001021976 001__ 1021976
001021976 005__ 20240130202645.0
001021976 0247_ $$2doi$$a10.1101/2023.12.31.573801
001021976 0247_ $$2datacite_doi$$a10.34734/FZJ-2024-01115
001021976 037__ $$aFZJ-2024-01115
001021976 1001_ $$00009-0005-3974-7528$$aWulan, Naren$$b0
001021976 245__ $$aTranslating phenotypic prediction models from big to small anatomical MRI data using meta-matching
001021976 260__ $$c2024
001021976 3367_ $$0PUB:(DE-HGF)25$$2PUB:(DE-HGF)$$aPreprint$$bpreprint$$mpreprint$$s1706624600_7629
001021976 3367_ $$2ORCID$$aWORKING_PAPER
001021976 3367_ $$028$$2EndNote$$aElectronic Article
001021976 3367_ $$2DRIVER$$apreprint
001021976 3367_ $$2BibTeX$$aARTICLE
001021976 3367_ $$2DataCite$$aOutput Types/Working Paper
001021976 520__ $$aIndividualized phenotypic prediction based on structural MRI is an important goal in neuroscience. Prediction performance increases with larger samples, but small-scale datasets with fewer than 200 participants are often unavoidable. We have previously proposed a "meta-matching" framework to translate models trained from large datasets to improve the prediction of new unseen phenotypes in small collection efforts. Meta-matching exploits correlations between phenotypes, yielding large improvement over classical machine learning when applied to prediction models using resting-state functional connectivity as input features. Here, we adapt the two best performing meta-matching variants ("meta-matching finetune" and "meta-matching stacking") from our previous study to work with T1-weighted MRI data by changing the base neural network architecture to a 3D convolution neural network. We compare the two meta-matching variants with elastic net and classical transfer learning using the UK Biobank (N = 36,461), Human Connectome Project Young Adults (HCP-YA) dataset (N = 1,017) and HCP-Aging dataset (N = 656). We find that meta-matching outperforms elastic net and classical transfer learning by a large margin, both when translating models within the same dataset, as well as translating models across datasets with different MRI scanners, acquisition protocols and demographics. For example, when translating a UK Biobank model to 100 HCP-YA participants, meta-matching finetune yielded a 136% improvement in variance explained over transfer learning, with an average absolute gain of 2.6% (minimum = -0.9%, maximum = 17.6%) across 35 phenotypes. Overall, our results highlight the versatility of the meta-matching framework.
001021976 536__ $$0G:(DE-HGF)POF4-5251$$a5251 - Multilevel Brain Organization and Variability (POF4-525)$$cPOF4-525$$fPOF IV$$x0
001021976 588__ $$aDataset connected to CrossRef
001021976 7001_ $$00000-0003-1030-4625$$aAn, Lijun$$b1
001021976 7001_ $$aZhang, Chen$$b2
001021976 7001_ $$00000-0001-7842-0329$$aKong, Ru$$b3
001021976 7001_ $$00009-0009-8881-8643$$aChen, Pansheng$$b4
001021976 7001_ $$0P:(DE-Juel1)136848$$aBzdok, Danilo$$b5
001021976 7001_ $$0P:(DE-Juel1)131678$$aEickhoff, Simon B$$b6
001021976 7001_ $$00000-0001-6583-803X$$aHolmes, Avram J$$b7
001021976 7001_ $$00000-0002-0119-3276$$aYeo, B. T. Thomas$$b8
001021976 773__ $$a10.1101/2023.12.31.573801
001021976 8564_ $$uhttps://juser.fz-juelich.de/record/1021976/files/2023.12.31.573801v1.full.pdf$$yOpenAccess
001021976 8564_ $$uhttps://juser.fz-juelich.de/record/1021976/files/2023.12.31.573801v1.full.gif?subformat=icon$$xicon$$yOpenAccess
001021976 8564_ $$uhttps://juser.fz-juelich.de/record/1021976/files/2023.12.31.573801v1.full.jpg?subformat=icon-1440$$xicon-1440$$yOpenAccess
001021976 8564_ $$uhttps://juser.fz-juelich.de/record/1021976/files/2023.12.31.573801v1.full.jpg?subformat=icon-180$$xicon-180$$yOpenAccess
001021976 8564_ $$uhttps://juser.fz-juelich.de/record/1021976/files/2023.12.31.573801v1.full.jpg?subformat=icon-640$$xicon-640$$yOpenAccess
001021976 909CO $$ooai:juser.fz-juelich.de:1021976$$popenaire$$popen_access$$pVDB$$pdriver$$pdnbdelivery
001021976 9101_ $$0I:(DE-588b)5008462-8$$6P:(DE-Juel1)131678$$aForschungszentrum Jülich$$b6$$kFZJ
001021976 9131_ $$0G:(DE-HGF)POF4-525$$1G:(DE-HGF)POF4-520$$2G:(DE-HGF)POF4-500$$3G:(DE-HGF)POF4$$4G:(DE-HGF)POF$$9G:(DE-HGF)POF4-5251$$aDE-HGF$$bKey Technologies$$lNatural, Artificial and Cognitive Information Processing$$vDecoding Brain Organization and Dysfunction$$x0
001021976 9141_ $$y2024
001021976 915__ $$0StatID:(DE-HGF)0510$$2StatID$$aOpenAccess
001021976 915__ $$0LIC:(DE-HGF)CCBYNCND4$$2HGFVOC$$aCreative Commons Attribution-NonCommercial-NoDerivs CC BY-NC-ND 4.0
001021976 920__ $$lyes
001021976 9201_ $$0I:(DE-Juel1)INM-7-20090406$$kINM-7$$lGehirn & Verhalten$$x0
001021976 980__ $$apreprint
001021976 980__ $$aVDB
001021976 980__ $$aUNRESTRICTED
001021976 980__ $$aI:(DE-Juel1)INM-7-20090406
001021976 9801_ $$aFullTexts