PSYCH OpenIR  > 中国科学院心理健康重点实验室
Generalizability of machine learning for classification of schizophrenia based on resting-state functional MRI data
Cai, Xin-Lu1,2,3; Xie, Dong-Jie1,4; Madsen, Kristoffer H.3,5,6; Wang, Yong-Ming1,2,3; Bogemann, Sophie Alida1,2,3; Cheung, Eric F. C.7; Moller, Arne3,8,9; Chan, Raymond C. K.1,2,3,10
First AuthorCai, Xin-Lu
Correspondent Emailrckchan@psych.ac.cn(chan, raymond c. k.)
2019-10-01
Source PublicationHUMAN BRAIN MAPPING
ISSN1065-9471
SubtypeArticle
Pages13
Contribution Rank1
Abstract

Machine learning has increasingly been applied to classification of schizophrenia in neuroimaging research. However, direct replication studies and studies seeking to investigate generalizability are scarce. To address these issues, we assessed within-site and between-site generalizability of a machine learning classification framework which achieved excellent performance in a previous study using two independent resting-state functional magnetic resonance imaging data sets collected from different sites and scanners. We established within-site generalizability of the classification framework in the main data set using cross-validation. Then, we trained a model in the main data set and investigated between-site generalization in the validated data set using external validation. Finally, recognizing the poor between-site generalization performance, we updated the unsupervised algorithm to investigate if transfer learning using additional unlabeled data were able to improve between-site classification performance. Cross-validation showed that the published classification procedure achieved an accuracy of 0.73 using majority voting across all selected components. External validation found a classification accuracy of 0.55 (not significant) and 0.70 (significant) using the direct and transfer learning procedures, respectively. The failure of direct generalization from one site to another demonstrates the limitation of within-site cross-validation and points toward the need to incorporate efforts to facilitate application of machine learning across multiple data sets. The improvement in performance with transfer learning highlights the importance of taking into account the properties of data when constructing predictive models across samples and sites. Our findings suggest that machine learning classification result based on a single study should be interpreted cautiously.

Keywordgeneralizability machine learning reproducibility schizophrenia spectrum disorders
DOI10.1002/hbm.24797
Indexed BySCI
Language英语
Funding OrganizationBeijing Municipal Science & Technology Commission Grant ; National Key Research and Development Programme ; National Natural Science Foundation of China ; CAS key Laboratory of Mental Health
Funding ProjectBeijing Municipal Science & Technology Commission Grant[Z161100000216138] ; National Key Research and Development Programme[2016YFC0906402] ; National Natural Science Foundation of China[81571317] ; CAS key Laboratory of Mental Health
WOS Research AreaNeurosciences & Neurology ; Radiology, Nuclear Medicine & Medical Imaging
WOS SubjectNeurosciences ; Neuroimaging ; Radiology, Nuclear Medicine & Medical Imaging
WOS IDWOS:000488298800001
PublisherWILEY
WOS KeywordCOMPUTER-AIDED DIAGNOSIS ; CONNECTIVITY PATTERNS ; TEMPORAL GYRUS ; NETWORK ; ONSET ; ABNORMALITIES ; METAANALYSIS ; BIOMARKERS ; MOVEMENT ; FEATURES
Citation statistics
Cited Times:1[WOS]   [WOS Record]     [Related Records in WOS]
Document Type期刊论文
Identifierhttp://ir.psych.ac.cn/handle/311026/30156
Collection中国科学院心理健康重点实验室
Corresponding AuthorChan, Raymond C. K.
Affiliation1.Inst Psychol, Neuropsychol & Appl Cognit Neurosci Lab, CAS Key Lab Mental Hlth, Beijing, Peoples R China
2.Univ Chinese Acad Sci, Sinodanish Coll, Beijing, Peoples R China
3.Sinodanish Ctr Educ & Res, Beijing, Peoples R China
4.Zhejiang Normal Univ, Hangzhou Coll Presch Teacher Educ, Hangzhou, Zhejiang, Peoples R China
5.Univ Copenhagen, Hosp Hvidovre, Danish Res Ctr Magnet Resonance, Ctr Funct & Diagnost Imaging & Res, Copenhagen, Denmark
6.Tech Univ Denmark, Dept Appl Math & Comp Sci, Lyngby, Denmark
7.Castle Peak Hosp, Hong Kong, Peoples R China
8.Aarhus Univ Hosp, Dept Nucl Med, Aarhus, Denmark
9.Aarhus Univ Hosp, PET Ctr, Aarhus, Denmark
10.Univ Chinese Acad Sci, Dept Psychol, Beijing, Peoples R China
Recommended Citation
GB/T 7714
Cai, Xin-Lu,Xie, Dong-Jie,Madsen, Kristoffer H.,et al. Generalizability of machine learning for classification of schizophrenia based on resting-state functional MRI data[J]. HUMAN BRAIN MAPPING,2019:13.
APA Cai, Xin-Lu.,Xie, Dong-Jie.,Madsen, Kristoffer H..,Wang, Yong-Ming.,Bogemann, Sophie Alida.,...&Chan, Raymond C. K..(2019).Generalizability of machine learning for classification of schizophrenia based on resting-state functional MRI data.HUMAN BRAIN MAPPING,13.
MLA Cai, Xin-Lu,et al."Generalizability of machine learning for classification of schizophrenia based on resting-state functional MRI data".HUMAN BRAIN MAPPING (2019):13.
Files in This Item:
File Name/Size DocType Version Access License
Generalizability of (1595KB)期刊论文出版稿限制开放CC BY-NC-SAApplication Full Text
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[Cai, Xin-Lu]'s Articles
[Xie, Dong-Jie]'s Articles
[Madsen, Kristoffer H.]'s Articles
Baidu academic
Similar articles in Baidu academic
[Cai, Xin-Lu]'s Articles
[Xie, Dong-Jie]'s Articles
[Madsen, Kristoffer H.]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[Cai, Xin-Lu]'s Articles
[Xie, Dong-Jie]'s Articles
[Madsen, Kristoffer H.]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.
 

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.