{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,7]],"date-time":"2026-07-07T06:20:57Z","timestamp":1783405257327,"version":"3.54.6"},"reference-count":35,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,7,7]],"date-time":"2026-07-07T00:00:00Z","timestamp":1783382400000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Digit. Health"],"abstract":"<jats:sec>\n                    <jats:title>Introduction<\/jats:title>\n                    <jats:p>Speech and voice changes affect up to 90% of people with Parkinson's disease (PD), a progressive neurodegenerative disorder affecting approximately 10 million people worldwide. Although continuous monitoring of disease severity is clinically important, most existing voice-based computational approaches focus on binary PD-versus-control classification and do not model longitudinal symptom progression. To address this gap, we propose a domain-adaptive transformer model, DAT-PD, for predicting continuous PD severity trajectories from real-world smartphone voice recordings.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>DAT-PD was developed using the public mPower dataset, comprising 58,247 voice recordings from 5,800 participants. The proposed pipeline included noise-aware acoustic preprocessing, extraction of extended Geneva Minimalistic Acoustic Parameter Set (eGeMAPS) features, a domain-adaptive attention mechanism to reduce cross-device and cross-environment variability and a longitudinal trajectory decoder. The model was trained, inferred, and evaluated using continuous MDS-UPDRS Part-II scores as the sole prediction target. Confounder-aware domain adaptation was incorporated to address demographic imbalance, including the age gap between the PD cohort and healthy controls. Robustness was further evaluated under harsh acoustic conditions with signal-to-noise ratios as low as 0 dB.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>On the held-out test set, DAT-PD achieved a mean absolute error (MAE) of 2.74 MDS-UPDRS units (95% CI: 2.44\u20133.01), root mean squared error (RMSE) of 3.61 (95% CI: 3.18\u20134.04), and R\u00b2 of 0.93 (95% CI: 0.91\u20130.95), outperforming six state-of-the-art baseline models. eGeMAPS features substantially outperformed MFCC-only representations, reducing MAE from 5.21 to 2.74. SHAP-based explainability identified MFCC-2, Shimmer (APQ5) and Jitter as the most influential longitudinal voice biomarkers.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Discussion<\/jats:title>\n                    <jats:p>The superior performance of DAT-PD suggests that domain-adaptive longitudinal modeling can effectively capture clinically meaningful voice-based severity trajectories in PD. The advantage of eGeMAPS over MFCC-only features is likely due to its ability to represent phonatory and prosodic characteristics relevant to PD dysarthria, including F0 dynamics, loudness contour, shimmer, jitter and spectral flux. By maintaining robustness under noisy real-world acoustic conditions, DAT-PD supports unsupervised home-based monitoring using standard smartphones. These findings align with the Bridge2AI-Voice research agenda and position DAT-PD as a clinically implementable, non-invasive tool for continuous PD severity assessment.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.3389\/fdgth.2026.1864460","type":"journal-article","created":{"date-parts":[[2026,7,7]],"date-time":"2026-07-07T05:47:04Z","timestamp":1783403224000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Longitudinal voice biomarker trajectory modelling for Parkinson's disease severity: domain-adaptive transfer learning on mPower real-world smartphone data"],"prefix":"10.3389","volume":"8","author":[{"given":"Deepika Roselind","family":"Johnson","sequence":"first","affiliation":[{"name":"School of Computer Science and Engineering, Vellore Institute of Technology, Chennai Campus","place":["Chennai, India"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"G.","family":"Logeswari","sequence":"additional","affiliation":[{"name":"School of Computer Science and Engineering, Vellore Institute of Technology, Chennai Campus","place":["Chennai, India"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,7,7]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"344","DOI":"10.1016\/S1474-4422(24)00038-3","article-title":"Global, regional, and national burden of neurological disorders, 1990-2021: a systematic analysis for the Globall Burden of Disease Study 2021","volume":"23","author":"Steinmetz","year":"2024","journal-title":"Lancet Neurol"},{"key":"B2","doi-asserted-by":"crossref","DOI":"10.1038\/s41531-020-0117-1","article-title":"Current and projected future economic burden of Parkinson\u2019s disease in the U.S.","volume":"6","author":"Yang","year":"2020","journal-title":"NPJ Parkinsons Dis"},{"key":"B3","doi-asserted-by":"publisher","first-page":"131","DOI":"10.1155\/1999\/327643","article-title":"Speech impairment in a large sample of patients with Parkinson\u2019s disease","volume":"11","author":"Ho","year":"1998","journal-title":"Behav Neurol"},{"key":"B4","doi-asserted-by":"publisher","DOI":"10.1038\/sdata.2016.11","article-title":"The mPower study, Parkinson disease mobile data collected using ResearchKit","volume":"3","author":"Bot","year":"2016","journal-title":"Sci Data"},{"key":"B5","doi-asserted-by":"publisher","DOI":"10.1038\/s41598-024-81824-x","article-title":"Transformer-based transfer learning on self-reported voice recordings for Parkinson\u2019s disease diagnosis","volume":"14","author":"Tougui","year":"2024","journal-title":"Sci Rep"},{"key":"B6","author":"Zhan","year":"2016"},{"key":"B7","author":"Gelderen","year":"2024"},{"key":"B8","doi-asserted-by":"publisher","first-page":"1015","DOI":"10.1109\/TBME.2008.2005954","article-title":"Suitability of dysphonia measurements for telemonitoring of Parkinson\u2019s disease","volume":"56","author":"Little","year":"2008","journal-title":"IEEE Trans Biomed Eng"},{"key":"B9","doi-asserted-by":"publisher","first-page":"842","DOI":"10.1098\/rsif.2010.0456","article-title":"Nonlinear speech analysis algorithms mapped to a standard metric achieve clinically useful quantification of average Parkinson\u2019s disease symptom severity","volume":"8","author":"Tsanas","year":"2011","journal-title":"J R Soc Interface"},{"key":"B10","doi-asserted-by":"publisher","first-page":"102418","DOI":"10.1016\/j.bspc.2021.102418","article-title":"Advances in Parkinson\u2019s disease detection and assessment using voice and speech: a review of the articulatory and phonatory aspects","volume":"66","author":"Moro-Vel\u00e1zquez","year":"2021","journal-title":"Biomed Signal Process Control"},{"key":"B11","doi-asserted-by":"publisher","first-page":"103109","DOI":"10.1016\/j.artmed.2025.103109","article-title":"Voice analysis in Parkinson\u2019s disease \u2013 a systematic literature review","volume":"163","author":"Xavier","year":"2025","journal-title":"Artif Intell Med"},{"key":"B12","first-page":"66","article-title":"Predicting Parkinson\u2019s disease progression: a non-invasive method leveraging voice inputs","volume":"8","author":"Hassan","year":"2023","journal-title":"J Comput Sci"},{"key":"B13","doi-asserted-by":"crossref","first-page":"11","DOI":"10.1016\/j.specom.2018.05.007","article-title":"Speaker models for monitoring Parkinson\u2019s disease progression considering different communication channels and acoustic conditions","volume":"101","author":"Arias-Vergara","year":"2018","journal-title":"Speech Commun"},{"key":"B14","doi-asserted-by":"crossref","first-page":"103379","DOI":"10.1016\/j.specom.2026.103379","article-title":"Towards automating the frenchay dysarthria assessment: can neural phoneme posteriorgrams inform the analysis of dysarthric speech?","author":"Leung","year":"2026","journal-title":"Speech Commun"},{"key":"B15","doi-asserted-by":"publisher","first-page":"876","DOI":"10.1001\/jamaneurol.2018.0809","article-title":"Using smartphones and machine learning to quantify Parkinson disease severity: the mobile Parkinson disease score","volume":"75","author":"Zhan","year":"2018","journal-title":"JAMA Neurol"},{"key":"B16","doi-asserted-by":"publisher","DOI":"10.1038\/s41746-021-00514-4","article-title":"A systematic review of smartphone-based human activity recognition methods for health research","volume":"4","author":"Straczkiewicz","year":"2021","journal-title":"NPJ Digit Med"},{"key":"B17","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1109\/ACCESS.2022.3156659","article-title":"Smartphone-based digital biomarkers for Parkinson\u2019s disease in a remotely-administered setting","author":"Go\u00f1i","year":"2022","journal-title":"IEEE Access"},{"key":"B18","author":"Vaswani","year":"2017"},{"key":"B19","author":"Nie","year":"2022"},{"key":"B20","author":"Zhou","year":"2020"},{"key":"B21","first-page":"491","article-title":"Big Transfer (BiT): general Visual Representation Learning","author":"Kolesnikov","year":"2019"},{"key":"B22","author":"Pepino","year":"2021"},{"key":"B23","doi-asserted-by":"publisher","first-page":"766","DOI":"10.1109\/TPAMI.2019.2945942","article-title":"A review of domain adaptation without target labels","volume":"43","author":"Kouw","year":"2021","journal-title":"IEEE Trans Pattern Anal Mach Intell"},{"key":"B24","first-page":"1180","article-title":"Unsupervised domain adaptation by backpropagation","volume":"37","author":"Ganin","year":"2015","journal-title":"Int Conf Machine Learning"},{"key":"B25","first-page":"443","article-title":"Deep CORAL: correlation alignment for deep domain adaptation","volume":"9915","author":"Sun","year":"2016","journal-title":"ECCV Workshops"},{"key":"B26","doi-asserted-by":"publisher","first-page":"446","DOI":"10.1609\/aaai.v29i1.9209","article-title":"A multivariate timeseries modeling approach to severity of illness assessment and forecasting in ICU with sparse, heterogeneous clinical data","volume":"2015","author":"Ghassemi","year":"2015","journal-title":"Proc AAAI Conf Artif Intell"},{"key":"B27","doi-asserted-by":"publisher","first-page":"e555","DOI":"10.1016\/S2589-7500(21)00101-1","article-title":"Discovery of Parkinson\u2019s disease states and disease progression modelling: a longitudinal data study using machine learning","volume":"3","author":"Severson","year":"2021","journal-title":"Lancet Digit Health"},{"key":"B28","doi-asserted-by":"publisher","DOI":"10.3390\/bioengineering11030295","article-title":"Leveraging deep learning for fine-grained categorization of Parkinson\u2019s disease progression levels through analysis of vocal acoustic patterns","volume":"11","author":"Malekroodi","year":"2024","journal-title":"Bioengineering"},{"key":"B29","first-page":"4765","article-title":"A unified approach to interpreting model predictions","volume":"30","author":"Lundberg","year":"2017","journal-title":"Neural Inf Process Syst"},{"key":"B30","first-page":"200","article-title":"Interpreting deep representations of phonetic features via neuro-based concept detector: application to speech disorders due to head and neck cancer. IEEE\/ACM transactions on audio","volume":"31","author":"Abderrazek","year":"2023","journal-title":"Speech Lang Proc"},{"key":"B31","doi-asserted-by":"publisher","first-page":"717","DOI":"10.1109\/JSTSP.2025.3539845","article-title":"Unveiling interpretability in self-supervised speech representations for Parkinson\u2019s diagnosis","volume":"19","author":"Gimeno-G\u00f3mez","year":"2025","journal-title":"IEEE J Sel Top Signal Process"},{"key":"B32","doi-asserted-by":"publisher","first-page":"1672971","DOI":"10.3389\/fnagi.2025.1672971","article-title":"Explainable machine learning for early detection of Parkinson\u2019s disease in aging populations using vocal biomarkers","volume":"17","author":"Egbo","year":"2025","journal-title":"Front Aging Neurosci"},{"key":"B33","doi-asserted-by":"publisher","first-page":"190","DOI":"10.1109\/TAFFC.2015.2457417","article-title":"The Geneva minimalistic acoustic parameter set (GeMAPS) for voice research and affective computing","volume":"7","author":"Eyben","year":"2016","journal-title":"IEEE Trans Affect Comput"},{"key":"B34","year":"2019"},{"key":"B35","doi-asserted-by":"publisher","first-page":"21","DOI":"10.1016\/j.jcomdis.2018.08.002","article-title":"Towards an automatic evaluation of the dysarthria level of patients with Parkinson\u2019s disease","volume":"76","author":"V\u00e1squez-Correa","year":"2018","journal-title":"J Commun Disord"}],"container-title":["Frontiers in Digital Health"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fdgth.2026.1864460\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,7,7]],"date-time":"2026-07-07T05:47:06Z","timestamp":1783403226000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fdgth.2026.1864460\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,7,7]]},"references-count":35,"alternative-id":["10.3389\/fdgth.2026.1864460"],"URL":"https:\/\/doi.org\/10.3389\/fdgth.2026.1864460","relation":{},"ISSN":["2673-253X"],"issn-type":[{"value":"2673-253X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,7,7]]},"article-number":"1864460"}}