Incomplete multi-modal representation learning for Alzheimer’s disease diagnosis

• Published in: Medical image analysis 2021-04, Vol.69, p.101953-101953, Article 101953
• Main Authors: Liu, Yanbei, Fan, Lianxi, Zhang, Changqing, Zhou, Tao, Xiao, Zhitao, Geng, Lei, Shen, Dinggang
• Format: Article
• Language: English
• Subjects: Algorithms; Alzheimer's disease; Alzheimers disease diagnosis; auto-encoder network; Coders; Diagnosis; incomplete multi-modality data; kernel completion; Kernels; Learning algorithms; Machine learning; Medical diagnosis; Medical imaging; Medical treatment; Missing data; multi-modal representation learning; Neurodegenerative diseases; Neuroimaging; Regularization; Representations
• ISSN: 1361-8415; 1361-8423
• DOI: 10.1016/j.media.2020.101953

•We study a multi-modal representation learning problem for Alzheimers disease diagnosis with incomplete modalities.•We propose an Auto-Encoder based Multi-View missing data Completion framework called AEMVC, that is able to complement the missing data in the kernel space.•Our framework can jointly optimize two stages of completion and learning representations while considering the structural information of data and the inherent association between multiple modalities.•Extensive experiments are conducted to evaluate the effectiveness of the proposed method. [Display omitted] Alzheimers disease (AD) is a complex neurodegenerative disease. Its early diagnosis and treatment have been a major concern of researchers. Currently, the multi-modality data representation learning of this disease is gradually becoming an emerging research field, attracting widespread attention. However, in practice, data from multiple modalities are only partially available, and most of the existing multi-modal learning algorithms can not deal with the incomplete multi-modality data. In this paper, we propose an Auto-Encoder based Multi-View missing data Completion framework (AEMVC) to learn common representations for AD diagnosis. Specifically, we firstly map the original complete view to a latent space using an auto-encoder network framework. Then, the latent representations measuring statistical dependence learned from the complete view are used to complement the kernel matrix of the incomplete view in the kernel space. Meanwhile, the structural information of original data and the inherent association between views are maintained by graph regularization and Hilbert-Schmidt Independence Criterion (HSIC) constraints. Finally, a kernel based multi-view method is applied to the learned kernel matrix for the acquisition of common representations. Experimental results achieved on Alzheimers Disease Neuroimaging Initiative (ADNI) datasets validate the effectiveness of the proposed method.