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 proceedingConference contribution

9 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 languageEnglish
Title of host publicationDeep 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
PublisherSpringer Verlag
Pages160-168
Number of pages9
Volume10553 LNCS
ISBN (Print)9783319675572
DOIs
Publication statusPublished - 2017
Externally publishedYes
Event3rd 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 Sep 142017 Sep 14

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume10553 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other3rd 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
CountryCanada
CityQuebec City
Period17/9/1417/9/14

Fingerprint

Multi-task Learning
Alzheimer's Disease
Neurodegenerative diseases
Modality
Incomplete Data
Labels
Availability
Matrix Completion
Diagnostic Accuracy
Multimodality
Imputation
Missing Data
Updating
Limiting
Deep learning
Prediction
Output
Experimental Results

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

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 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 (Vol. 10553 LNCS, 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. Vol. 10553 LNCS 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 proceedingConference 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 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. vol. 10553 LNCS, 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 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. Vol. 10553 LNCS. Springer Verlag. 2017. p. 160-168. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/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. Vol. 10553 LNCS 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’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",
year = "2017",
doi = "10.1007/978-3-319-67558-9_19",
language = "English",
isbn = "9783319675572",
volume = "10553 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "160--168",
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

UR - http://www.scopus.com/inward/citedby.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

VL - 10553 LNCS

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

PB - Springer Verlag

ER -