{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,12]],"date-time":"2026-10-12T22:33:22Z","timestamp":1791844402565,"version":"4.3.4"},"reference-count":37,"publisher":"Public Library of Science (PLoS)","issue":"5","license":[{"start":{"date-parts":[[2021,5,26]],"date-time":"2021-05-26T00:00:00Z","timestamp":1621987200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/publicdomain\/zero\/1.0\/"}],"funder":[{"DOI":"10.13039\/100000001","name":"National Science Foundation","doi-asserted-by":"publisher","award":["1545287"],"award-info":[{"award-number":["1545287"]}],"id":[{"id":"10.13039\/100000001","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000070","name":"National Institute of Biomedical Imaging and Bioengineering","doi-asserted-by":"publisher","award":["1R01EB023808"],"award-info":[{"award-number":["1R01EB023808"]}],"id":[{"id":"10.13039\/100000070","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100010194","name":"Natick Soldier Research, Development and Engineering Center","doi-asserted-by":"publisher","award":["W911QY18C0140"],"award-info":[{"award-number":["W911QY18C0140"]}],"id":[{"id":"10.13039\/100010194","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000049","name":"National Institute on Aging","doi-asserted-by":"publisher","award":["R0106052017"],"award-info":[{"award-number":["R0106052017"]}],"id":[{"id":"10.13039\/100000049","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["www.plosone.org"],"crossmark-restriction":false},"short-container-title":["PLoS ONE"],"abstract":"\n Background and objective<\/jats:title>\n \n Dynamic muscle fascicle length measurements through B-mode ultrasound have become popular for the non-invasive physiological insights they provide regarding musculoskeletal structure-function. However, current practices typically require time consuming post-processing to track muscle length changes from B-mode images. A real-time measurement tool would not only save processing time but would also help pave the way toward closed-loop applications based on feedback signals driven by\n in vivo<\/jats:italic>\n muscle length change patterns. In this paper, we benchmark an approach that combines traditional machine learning (ML) models with B-mode ultrasound recordings to obtain muscle fascicle length changes in real-time. To gauge the utility of this framework for \u2018in-the-loop\u2019 applications, we evaluate accuracy of the extracted muscle length change signals against time-series\u2019 derived from a standard, post-hoc automated tracking algorithm.\n <\/jats:p>\n <\/jats:sec>\n \n Methods<\/jats:title>\n \n We collected B-mode ultrasound data from the soleus muscle of six participants performing five defined ankle motion tasks: (a) seated, constrained ankle plantarflexion, (b) seated, free ankle dorsi\/plantarflexion, (c) weight-bearing, calf raises (d) walking, and then a (e) mix. We trained machine learning (ML) models by pairing muscle fascicle lengths obtained from standardized automated tracking software (UltraTrack) with the respective B-mode ultrasound image input to the tracker, frame-by-frame. Then we conducted hyperparameter optimizations for five different ML models using a grid search to find the best performing parameters for a combination of high correlation and low RMSE between ML and UltraTrack processed muscle fascicle length trajectories. Finally, using the global best model\/hyperparameter settings, we comprehensively evaluated training-testing outcomes within subject (\n i<\/jats:italic>\n .\n e<\/jats:italic>\n ., train and test on same subject), cross subject (\n i<\/jats:italic>\n .\n e<\/jats:italic>\n ., train on one subject, test on another) and within\/direct cross task (\n i<\/jats:italic>\n .\n e<\/jats:italic>\n ., train and test on same subject, but different task).\n <\/jats:p>\n <\/jats:sec>\n \n Results<\/jats:title>\n \n Support vector machine (SVM) was the best performing model with an average\n r<\/jats:italic>\n = 0.70 \u00b10.34 and average RMSE = 2.86 \u00b12.55 mm across all direct training conditions and average\n r<\/jats:italic>\n = 0.65 \u00b10.35 and average RMSE = 3.28 \u00b12.64 mm when optimized for all cross-participant conditions. Comparisons between ML vs. UltraTrack (\n i<\/jats:italic>\n .\n e<\/jats:italic>\n ., ground truth) tracked muscle fascicle length versus time data indicated that ML tracked images reliably capture the salient qualitative features in ground truth length change data, even when correlation values are on the lower end. Furthermore, in the direct training, calf raises condition, which is most comparable to previous studies validating automated tracking performance during isolated contractions on a dynamometer, our ML approach yielded 0.90 average correlation, in line with other accepted tracking methods in the field.\n <\/jats:p>\n <\/jats:sec>\n \n Conclusions<\/jats:title>\n By combining B-mode ultrasound and classical ML models, we demonstrate it is possible to achieve real-time tracking of human soleus muscle fascicles across a number of functionally relevant contractile conditions. This novel sensing modality paves the way for muscle physiology in-the-loop applications that could be used to modify gait via biofeedback or unlock novel wearable device control techniques that could enable restored or augmented locomotion performance.<\/jats:p>\n <\/jats:sec>","DOI":"10.1371\/journal.pone.0246611","type":"journal-article","created":{"date-parts":[[2021,5,26]],"date-time":"2021-05-26T13:42:44Z","timestamp":1622036564000},"page":"e0246611","update-policy":"https:\/\/doi.org\/10.1371\/journal.pone.corrections_policy","source":"Crossref","is-referenced-by-count":34,"title":["Machine learning to extract muscle fascicle length changes from dynamic ultrasound images in real-time"],"prefix":"10.1371","volume":"16","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-9879-2223","authenticated-orcid":true,"given":"Luis G.","family":"Rosa","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jonathan S.","family":"Zia","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Omer T.","family":"Inan","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Gregory S.","family":"Sawicki","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"340","published-online":{"date-parts":[[2021,5,26]]},"reference":[{"key":"pone.0246611.ref001","article-title":"Machine Learning in Ultrasound Computer-Aided Diagnostic Systems: A Survey","author":"Q Huang","year":"2018","journal-title":"BioMed Research International"},{"key":"pone.0246611.ref002","first-page":"933","volume-title":"Medical Image Analysis","author":"S Wang","year":"2012"},{"key":"pone.0246611.ref003","first-page":"261","volume-title":"Engineering","author":"S Liu","year":"2019"},{"key":"pone.0246611.ref004","doi-asserted-by":"crossref","first-page":"786","DOI":"10.1007\/s00261-018-1517-0","article-title":"Machine learning for medical ultrasound: status, methods, and future opportunities","volume":"43","author":"LJ Brattain","year":"2018","journal-title":"Abdom Radiol"},{"key":"pone.0246611.ref005","doi-asserted-by":"crossref","first-page":"1318","DOI":"10.1016\/j.jacr.2019.06.004","article-title":"A Survey of Deep-Learning Applications in Ultrasound: Artificial Intelligence\u2013Powered Ultrasound for Improving Clinical Workflow","volume":"16","author":"Z Akkus","year":"2019","journal-title":"J Am Coll Radiol"},{"key":"pone.0246611.ref006","doi-asserted-by":"crossref","first-page":"978","DOI":"10.1152\/japplphysiol.00835.2019","article-title":"Ultrasound imaging to assess skeletal muscle architecture during movements: a systematic review of methods, reliability, and challenges","volume":"128","author":"B Van Hooren","year":"2020","journal-title":"J Appl Physiol"},{"key":"pone.0246611.ref007","doi-asserted-by":"crossref","first-page":"20192560","DOI":"10.1098\/rspb.2019.2560","article-title":"The force\u2013length\u2013velocity potential of the human soleus muscle is related to the energetic cost of running","volume":"286","author":"S Bohm","year":"2019","journal-title":"Proc R Soc B Biol Sci"},{"key":"pone.0246611.ref008","doi-asserted-by":"crossref","first-page":"579","DOI":"10.1152\/japplphysiol.00253.2013","article-title":"Elastic ankle exoskeletons reduce soleus muscle force but not work in human hopping","volume":"115","author":"DJ Farris","year":"2013","journal-title":"J Appl Physiol"},{"key":"pone.0246611.ref009","doi-asserted-by":"crossref","first-page":"977","DOI":"10.1073\/pnas.1107972109","article-title":"Human medial gastrocnemius force-velocity behavior shifts with locomotion speed and gait","volume":"109","author":"DJ Farris","year":"2012","journal-title":"Proc Natl Acad Sci U S A"},{"key":"pone.0246611.ref010","doi-asserted-by":"crossref","first-page":"1862","DOI":"10.1152\/japplphysiol.00802.2012","article-title":"Linking the mechanics and energetics of hopping with elastic ankle exoskeletons","volume":"113","author":"DJ Farris","year":"2012","journal-title":"J Appl Physiol"},{"key":"pone.0246611.ref011","article-title":"Limited fascicle shortening and fascicle rotation may be associated with impaired voluntary force-generating capacity in pennate muscles of chronic stroke survivors","volume":"75","author":"J Son","year":"2020","journal-title":"Clin Biomech"},{"key":"pone.0246611.ref012","doi-asserted-by":"crossref","first-page":"558","DOI":"10.1016\/j.medengphy.2008.11.004","article-title":"Automatic detection method of muscle fiber movement as revealed by ultrasound images","volume":"31","author":"T Miyoshi","year":"2009","journal-title":"Med Eng Phys"},{"key":"pone.0246611.ref013","first-page":"761","article-title":"Reliability and validity of ultrasound measurements of muscle fascicle length and pennation in humans: A systematic review","author":"LK Kwah","year":"2013","journal-title":"Journal of Applied Physiology. American Physiological Society Bethesda, MD"},{"key":"pone.0246611.ref014","doi-asserted-by":"crossref","first-page":"e7120","DOI":"10.7717\/peerj.7120","article-title":"An automatic fascicle tracking algorithm quantifying gastrocnemius architecture during maximal effort contractions","volume":"2019","author":"JF Drazan","year":"2019","journal-title":"PeerJ"},{"key":"pone.0246611.ref015","doi-asserted-by":"crossref","first-page":"105583","DOI":"10.1016\/j.cmpb.2020.105583","article-title":"Using deep learning to generate synthetic B-mode musculoskeletal ultrasound images","volume":"196","author":"NJ Cronin","year":"2020","journal-title":"Comput Methods Programs Biomed"},{"key":"pone.0246611.ref016","doi-asserted-by":"crossref","first-page":"29","DOI":"10.3390\/jimaging4020029","article-title":"Estimating Full Regional Skeletal Muscle Fibre Orientation from B-Mode Ultrasound Images Using Convolutional, Residual, and Deconvolutional Neural Networks","volume":"4","author":"R Cunningham","year":"2018","journal-title":"J Imaging"},{"key":"pone.0246611.ref017","doi-asserted-by":"crossref","unstructured":"Cunningham RJ, Harding PJ, Loram ID. The application of deep convolutional neural networks to ultrasound for modelling of dynamic states within human skeletal muscle. 2017 [cited 17 Jul 2020]. http:\/\/arxiv.org\/abs\/1706.09450.","DOI":"10.1101\/157479"},{"key":"pone.0246611.ref018","unstructured":"Cronin NJ, Finni T, Seynnes O. Fully Automated Analysis of Muscle Architecture from B-Mode Ultrasound Images with Deep Learning A PREPRINT. 2020. https:\/\/github.com\/njcronin\/DL_Track."},{"key":"pone.0246611.ref019","doi-asserted-by":"crossref","first-page":"82","DOI":"10.1016\/j.jneumeth.2007.09.022","article-title":"Machine learning for real-time single-trial EEG-analysis: From brain-computer interfacing to mental state monitoring","volume":"167","author":"KR M\u00fcller","year":"2008","journal-title":"J Neurosci Methods"},{"key":"pone.0246611.ref020","first-page":"430","volume-title":"Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)","author":"E Rosten","year":"2006"},{"key":"pone.0246611.ref021","doi-asserted-by":"crossref","first-page":"111","DOI":"10.1016\/j.cmpb.2016.02.016","article-title":"UltraTrack: Software for semi-automated tracking of muscle fascicles in sequences of B-mode ultrasound images","volume":"128","author":"DJ Farris","year":"2016","journal-title":"Comput Methods Programs Biomed"},{"key":"pone.0246611.ref022","doi-asserted-by":"crossref","first-page":"1491","DOI":"10.1152\/japplphysiol.00530.2011","article-title":"Automatic tracking of medial gastrocnemius fascicle length during human locomotion","volume":"111","author":"NJ Cronin","year":"2011","journal-title":"J Appl Physiol"},{"key":"pone.0246611.ref023","doi-asserted-by":"crossref","first-page":"678","DOI":"10.1080\/10255842.2011.633516","article-title":"Reliability and accuracy of an automated tracking algorithm to measure controlled passive and active muscle fascicle length changes from ultrasound","volume":"16","author":"JG Gillett","year":"2013","journal-title":"Comput Methods Biomech Biomed Engin"},{"key":"pone.0246611.ref024","article-title":"Exoskeletons improve locomotion economy by reducing active muscle volume","volume":"47","author":"ON Beck","year":"2019","journal-title":"Exerc Sport Sci Rev"},{"key":"pone.0246611.ref025","doi-asserted-by":"crossref","first-page":"200","DOI":"10.1016\/j.jbiomech.2018.03.013","article-title":"Ultrasound estimates of Achilles tendon exhibit unexpected shortening during ankle plantarflexion","volume":"72","author":"ES Matijevich","year":"2018","journal-title":"J Biomech"},{"key":"pone.0246611.ref026","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1038\/s41598-020-60360-4","article-title":"Ultrasound imaging links soleus muscle neuromechanics and energetics during human walking with elastic ankle exoskeletons","volume":"10","author":"RW Nuckols","year":"2020","journal-title":"Sci Rep"},{"key":"pone.0246611.ref027","doi-asserted-by":"crossref","first-page":"e453","DOI":"10.7717\/peerj.453","article-title":"Scikit-image: Image processing in python","volume":"2014","author":"S Van Der Walt","year":"2014","journal-title":"PeerJ"},{"key":"pone.0246611.ref028","article-title":"Scikit-learn: Machine Learning in Python","author":"F Pedregosa","year":"2011","journal-title":"J Mach Learn Res"},{"key":"pone.0246611.ref029","article-title":"Analysis of the Mean Absolute Error (MAE) and the Root Mean Square Error (RMSE) in Assessing Rounding Model","volume":"324","author":"M-S Shang","year":"2018","journal-title":"IOP Conf Ser Mater Sci Eng"},{"key":"pone.0246611.ref030","doi-asserted-by":"crossref","first-page":"35","DOI":"10.1177\/875647939000600106","article-title":"Interpretation of the Correlation Coefficient: A Basic Review","volume":"6","author":"R Taylor","year":"1990","journal-title":"J Diagnostic Med Sonogr"},{"key":"pone.0246611.ref031","unstructured":"Physical Therapy Research: Principles and Applications, ed 2\u2014ProQuest. [cited 17 Jul 2020]. https:\/\/search.proquest.com\/docview\/223117471?pq-origsite=gscholar&fromopenview=true."},{"key":"pone.0246611.ref032","author":"GS Sawicki","journal-title":"The exoskeleton expansion: improving walking and running economy"},{"key":"pone.0246611.ref033","doi-asserted-by":"crossref","first-page":"520","DOI":"10.1152\/jappl.1993.74.2.520","article-title":"Muscle damage is not a function of muscle force but active muscle strain","volume":"74","author":"RL Lieber","year":"1993","journal-title":"J Appl Physiol"},{"key":"pone.0246611.ref034","doi-asserted-by":"crossref","first-page":"521","DOI":"10.1249\/00005768-199205000-00005","article-title":"Structural and mechanical basis of exercise-induced muscle injury","volume":"24","author":"J Frid\u00e9n","year":"1992","journal-title":"Med Sci Sports Exerc"},{"key":"pone.0246611.ref035","doi-asserted-by":"crossref","first-page":"441","DOI":"10.1080\/03091902.2020.1822940","article-title":"Role of machine learning in gait analysis: a review","volume":"44","author":"P Khera","year":"2020","journal-title":"J Med Eng Technol"},{"key":"pone.0246611.ref036","doi-asserted-by":"crossref","first-page":"69","DOI":"10.3389\/fspor.2020.00069","article-title":"Shear Wave Tensiometry Reveals an Age-Related Deficit in Triceps Surae Work at Slow and Fast Walking Speeds","volume":"2","author":"A Ebrahimi","year":"2020","journal-title":"Front Sport Act Living"},{"key":"pone.0246611.ref037","doi-asserted-by":"crossref","first-page":"e0241339","DOI":"10.1371\/journal.pone.0241339","article-title":"Mechanisms of gait phase entrainment in healthy subjects during rhythmic electrical stimulation of the medial gastrocnemius","volume":"15","author":"JE Thorp","year":"2020","journal-title":"PLoS One"}],"container-title":["PLOS ONE"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pone.0246611","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2021,5,26]],"date-time":"2021-05-26T13:43:27Z","timestamp":1622036607000},"score":1,"resource":{"primary":{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pone.0246611"}},"subtitle":[],"editor":[{"given":"Yih-Kuen","family":"Jan","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"editor"}]}],"short-title":[],"issued":{"date-parts":[[2021,5,26]]},"references-count":37,"journal-issue":{"issue":"5","published-online":{"date-parts":[[2021,5,26]]}},"URL":"https:\/\/doi.org\/10.1371\/journal.pone.0246611","relation":{"has-preprint":[{"id-type":"doi","id":"10.1101\/2021.01.25.428061","asserted-by":"object"}]},"ISSN":["1932-6203"],"issn-type":[{"value":"1932-6203","type":"electronic"}],"subject":[],"published":{"date-parts":[[2021,5,26]]}}}