TY - GEN
T1 - Alzheimer’s Disease Diagnosis via Specific-Shared Representation Learning in Multimodal Neuroimaging
AU - Wang, Mingxia
AU - Yang, Yun
AU - Qi, Jun
AU - Nan, Fengtao
AU - Li, Shunbao
AU - Yang, Po
N1 - Publisher Copyright:
© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.
PY - 2024
Y1 - 2024
N2 - Mild cognitive impairment (MCI) is a potential prodromal state of Alzheimer’s disease (AD), an irreversible progressive neurodegenerative disease. Early diagnosis and intervention of AD are crucial. Recent studies have shown that deep learning methods exhibit excellent performance in predicting AD from multimodal neuroimaging. However, previous studies tend to focus on one aspect of the analysis of intra-modal information or inter-modal shared information, while ignoring the integration of the two. To solve this problem, this paper proposes a fusion network that combines modality-specific representations with shared representations between modalities to assist AD diagnosis. The network first extracts modality-specific representations from each modality data, then introduces a modality shared representation extraction network to supplement modality shared information, and finally performs modality representation fusion and performs AD prediction. In addition, in order to reduce the interference of non-disease-related regions between modalities on shared information, the network combines the shallow representation extracted by the single-modal network to guide the extraction process of shared representation. Combining these representations can provide a comprehensive view of multimodal data, thereby more accurately assisting in AD prediction tasks. The experimental results of this article on the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database prove the effectiveness and excellent performance of this model in assisting in the diagnosis of AD.
AB - Mild cognitive impairment (MCI) is a potential prodromal state of Alzheimer’s disease (AD), an irreversible progressive neurodegenerative disease. Early diagnosis and intervention of AD are crucial. Recent studies have shown that deep learning methods exhibit excellent performance in predicting AD from multimodal neuroimaging. However, previous studies tend to focus on one aspect of the analysis of intra-modal information or inter-modal shared information, while ignoring the integration of the two. To solve this problem, this paper proposes a fusion network that combines modality-specific representations with shared representations between modalities to assist AD diagnosis. The network first extracts modality-specific representations from each modality data, then introduces a modality shared representation extraction network to supplement modality shared information, and finally performs modality representation fusion and performs AD prediction. In addition, in order to reduce the interference of non-disease-related regions between modalities on shared information, the network combines the shallow representation extracted by the single-modal network to guide the extraction process of shared representation. Combining these representations can provide a comprehensive view of multimodal data, thereby more accurately assisting in AD prediction tasks. The experimental results of this article on the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database prove the effectiveness and excellent performance of this model in assisting in the diagnosis of AD.
KW - Alzheimer’s Disease Diagnosis
KW - Multimodal Neuroimaging
KW - Multimodal Representation Learning
UR - http://www.scopus.com/inward/record.url?scp=85200945769&partnerID=8YFLogxK
U2 - 10.1007/978-981-97-5689-6_1
DO - 10.1007/978-981-97-5689-6_1
M3 - Conference Proceeding
AN - SCOPUS:85200945769
SN - 9789819756889
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 3
EP - 14
BT - Advanced Intelligent Computing in Bioinformatics - 20th International Conference, ICIC 2024, Proceedings
A2 - Huang, De-Shuang
A2 - Zhang, Qinhu
A2 - Guo, Jiayang
PB - Springer Science and Business Media Deutschland GmbH
T2 - 20th International Conference on Intelligent Computing , ICIC 2024
Y2 - 5 August 2024 through 8 August 2024
ER -