Schizophrenia recognition based on three-dimensional adaptive graph convolutional neural network
Abstract Previous deep learning-based brain network research has made significant progress in understanding the pathophysiology of schizophrenia. However, it ignores the three-dimensional spatial characteristics of EEG signals and cannot dynamically learn the interactions between nodes. To address t...
Saved in:
| Main Authors: | , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-02-01
|
| Series: | Scientific Reports |
| Subjects: | |
| Online Access: | https://doi.org/10.1038/s41598-024-84497-8 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1823862504956100608 |
|---|---|
| author | Guimei Yin Jie Yuan Yanjun Chen Guangxing Guo Dongli Shi Lin Wang Zilong Zhao Yanli Zhao Manjie Zhang Yuan Dong Bin Wang Shuping Tan |
| author_facet | Guimei Yin Jie Yuan Yanjun Chen Guangxing Guo Dongli Shi Lin Wang Zilong Zhao Yanli Zhao Manjie Zhang Yuan Dong Bin Wang Shuping Tan |
| author_sort | Guimei Yin |
| collection | DOAJ |
| description | Abstract Previous deep learning-based brain network research has made significant progress in understanding the pathophysiology of schizophrenia. However, it ignores the three-dimensional spatial characteristics of EEG signals and cannot dynamically learn the interactions between nodes. To address this issue, a schizophrenia classification model based on a three-dimensional adaptive graph convolutional neural network (3D-AGCN) is proposed. Each subject’s EEG data is divided into various segment lengths and frequency bands for the experiment. The attention mechanism is then used to integrate the node features in the spatial, feature, and frequency band dimensions. The resulting adaptive brain functional network features are then constructed and fed into the GAT + GCN model. This adaptive approach eliminates the human-specified criteria for feature selection and brain network construction. The trial results demonstrated that, when using a 6-second segment length and time-domain and frequency-domain features, patients with first-episode schizophrenia achieved the highest classification accuracy of 87.64% This method outperforms other feature selection and brain network modeling approaches, providing new insights and directions for the early diagnosis and recognition of schizophrenia. |
| format | Article |
| id | doaj-art-a614ba2481dd47f9b6ebf42ecc815f9d |
| institution | Kabale University |
| issn | 2045-2322 |
| language | English |
| publishDate | 2025-02-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj-art-a614ba2481dd47f9b6ebf42ecc815f9d2025-02-09T12:31:20ZengNature PortfolioScientific Reports2045-23222025-02-0115111110.1038/s41598-024-84497-8Schizophrenia recognition based on three-dimensional adaptive graph convolutional neural networkGuimei Yin0Jie Yuan1Yanjun Chen2Guangxing Guo3Dongli Shi4Lin Wang5Zilong Zhao6Yanli Zhao7Manjie Zhang8Yuan Dong9Bin Wang10Shuping Tan11School of Computer Science and Technology, Taiyuan Normal UniversityDepartment of Radiology, Shanxi Provincial People’s HospitalSchool of Computer Science and Technology, Taiyuan Normal UniversityInstitute of Big Data Technology Analysis and Application, Taiyuan Normal UniversitySchool of Computer Science and Technology, Taiyuan Normal UniversitySchool of Computer Science and Technology, Taiyuan Normal UniversitySchool of Chemical Engineering and Technology, Sun Yat-sen UniversityPsychiatry Research Center, Beijing Huilongguan HospitalSchool of Computer Science and Technology, Taiyuan Normal UniversitySchool of Computer Science and Technology, Taiyuan Normal UniversityCollege of Computer Science and Technology, Taiyuan University of TechnologyPsychiatry Research Center, Beijing Huilongguan HospitalAbstract Previous deep learning-based brain network research has made significant progress in understanding the pathophysiology of schizophrenia. However, it ignores the three-dimensional spatial characteristics of EEG signals and cannot dynamically learn the interactions between nodes. To address this issue, a schizophrenia classification model based on a three-dimensional adaptive graph convolutional neural network (3D-AGCN) is proposed. Each subject’s EEG data is divided into various segment lengths and frequency bands for the experiment. The attention mechanism is then used to integrate the node features in the spatial, feature, and frequency band dimensions. The resulting adaptive brain functional network features are then constructed and fed into the GAT + GCN model. This adaptive approach eliminates the human-specified criteria for feature selection and brain network construction. The trial results demonstrated that, when using a 6-second segment length and time-domain and frequency-domain features, patients with first-episode schizophrenia achieved the highest classification accuracy of 87.64% This method outperforms other feature selection and brain network modeling approaches, providing new insights and directions for the early diagnosis and recognition of schizophrenia.https://doi.org/10.1038/s41598-024-84497-8Schizophrenia3D spacesAttention mechanismsAdaptive brain networksGraph convolutional neural network |
| spellingShingle | Guimei Yin Jie Yuan Yanjun Chen Guangxing Guo Dongli Shi Lin Wang Zilong Zhao Yanli Zhao Manjie Zhang Yuan Dong Bin Wang Shuping Tan Schizophrenia recognition based on three-dimensional adaptive graph convolutional neural network Scientific Reports Schizophrenia 3D spaces Attention mechanisms Adaptive brain networks Graph convolutional neural network |
| title | Schizophrenia recognition based on three-dimensional adaptive graph convolutional neural network |
| title_full | Schizophrenia recognition based on three-dimensional adaptive graph convolutional neural network |
| title_fullStr | Schizophrenia recognition based on three-dimensional adaptive graph convolutional neural network |
| title_full_unstemmed | Schizophrenia recognition based on three-dimensional adaptive graph convolutional neural network |
| title_short | Schizophrenia recognition based on three-dimensional adaptive graph convolutional neural network |
| title_sort | schizophrenia recognition based on three dimensional adaptive graph convolutional neural network |
| topic | Schizophrenia 3D spaces Attention mechanisms Adaptive brain networks Graph convolutional neural network |
| url | https://doi.org/10.1038/s41598-024-84497-8 |
| work_keys_str_mv | AT guimeiyin schizophreniarecognitionbasedonthreedimensionaladaptivegraphconvolutionalneuralnetwork AT jieyuan schizophreniarecognitionbasedonthreedimensionaladaptivegraphconvolutionalneuralnetwork AT yanjunchen schizophreniarecognitionbasedonthreedimensionaladaptivegraphconvolutionalneuralnetwork AT guangxingguo schizophreniarecognitionbasedonthreedimensionaladaptivegraphconvolutionalneuralnetwork AT donglishi schizophreniarecognitionbasedonthreedimensionaladaptivegraphconvolutionalneuralnetwork AT linwang schizophreniarecognitionbasedonthreedimensionaladaptivegraphconvolutionalneuralnetwork AT zilongzhao schizophreniarecognitionbasedonthreedimensionaladaptivegraphconvolutionalneuralnetwork AT yanlizhao schizophreniarecognitionbasedonthreedimensionaladaptivegraphconvolutionalneuralnetwork AT manjiezhang schizophreniarecognitionbasedonthreedimensionaladaptivegraphconvolutionalneuralnetwork AT yuandong schizophreniarecognitionbasedonthreedimensionaladaptivegraphconvolutionalneuralnetwork AT binwang schizophreniarecognitionbasedonthreedimensionaladaptivegraphconvolutionalneuralnetwork AT shupingtan schizophreniarecognitionbasedonthreedimensionaladaptivegraphconvolutionalneuralnetwork |