{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"institution":[{"name":"medRxiv"}],"indexed":{"date-parts":[[2026,1,16]],"date-time":"2026-01-16T13:19:14Z","timestamp":1768569554832,"version":"3.49.0"},"posted":{"date-parts":[[2025,3,4]]},"group-title":"Neurology","reference-count":41,"publisher":"openRxiv","license":[{"start":{"date-parts":[[2025,3,4]],"date-time":"2025-03-04T00:00:00Z","timestamp":1741046400000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/www.medrxiv.org\/about\/FAQ#license"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"accepted":{"date-parts":[[2025,9,10]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                <jats:sec>\n                  <jats:title>Objective<\/jats:title>\n                  <jats:p>Frequent and objective assessment of ataxia severity is essential for tracking disease progression and evaluating the effectiveness of potential treatments. Wearable-based assessments have emerged as a promising solution. However, existing methods rely on inertial data features directly correlated with subjective and coarse clinician-evaluated rating scales, which serve as imperfect gold standards. This approach may introduce biases and restrict flexibility in feature design. To address these limitations, this study introduces a novel contrastive learning-based model that leverages motor severity differences in wearable inertial data to learn relevant features.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Methods<\/jats:title>\n                  <jats:p>The model was trained on inertial data collected from 87 individuals with diagnostically heterogeneous ataxias and 44 healthy participants performing the finger-to-nose task. A pairwise contrastive loss function was proposed to learn representations capturing relative differences in ataxia severity, which were evaluated through downstream regression and classification tasks.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Results<\/jats:title>\n                  <jats:p>The learned features demonstrated strong cross-sectional (r = 0.84) and longitudinal (r = 0.68) associations with clinical scores and robust measurement reliability (intraclass correlation coefficient = 0.96). Additionally, the model exhibited strong known-group validity, distinguishing between ataxia and healthy phenotypes with an area under the receiver operating characteristic curve of 0.95.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Conclusion<\/jats:title>\n                  <jats:p>The proposed contrastive model captures robust representations of disease severity with reduced reliance on clinical scales, outperforming state-of-the-art methods that derive features directly from clinical scores.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Significance<\/jats:title>\n                  <jats:p>Combining wearable sensors with contrastive learning enables a more objective, scalable, and frequent method for assessing ataxia severity, with the potential to enhance patient monitoring and improve clinical trial efficiency.<\/jats:p>\n                <\/jats:sec>","DOI":"10.1101\/2025.02.28.25323114","type":"posted-content","created":{"date-parts":[[2025,3,5]],"date-time":"2025-03-05T00:40:18Z","timestamp":1741135218000},"source":"Crossref","is-referenced-by-count":0,"title":["Contrastive Learning Model for Wearable-based Ataxia Assessment"],"prefix":"10.64898","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-6877-6329","authenticated-orcid":false,"given":"Juhyeon","family":"Lee","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-1807-4775","authenticated-orcid":false,"given":"Brandon","family":"Oubre","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8530-1827","authenticated-orcid":false,"given":"Jean-Francois","family":"Daneault","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Christopher D.","family":"Stephen","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Jeremy D.","family":"Schmahmann","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Anoopum S.","family":"Gupta","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-5935-125X","authenticated-orcid":false,"given":"Sunghoon Ivan","family":"Lee","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"54368","reference":[{"issue":"4","key":"2025091303450465000_2025.02.28.25323114v2.1","doi-asserted-by":"crossref","first-page":"1208","DOI":"10.1212\/CON.0000000000000362","article-title":"Ataxia","volume":"22","year":"2016","journal-title":"Continuum: Lifelong Learn. Neurology"},{"key":"2025091303450465000_2025.02.28.25323114v2.2","doi-asserted-by":"publisher","DOI":"10.1136\/practneurol-2018-002096"},{"issue":"3","key":"2025091303450465000_2025.02.28.25323114v2.3","first-page":"1","article-title":"The comprehensive management of cerebellar ataxia in adults","volume":"21","year":"2019","journal-title":"Current Treatment Options Neurology"},{"issue":"2","key":"2025091303450465000_2025.02.28.25323114v2.4","doi-asserted-by":"crossref","first-page":"101","DOI":"10.1080\/14737175.2022.2029703","article-title":"Current and emerging treatment modalities for spinocerebellar ataxias","volume":"22","year":"2022","journal-title":"Expert review of neurotherapeutics"},{"issue":"6","key":"2025091303450465000_2025.02.28.25323114v2.5","first-page":"707","article-title":"Oligonucleotide therapeutics in neurodegenerative diseases","volume":"15","year":"2018","journal-title":"RNA biology"},{"key":"2025091303450465000_2025.02.28.25323114v2.6","doi-asserted-by":"crossref","first-page":"488","DOI":"10.1007\/s12311-011-0316-8","article-title":"Ataxia rating scales\u2014psychometric profiles, natural history and their application in clinical trials","volume":"11","year":"2012","journal-title":"The Cerebellum"},{"key":"2025091303450465000_2025.02.28.25323114v2.7","first-page":"1","article-title":"Guidelines on the diagnosis and management of the progressive ataxias","volume":"14","year":"2019","journal-title":"Orphanet journal of rare diseases"},{"key":"2025091303450465000_2025.02.28.25323114v2.8","doi-asserted-by":"publisher","DOI":"10.1002\/mds.22681"},{"key":"2025091303450465000_2025.02.28.25323114v2.9","doi-asserted-by":"publisher","DOI":"10.1212\/01.wnl.0000219042.60538.92"},{"key":"2025091303450465000_2025.02.28.25323114v2.10","unstructured":"H. Zhou et al., \u201cAssessment of gait and balance impairment in people with spinocerebellar ataxia using wearable sensors,\u201d Neurological Sciences, pp. 1\u201311, 2022."},{"key":"2025091303450465000_2025.02.28.25323114v2.11","doi-asserted-by":"crossref","unstructured":"R. B. Wilson , Experimental therapeutics for Friedreich ataxia. Elsevier: Burlington, MA, 2006.","DOI":"10.1016\/B978-012369462-1\/50019-3"},{"key":"2025091303450465000_2025.02.28.25323114v2.12","unstructured":"B. Oubre et al., \u201cDecomposition of reaching movements enables detection and measurement of ataxia,\u201d Cerebellum, pp. 1\u201312, 2021."},{"issue":"7","key":"2025091303450465000_2025.02.28.25323114v2.13","doi-asserted-by":"crossref","first-page":"2314","DOI":"10.1109\/TBME.2022.3142504","article-title":"Analysis of gait sub-movements to estimate ataxia severity using ankle inertial data","volume":"69","year":"2022","journal-title":"IEEE Trans. Biomed. Eng"},{"key":"2025091303450465000_2025.02.28.25323114v2.14","doi-asserted-by":"publisher","DOI":"10.1038\/s41598-020-59121-0"},{"key":"2025091303450465000_2025.02.28.25323114v2.15","doi-asserted-by":"crossref","unstructured":"R. Manohar et al., \u201cAt-home wearables and machine learning capture motor impairment and progression in adult ataxias,\u201d medRxiv, pp. 2024\u201310, 2024.","DOI":"10.1101\/2024.10.27.24316161"},{"issue":"2","key":"2025091303450465000_2025.02.28.25323114v2.16","doi-asserted-by":"crossref","first-page":"fcad064","DOI":"10.1093\/braincomms\/fcad064","article-title":"Real-life ankle submovements and computer mouse use reflect patient-reported function in adult ataxias","volume":"5","year":"2023","journal-title":"Brain Communications"},{"issue":"10","key":"2025091303450465000_2025.02.28.25323114v2.17","doi-asserted-by":"crossref","first-page":"5097","DOI":"10.1007\/s00415-023-11786-z","article-title":"Estimation of ataxia severity in children with ataxiatelangiectasia using ankle-worn sensors","volume":"270","year":"2023","journal-title":"Journal of Neurology"},{"issue":"2","key":"2025091303450465000_2025.02.28.25323114v2.18","first-page":"261","article-title":"Real-life wrist movement patterns capture motor impairment in individuals with ataxia-telangiectasia","volume":"22","year":"2023","journal-title":"The Cerebellum"},{"key":"2025091303450465000_2025.02.28.25323114v2.19","unstructured":"N. C. Khan et al., \u201cFree-living motor activity monitoring in ataxiatelangiectasia,\u201d The Cerebellum, pp. 1\u201312, 2022."},{"key":"2025091303450465000_2025.02.28.25323114v2.20","doi-asserted-by":"crossref","unstructured":"J. Goschenhofer et al., \u201cWearable-based parkinson\u2019s disease severity monitoring using deep learning,\u201d in Machine Learning and Knowledge Discovery in Databases: European Conference, ECML PKDD 2019, W\u00fcrzburg, Germany, September 16\u201320, 2019, Proceedings, Part III. Springer, 2020, pp. 400\u2013415.","DOI":"10.1007\/978-3-030-46133-1_24"},{"key":"2025091303450465000_2025.02.28.25323114v2.21","doi-asserted-by":"crossref","unstructured":"\u00c7. Berke Erda\u015f et al., \u201cCnn-based severity prediction of neurodegenerative diseases using gait data,\u201d Digital Health, vol. 8, p. 20552076221075147, 2022.","DOI":"10.1177\/20552076221075147"},{"key":"2025091303450465000_2025.02.28.25323114v2.22","first-page":"006","article-title":"Technological evolution in the instrumentation of ataxia severity measurement","volume":"11","year":"2023","journal-title":"IEEE Access"},{"key":"2025091303450465000_2025.02.28.25323114v2.23","first-page":"661","volume-title":"in Advances in Neural Information Processing Systems","volume":"33","year":"2020"},{"key":"2025091303450465000_2025.02.28.25323114v2.24","unstructured":"F. Graf et al., \u201cDissecting supervised contrastive learning,\u201d in International Conference on Machine Learning. PMLR, 2021, pp. 3821\u20133830."},{"key":"2025091303450465000_2025.02.28.25323114v2.25","unstructured":"Y. Xue et al., \u201cInvestigating why contrastive learning benefits robustness against label noise,\u201d in International Conference on Machine Learning. PMLR, 2022, pp. 24 851\u201324 871."},{"key":"2025091303450465000_2025.02.28.25323114v2.26","unstructured":"L. Holmstrom , \u201cHow is the orientation of an opal estimated?\u201d April 2021, accessed: 2025-01-31. [Online]. Available: https:\/\/support.apdm.com\/hc\/en-us\/articles\/115000390803"},{"key":"2025091303450465000_2025.02.28.25323114v2.27","doi-asserted-by":"publisher","DOI":"10.1109\/10.554760"},{"issue":"8","key":"2025091303450465000_2025.02.28.25323114v2.28","doi-asserted-by":"crossref","first-page":"8980","DOI":"10.1609\/aaai.v36i8.20881","article-title":"Ts2vec: Towards universal representation of time series","volume":"36","year":"2022","journal-title":"in Proceedings of the AAAI Conference on Artificial Intelligence"},{"key":"2025091303450465000_2025.02.28.25323114v2.29","unstructured":"Y. Yang et al., \u201cSimper: Simple self-supervised learning of periodic targets,\u201d in The Eleventh International Conference on Learning Representations, 2023."},{"issue":"2","key":"2025091303450465000_2025.02.28.25323114v2.30","doi-asserted-by":"crossref","first-page":"354","DOI":"10.1002\/mds.27915","article-title":"Computer mouse use captures ataxia and parkinsonism, enabling accurate measurement and detection","volume":"35","year":"2020","journal-title":"Movement Disorders"},{"key":"2025091303450465000_2025.02.28.25323114v2.31","unstructured":"I. Loshchilov and F. Hutter , \u201cDecoupled weight decay regularization,\u201d in International Conference on Learning Representations, 2019."},{"key":"2025091303450465000_2025.02.28.25323114v2.32","doi-asserted-by":"publisher","DOI":"10.1016\/J.JCM.2016.02.012"},{"key":"2025091303450465000_2025.02.28.25323114v2.33","doi-asserted-by":"crossref","unstructured":"D. W. Hosmer Jr et al., Applied logistic regression. John Wiley & Sons, 2013.","DOI":"10.1002\/9781118548387"},{"issue":"5","key":"2025091303450465000_2025.02.28.25323114v2.34","doi-asserted-by":"crossref","first-page":"1147","DOI":"10.1002\/acr.24821","article-title":"Reliability, validity, responsiveness, and minimum important change of the stair climb test in adults with hip and knee osteoarthritis","volume":"75","year":"2023","journal-title":"Arthritis care & research"},{"key":"2025091303450465000_2025.02.28.25323114v2.35","doi-asserted-by":"crossref","unstructured":"J. Cohen , Statistical power analysis for the behavioral sciences. routledge, 2013.","DOI":"10.4324\/9780203771587"},{"key":"2025091303450465000_2025.02.28.25323114v2.36","first-page":"1","article-title":"A comprehensive scheme for the objective upper body assessments of subjects with cerebellar ataxia","volume":"17","year":"2020","journal-title":"Journal of NeuroEngineering and Rehabilitation"},{"issue":"1","key":"2025091303450465000_2025.02.28.25323114v2.37","doi-asserted-by":"crossref","first-page":"5080","DOI":"10.1038\/s41467-023-40917-3","article-title":"At-home wearables and machine learning sensitively capture disease progression in amyotrophic lateral sclerosis","volume":"14","year":"2023","journal-title":"Nature Communications"},{"issue":"9","key":"2025091303450465000_2025.02.28.25323114v2.38","doi-asserted-by":"crossref","first-page":"4221","DOI":"10.3390\/s23094221","article-title":"Self-supervised contrastive learning for medical time series: A systematic review","volume":"23","year":"2023","journal-title":"Sensors"},{"key":"2025091303450465000_2025.02.28.25323114v2.39","doi-asserted-by":"crossref","unstructured":"M. Shuqair et al., \u201cShared-task self-supervised learning for estimating free movement unified parkinson\u2019s disease rating scale iii,\u201d in 2024 46th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC). IEEE, 2024, pp. 1\u20134.","DOI":"10.1109\/EMBC53108.2024.10782431"},{"issue":"3","key":"2025091303450465000_2025.02.28.25323114v2.40","doi-asserted-by":"crossref","first-page":"250","DOI":"10.1038\/s42256-023-00620-w","article-title":"Evaluation of post-hoc interpretability methods in timeseries classification","volume":"5","year":"2023","journal-title":"Nature Machine Intelligence"},{"key":"2025091303450465000_2025.02.28.25323114v2.41","unstructured":"A. Vaswani et al., \u201cAttention is all you need,\u201d Advances in neural information processing systems, vol. 30, 2017."}],"container-title":[],"original-title":[],"link":[{"URL":"https:\/\/syndication.highwire.org\/content\/doi\/10.1101\/2025.02.28.25323114","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,1,15]],"date-time":"2026-01-15T20:29:39Z","timestamp":1768508979000},"score":1,"resource":{"primary":{"URL":"http:\/\/medrxiv.org\/lookup\/doi\/10.1101\/2025.02.28.25323114"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,3,4]]},"references-count":41,"URL":"https:\/\/doi.org\/10.1101\/2025.02.28.25323114","relation":{"is-preprint-of":[{"id-type":"doi","id":"10.1109\/TBME.2025.3608674","asserted-by":"subject"}]},"subject":[],"published":{"date-parts":[[2025,3,4]]},"subtype":"preprint"}}