDate | 2023-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Transactions on Neural Systems and Rehabilitation Engineering |
| spelling | doaj.art-7f52079288154160a82f466dd1fe919d2023-09-20T23:00:05ZengIEEEIEEE Transactions on Neural Systems and Rehabilitation Engineering1558-02102023-01-01313664367410.1109/TNSRE.2023.331451610247643Multimodal Autism Spectrum Disorder Diagnosis Method Based on DeepGCNMingzhi Wang0https://orcid.org/0000-0002-3453-6474Jifeng Guo1https://orcid.org/0000-0002-8692-6255Yongjie Wang2https://orcid.org/0000-0002-1456-8428Ming Yu3https://orcid.org/0009-0007-6616-5093Jingtan Guo4https://orcid.org/0009-0001-5983-6430College of Computer and Control Engineering, Northeast Forestry University, Harbin, ChinaCollege of Computer Science and Engineering, Guilin University of Aerospace Technology, Guilin, ChinaCollege of Computer and Control Engineering, Northeast Forestry University, Harbin, ChinaCollege of Computer and Control Engineering, Northeast Forestry University, Harbin, ChinaCollege of Computer and Control Engineering, Northeast Forestry University, Harbin, ChinaMultimodal data play an important role in the diagnosis of brain diseases. This study constructs a whole-brain functional connectivity network based on functional MRI data, uses non-imaging data with demographic information to complement the classification task for diagnosing subjects, and proposes a multimodal and across-site WL-DeepGCN-based method for classification to diagnose autism spectrum disorder (ASD). This method is used to resolve the existing problem that deep learning ASD identification cannot efficiently utilize multimodal data. In the WL-DeepGCN, a weight-learning network is used to represent the similarity of non-imaging data in the latent space, introducing a new approach for constructing population graph edge weights, and we find that it is beneficial and robust to define pairwise associations in the latent space rather than the input space. We propose a graph convolutional neural network residual connectivity approach to reduce the information loss due to convolution operations by introducing residual units to avoid gradient disappearance and gradient explosion. Furthermore, an EdgeDrop strategy makes the node connections sparser by randomly dropping edges in the raw graph, and its introduction can alleviate the overfitting and oversmoothing problems in the DeepGCN training process. We compare the WL-DeepGCN model with competitive models based on the same topics and nested 10-fold cross-validation show that our method achieves 77.27% accuracy and 0.83 AUC for ASD identification, bringing substantial performance gains.https://ieeexplore.ieee.org/document/10247643/Medical imagingautism spectrum disordersbrain networksgraph convolutional neural networkmultimodal |
| spellingShingle | Mingzhi Wang Jifeng Guo Yongjie Wang Ming Yu Jingtan Guo Multimodal Autism Spectrum Disorder Diagnosis Method Based on DeepGCN IEEE Transactions on Neural Systems and Rehabilitation Engineering Medical imaging autism spectrum disorders brain networks graph convolutional neural network multimodal |
| title | Multimodal Autism Spectrum Disorder Diagnosis Method Based on DeepGCN |
| title_full | Multimodal Autism Spectrum Disorder Diagnosis Method Based on DeepGCN |
| title_fullStr | Multimodal Autism Spectrum Disorder Diagnosis Method Based on DeepGCN |
| title_full_unstemmed | Multimodal Autism Spectrum Disorder Diagnosis Method Based on DeepGCN |
| title_short | Multimodal Autism Spectrum Disorder Diagnosis Method Based on DeepGCN |
| title_sort | multimodal autism spectrum disorder diagnosis method based on deepgcn |
| topic | Medical imaging autism spectrum disorders brain networks graph convolutional neural network multimodal |
| url | https://ieeexplore.ieee.org/document/10247643/ |
| work_keys_str_mv | AT mingzhiwang multimodalautismspectrumdisorderdiagnosismethodbasedondeepgcn AT jifengguo multimodalautismspectrumdisorderdiagnosismethodbasedondeepgcn AT yongjiewang multimodalautismspectrumdisorderdiagnosismethodbasedondeepgcn AT mingyu multimodalautismspectrumdisorderdiagnosismethodbasedondeepgcn AT jingtanguo multimodalautismspectrumdisorderdiagnosismethodbasedondeepgcn |