Journal article
Data-driven biomarkers better associate with stroke motor outcomes than theory-based biomarkers
Brain communications, Vol.6(4), fcae254
07/02/2024
DOI: 10.1093/braincomms/fcae254
PMCID: PMC11336660
PMID: 39171205
Abstract
Abstract Chronic motor impairments are a leading cause of disability after stroke. Previous studies have associated motor outcomes with the degree of damage to predefined structures in the motor system, such as the corticospinal tract. However, such theory-based approaches may not take full advantage of the information contained in clinical imaging data. The present study uses data-driven approaches to model chronic motor outcomes after stroke and compares the accuracy of these associations to previously-identified theory-based biomarkers. Using a cross-validation framework, regression models were trained using lesion masks and motor outcomes data from 789 stroke patients from the Enhancing NeuroImaging Genetics through Meta Analysis (ENIGMA) Stroke Recovery Working Group. Using the explained variance metric to measure the strength of the association between chronic motor outcomes and imaging biomarkers, we compared theory-based biomarkers, like lesion load to known motor tracts, to three data-driven biomarkers: lesion load of lesion-behaviour maps, lesion load of structural networks associated with lesion-behaviour maps, and measures of regional structural disconnection. In general, data-driven biomarkers had stronger associations with chronic motor outcomes accuracy than theory-based biomarkers. Data-driven models of regional structural disconnection performed the best of all models tested (R2 = 0.210, p < 0.001), performing significantly better than the theory-based biomarkers of lesion load of the corticospinal tract (R2 = 0.132, p< 0.001) and of multiple descending motor tracts (R2 = 0.180, p < 0.001). They also performed slightly, but significantly, better than other data-driven biomarkers including lesion load of lesion-behaviour maps (R2 =0.200, p < 0.001) and lesion load of structural networks associated with lesion-behaviour maps (R2 =0.167, p < 0.001). Ensemble models - combining basic demographic variables like age, sex, and time since stroke - improved the strength of associations for theory-based and data-driven biomarkers. Combining both theory-based and data-driven biomarkers with demographic variables improved predictions, and the best ensemble model achieved R2 = 0.241, p < 0.001. Overall, these results demonstrate that out-of-sample associations between chronic motor outcomes and data-driven imaging features, particularly when lesion data is represented in terms of structural disconnection, are stronger than associations between chronic motor outcomes and theory-based biomarkers. However, combining both theory-based and data-driven models provides the most robust associations.
Details
- Title: Subtitle
- Data-driven biomarkers better associate with stroke motor outcomes than theory-based biomarkers
- Creators
- Emily R Olafson - Weill Cornell MedicineChristoph Sperber - University Hospital of BernKeith W Jamison - Weill Cornell MedicineMark D Bowren Jr - University of Iowa, PsychiatryAaron D Boes - University of IowaJustin W Andrushko - University of British ColumbiaMichael R Borich - Emory UniversityLara A Boyd - University of British ColumbiaJessica M Cassidy - University of North Carolina at Chapel HillAdriana B Conforto - Hospital Israelita Albert EinsteinSteven C Cramer - UCLA HealthAdrienne N Dula - The University of Texas at AustinFatemeh Geranmayeh - Imperial College LondonBrenton Hordacre - University of South AustraliaNeda Jahanshad - University of Southern CaliforniaSteven A Kautz - Medical University of South CarolinaBethany P Tavenner - University of Southern CaliforniaBradley J MacIntosh - Oslo University HospitalFabrizio Piras - Fondazione Santa LuciaAndrew D Robertson - Sunnybrook Health Science CentreNa Jin Seo - Medical University of South CarolinaSurjo R Soekadar - Charité - Universitätsmedizin BerlinSophia I Thomopoulos - University of Southern CaliforniaDaniela Vecchio - Fondazione Santa LuciaTimothy B Weng - The University of Texas at AustinLars T Westlye - University of OsloCarolee J Winstein - University of Southern CaliforniaGeorge F Wittenberg - University of PittsburghKristin A Wong - The University of Texas at AustinPaul M Thompson - University of Southern CaliforniaSook-Lei Liew - University of Southern CaliforniaAmy F Kuceyeski - Weill Cornell Medicine
- Resource Type
- Journal article
- Publication Details
- Brain communications, Vol.6(4), fcae254
- Publisher
- OXFORD UNIV PRESS
- DOI
- 10.1093/braincomms/fcae254
- PMID
- 39171205
- PMCID
- PMC11336660
- ISSN
- 2632-1297
- eISSN
- 2632-1297
- Grant note
- Canadian Institutes of Health Research: PJT-148535, PJT-153330, MOP-106651 National Institute of General Medical Sciences: T32GM108540 National Institutes of Health: 1S10RR028821-01, R00 HD091375, R01 NS076348, IIEP-2250-14, UH3-NS121565, U01 NS120910, R01HD095457, R01 NS115845, U01 NS086872, R01 HD062744, R01 HD095457, P20GM109040, R01MH123163, R01MH121246, R01MH116147 Wellcome Trust: 093957 National Institutes of Health/National Institute of Child Health and Human Development: R01 HD094731 National Institutes of Health/National Institute of General Medical Sciences: P20 GM109040 European Research Council: 759370 DFG: 932/7 Federal Ministry of Education and Research: 01DR21025A Enhancing NeuroImaging Genetics through Meta Analysis Big Data to Knowledge (BD2K): U54 EB020403 European Research Council under the European Union: 802998 National Institutes of Health Big Data to Knowledge (BD2K) programme: U54 EB020403
J.W.A. is supported by the Canadian Institutes of Health Research and Michael Smith Foundation for Health Research Postdoctoral Fellowships. M.D.B. is supported by T32GM108540 (National Institute of General Medical Sciences) and 1S10RR028821-01 (National Institutes of Health). J.M.C. is supported by the National Institutes of Health R00 HD091375. A.B.C. is supported by National Institutes of Health R01 NS076348 and IIEP-2250-14. S.C.C. is supported by UH3-NS121565, U01 NS120910, R01HD095457, R01 NS115845, U01 NS086872, R01 HD062744, and R01 HD095457. A.N.D. is supported by the Lone Star Stroke Research Consortium. F.G. is supported by the Wellcome Trust (093957). S.A.K. is supported by National Institutes of Health P20GM109040. F.P. is supported by the Italian Ministry of Health, Ricerca Corrente 23. N.J.S. is supported by National Institutes of Health/National Institute of Child Health and Human Development R01 HD094731 and National Institutes of Health/National Institute of General Medical Sciences P20 GM109040. S.R.S. is supported by European Research Council 759370, DFG 932/7, and Federal Ministry of Education and Research (Bundesministerium fuer Bildung und Forschung) 01DR21025A. S.I.T. is supported in part by National Institutes of Health grants R01MH123163, R01MH121246, and R01MH116147. P.M.T. is supported by Enhancing NeuroImaging Genetics through Meta Analysis Big Data to Knowledge (BD2K) grant U54 EB020403. D.V. is supported by the Italian Ministry of Health-RC 2022/2023. T.B.W. is supported by Lonestar Stoke. L.T.W. is supported by The European Research Council under the European Union's Horizon 2020 research and Innovation program (grant 802998). C.J. Winstein is supported by grants HD065438 and HD104296. G.F.W. is supported by the Dept. of Veterans Affairs Rehabilitation Research & Development Service. Core funding for Enhancing NeuroImaging Genetics through Meta Analysis was provided by the National Institutes of Health Big Data to Knowledge (BD2K) programme under consortium grant U54 EB020403 to P.M.T. S.-L.L is supported by R01NS115845. L.A.B. is supported by Canadian Institutes of Health Research PJT-148535; PJT-153330; MOP-106651 (LAB PI). DAS:To protect the privacy of research participants, individual subject data used in this study are not available in a public repository. Participating research cohorts vary in public data-sharing restrictions as determined by the following: (1) ethical review board and consent documents; (2) national and transnational sharing laws; and (3) institutional processes that may require signed data transfer agreements for limited, predefined data use. However, data sharing is possible for new and participating ENIGMA Stroke Recovery Working Group members who agree to the consortium's ethical standards for data use and upon the submission of a secondary analysis plan for group review. Upon the approval of the proposed analysis plan, access to relevant data is provided contingent on local principal investigator approval, data availability, and compliance with supervening regulatory boards. Deidentified summary data used for this study can be made available upon reasonable request to the corresponding author.
- Language
- English
- Electronic publication date
- 07/31/2024
- Date published
- 07/02/2024
- Academic Unit
- Roy J. Carver Department of Biomedical Engineering; Neurology; Psychiatry; Stead Family Department of Pediatrics; Psychological and Brain Sciences; Iowa Neuroscience Institute; Neurology (Pediatrics); Neurosurgery
- Record Identifier
- 9984696658802771
Metrics
2 Record Views