Multi-stage diagnosis of Alzheimer’s disease with incomplete multimodal data via multi-task deep learning

Kim Han Thung; Pew Thian Yap; Dinggang Shen

doi:10.1007/978-3-319-67558-9_19

Multi-stage diagnosis of Alzheimer’s disease with incomplete multimodal data via multi-task deep learning

Kim Han Thung, Pew Thian Yap, Dinggang Shen

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

53 Citations (Scopus)

Abstract

Utilization of biomedical data from multiple modalities improves the diagnostic accuracy of neurodegenerative diseases. However, multi-modality data are often incomplete because not all data can be collected for every individual. When using such incomplete data for diagnosis, current approaches for addressing the problem of missing data, such as imputation, matrix completion and multi-task learning, implicitly assume linear data-to-label relationship, therefore limiting their performances. We thus propose multi-task deep learning for incomplete data, where prediction tasks that are associated with different modality combinations are learnt jointly to improve the performance of each task. Specifically, we devise a multi-input multi-output deep learning framework, and train our deep network subnet-wise, partially updating its weights based on the availability of modality data. The experimental results using the ADNI dataset show that our method outperforms the state-of-the-art methods.

Original language	English
Title of host publication	Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support - 3rd International Workshop, DLMIA 2017 and 7th International Workshop, ML-CDS 2017 Held in Conjunction with MICCAI 2017, Proceedings
Editors	Tal Arbel, M. Jorge Cardoso
Publisher	Springer Verlag
Pages	160-168
Number of pages	9
ISBN (Print)	9783319675572
DOIs	https://doi.org/10.1007/978-3-319-67558-9_19
Publication status	Published - 2017
Event	3rd International Workshop on Deep Learning in Medical Image Analysis, DLMIA 2017 and 7th International Workshop on Multimodal Learning for Clinical Decision Support, ML-CDS 2017 held in Conjunction with 20th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2017 - Quebec City, Canada Duration: 2017 Sept 14 → 2017 Sept 14

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	10553 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Other

Other	3rd International Workshop on Deep Learning in Medical Image Analysis, DLMIA 2017 and 7th International Workshop on Multimodal Learning for Clinical Decision Support, ML-CDS 2017 held in Conjunction with 20th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2017
Country/Territory	Canada
City	Quebec City
Period	17/9/14 → 17/9/14

Bibliographical note

Publisher Copyright:
© Springer International Publishing AG 2017.

ASJC Scopus subject areas

Theoretical Computer Science
General Computer Science

Access to Document

10.1007/978-3-319-67558-9_19

Cite this

Thung, K. H., Yap, P. T., & Shen, D. (2017). Multi-stage diagnosis of Alzheimer’s disease with incomplete multimodal data via multi-task deep learning. In T. Arbel, & M. J. Cardoso (Eds.), Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support - 3rd International Workshop, DLMIA 2017 and 7th International Workshop, ML-CDS 2017 Held in Conjunction with MICCAI 2017, Proceedings (pp. 160-168). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10553 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-67558-9_19

Multi-stage diagnosis of Alzheimer’s disease with incomplete multimodal data via multi-task deep learning. / Thung, Kim Han; Yap, Pew Thian; Shen, Dinggang.
Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support - 3rd International Workshop, DLMIA 2017 and 7th International Workshop, ML-CDS 2017 Held in Conjunction with MICCAI 2017, Proceedings. ed. / Tal Arbel; M. Jorge Cardoso. Springer Verlag, 2017. p. 160-168 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10553 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Thung, KH, Yap, PT & Shen, D 2017, Multi-stage diagnosis of Alzheimer’s disease with incomplete multimodal data via multi-task deep learning. in T Arbel & MJ Cardoso (eds), Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support - 3rd International Workshop, DLMIA 2017 and 7th International Workshop, ML-CDS 2017 Held in Conjunction with MICCAI 2017, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 10553 LNCS, Springer Verlag, pp. 160-168, 3rd International Workshop on Deep Learning in Medical Image Analysis, DLMIA 2017 and 7th International Workshop on Multimodal Learning for Clinical Decision Support, ML-CDS 2017 held in Conjunction with 20th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2017, Quebec City, Canada, 17/9/14. https://doi.org/10.1007/978-3-319-67558-9_19

Thung KH, Yap PT, Shen D. Multi-stage diagnosis of Alzheimer’s disease with incomplete multimodal data via multi-task deep learning. In Arbel T, Cardoso MJ, editors, Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support - 3rd International Workshop, DLMIA 2017 and 7th International Workshop, ML-CDS 2017 Held in Conjunction with MICCAI 2017, Proceedings. Springer Verlag. 2017. p. 160-168. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-319-67558-9_19

Thung, Kim Han ; Yap, Pew Thian ; Shen, Dinggang. / Multi-stage diagnosis of Alzheimer’s disease with incomplete multimodal data via multi-task deep learning. Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support - 3rd International Workshop, DLMIA 2017 and 7th International Workshop, ML-CDS 2017 Held in Conjunction with MICCAI 2017, Proceedings. editor / Tal Arbel ; M. Jorge Cardoso. Springer Verlag, 2017. pp. 160-168 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{7b7025a3226d466daaf08421b5895d1c,

title = "Multi-stage diagnosis of Alzheimer{\textquoteright}s disease with incomplete multimodal data via multi-task deep learning",

abstract = "Utilization of biomedical data from multiple modalities improves the diagnostic accuracy of neurodegenerative diseases. However, multi-modality data are often incomplete because not all data can be collected for every individual. When using such incomplete data for diagnosis, current approaches for addressing the problem of missing data, such as imputation, matrix completion and multi-task learning, implicitly assume linear data-to-label relationship, therefore limiting their performances. We thus propose multi-task deep learning for incomplete data, where prediction tasks that are associated with different modality combinations are learnt jointly to improve the performance of each task. Specifically, we devise a multi-input multi-output deep learning framework, and train our deep network subnet-wise, partially updating its weights based on the availability of modality data. The experimental results using the ADNI dataset show that our method outperforms the state-of-the-art methods.",

author = "Thung, {Kim Han} and Yap, {Pew Thian} and Dinggang Shen",

note = "Publisher Copyright: {\textcopyright} Springer International Publishing AG 2017.; 3rd International Workshop on Deep Learning in Medical Image Analysis, DLMIA 2017 and 7th International Workshop on Multimodal Learning for Clinical Decision Support, ML-CDS 2017 held in Conjunction with 20th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2017 ; Conference date: 14-09-2017 Through 14-09-2017",

year = "2017",

doi = "10.1007/978-3-319-67558-9_19",

language = "English",

isbn = "9783319675572",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "160--168",

editor = "Tal Arbel and Cardoso, {M. Jorge}",

booktitle = "Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support - 3rd International Workshop, DLMIA 2017 and 7th International Workshop, ML-CDS 2017 Held in Conjunction with MICCAI 2017, Proceedings",

}

TY - GEN

T1 - Multi-stage diagnosis of Alzheimer’s disease with incomplete multimodal data via multi-task deep learning

AU - Thung, Kim Han

AU - Yap, Pew Thian

AU - Shen, Dinggang

PY - 2017

Y1 - 2017

N2 - Utilization of biomedical data from multiple modalities improves the diagnostic accuracy of neurodegenerative diseases. However, multi-modality data are often incomplete because not all data can be collected for every individual. When using such incomplete data for diagnosis, current approaches for addressing the problem of missing data, such as imputation, matrix completion and multi-task learning, implicitly assume linear data-to-label relationship, therefore limiting their performances. We thus propose multi-task deep learning for incomplete data, where prediction tasks that are associated with different modality combinations are learnt jointly to improve the performance of each task. Specifically, we devise a multi-input multi-output deep learning framework, and train our deep network subnet-wise, partially updating its weights based on the availability of modality data. The experimental results using the ADNI dataset show that our method outperforms the state-of-the-art methods.

AB - Utilization of biomedical data from multiple modalities improves the diagnostic accuracy of neurodegenerative diseases. However, multi-modality data are often incomplete because not all data can be collected for every individual. When using such incomplete data for diagnosis, current approaches for addressing the problem of missing data, such as imputation, matrix completion and multi-task learning, implicitly assume linear data-to-label relationship, therefore limiting their performances. We thus propose multi-task deep learning for incomplete data, where prediction tasks that are associated with different modality combinations are learnt jointly to improve the performance of each task. Specifically, we devise a multi-input multi-output deep learning framework, and train our deep network subnet-wise, partially updating its weights based on the availability of modality data. The experimental results using the ADNI dataset show that our method outperforms the state-of-the-art methods.

UR - http://www.scopus.com/inward/record.url?scp=85029806638&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-67558-9_19

DO - 10.1007/978-3-319-67558-9_19

M3 - Conference contribution

AN - SCOPUS:85029806638

SN - 9783319675572

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 160

EP - 168

BT - Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support - 3rd International Workshop, DLMIA 2017 and 7th International Workshop, ML-CDS 2017 Held in Conjunction with MICCAI 2017, Proceedings

A2 - Arbel, Tal

A2 - Cardoso, M. Jorge

PB - Springer Verlag

T2 - 3rd International Workshop on Deep Learning in Medical Image Analysis, DLMIA 2017 and 7th International Workshop on Multimodal Learning for Clinical Decision Support, ML-CDS 2017 held in Conjunction with 20th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2017

Y2 - 14 September 2017 through 14 September 2017

ER -

Multi-stage diagnosis of Alzheimer’s disease with incomplete multimodal data via multi-task deep learning

Abstract

Publication series

Other

Bibliographical note

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this