{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,11,3]],"date-time":"2026-11-03T06:00:14Z","timestamp":1793685614867,"version":"4.5.1"},"reference-count":26,"publisher":"Springer Science and Business Media LLC","issue":"1","license":[{"start":{"date-parts":[[2025,12,11]],"date-time":"2025-12-11T00:00:00Z","timestamp":1765411200000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2026,1,15]],"date-time":"2026-01-15T00:00:00Z","timestamp":1768435200000},"content-version":"vor","delay-in-days":35,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["Sci Rep"],"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>\n                    Accurately forecasting the progression of Parkinson\u2019s disease (PD) motor symptoms in early-to-moderate stages is essential for timely intervention and personalized patient care but remains challenging due to heterogeneous and longitudinal symptom evolution. We present a novel dynamic context-aware multi-modal deep learning framework that predicts future motor symptom severity by integrating advanced voice biomarkers with signal processing techniques, clinical progression features, demographic metadata, and semantically enriched patient summary embeddings derived from comprehensive clinical narratives via state-of-the-art natural language processing. Leveraging bidirectional LSTMs augmented with multi-head self-attention, our architecture captures complex temporal dependencies while preventing information leakage. To ensure robust evaluation despite limited sample size (42 patients), we implemented repeated 5-fold cross-validation at the patient level (8 repetitions, 40 total folds), substantially exceeding standard evaluation rigor. Our approach achieves exceptional performance (\n                    <jats:inline-formula>\n                      <jats:tex-math>$$\\hbox {R}^2$$<\/jats:tex-math>\n                    <\/jats:inline-formula>\n                    = 0.9925 \u00b1 0.0027, RMSE = 0.67 \u00b1 0.19, MAE = 0.50 \u00b1 0.15) with all 40 folds achieving\n                    <jats:inline-formula>\n                      <jats:tex-math>$$\\hbox {R}^2$$<\/jats:tex-math>\n                    <\/jats:inline-formula>\n                    &gt; 0.989, significantly outperforming classical machine learning baselines (\n                    <jats:inline-formula>\n                      <jats:tex-math>$$p &lt; 1 \\times 10^{-5}$$<\/jats:tex-math>\n                    <\/jats:inline-formula>\n                    and 0.002785) and all previously published methods on this dataset. Cross-validated ablation studies (240 total model trainings across 6 configurations) reveal that clinical features establish a strong baseline (\n                    <jats:inline-formula>\n                      <jats:tex-math>$$\\hbox {R}^2$$<\/jats:tex-math>\n                    <\/jats:inline-formula>\n                    = 0.9887 \u00b1 0.0043), while text embeddings provide the largest incremental gain (3.82% RMSE reduction). Voice biomarkers contribute modestly to accuracy (2.72%) but substantially enhance stability (10-fold lower variability). The full multi-modal model achieves optimal performance (7.50% RMSE reduction vs. clinical-only) with the lowest variability (CV = 0.27%), demonstrating that dynamic cross-modal fusion enhances both accuracy and robustness. These findings, validated through 40 independent evaluations with each patient tested 8 times, demonstrate that integrating engineered temporal dynamics and contextual embeddings through advanced temporal modeling enables accurate longitudinal predictions of early-to-moderate PD progression. Complete code and implementation details are publicly available to ensure reproducibility.\n                  <\/jats:p>","DOI":"10.1038\/s41598-025-31898-y","type":"journal-article","created":{"date-parts":[[2025,12,11]],"date-time":"2025-12-11T13:24:10Z","timestamp":1765459450000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":7,"title":["Dynamic context-aware multi-modal deep learning for longitudinal prediction of Parkinson\u2019s disease progression"],"prefix":"10.1038","volume":"16","author":[{"given":"Amin","family":"Dehghanghanatkaman","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2025,12,11]]},"reference":[{"key":"31898_CR1","doi-asserted-by":"publisher","first-page":"35","DOI":"10.1038\/s41531-022-00288-w","volume":"8","author":"MB Makarious","year":"2022","unstructured":"Makarious, M. B. et al. Multi-modality machine learning predicting parkinson\u2019s disease. npj Parkinson\u2019s Disease 8, 35 (2022).","journal-title":"npj Parkinson\u2019s Disease"},{"key":"31898_CR2","doi-asserted-by":"publisher","first-page":"4","DOI":"10.3390\/diagnostics15010004","volume":"15","author":"S Benredjem","year":"2024","unstructured":"Benredjem, S. et al. Parkinson\u2019s disease prediction: An attention-based multimodal fusion framework using handwriting and clinical data. Diagnostics 15, 4 (2024).","journal-title":"Diagnostics"},{"key":"31898_CR3","doi-asserted-by":"publisher","first-page":"1761","DOI":"10.3390\/biom13121761","volume":"13","author":"J Zhang","year":"2023","unstructured":"Zhang, J. et al. Prediction of parkinson\u2019s disease using machine learning methods. Biomolecules 13, 1761 (2023).","journal-title":"Biomolecules"},{"key":"31898_CR4","doi-asserted-by":"publisher","first-page":"35","DOI":"10.1038\/s41531-022-00288-w","volume":"8","author":"MB Makarious","year":"2022","unstructured":"Makarious, M. B. et al. Multi-modality machine learning predicting parkinson\u2019s disease. npj Parkinson\u2019s Disease 8, 35 (2022).","journal-title":"npj Parkinson\u2019s Disease"},{"key":"31898_CR5","doi-asserted-by":"publisher","DOI":"10.1016\/j.neurot.2024.e00471","volume":"21","author":"B Chang","year":"2024","unstructured":"Chang, B. et al. Application of multimodal deep learning and multi-instance learning fusion techniques in predicting stn-dbs outcomes for parkinson\u2019s disease patients. Neurotherapeutics 21, e00471 (2024).","journal-title":"Neurotherapeutics"},{"key":"31898_CR6","doi-asserted-by":"crossref","unstructured":"Jiang, Y., Yang, H.-J., Kim, J., Tang, Z. & Ruan, X. Power of multi-modality variables in predicting parkinson\u2019s disease progression. IEEE J. Biomed. Health Inform. 29 1343\u20131356  (2024).","DOI":"10.1109\/JBHI.2024.3482180"},{"key":"31898_CR7","doi-asserted-by":"publisher","first-page":"20941","DOI":"10.1038\/s41598-024-70165-4","volume":"14","author":"V Dentamaro","year":"2024","unstructured":"Dentamaro, V., Impedovo, D., Musti, L., Pirlo, G. & Taurisano, P. Enhancing early parkinson\u2019s disease detection through multimodal deep learning and explainable ai: insights from the ppmi database. Sci. Rep. 14, 20941 (2024).","journal-title":"Sci. Rep."},{"key":"31898_CR8","unstructured":"Vaswani, A. et al. Attention is all you need. Adv. Neural Inf. Process. Syst. 30 5998\u20136008 (2017)."},{"key":"31898_CR9","doi-asserted-by":"publisher","DOI":"10.1016\/j.cmpb.2023.107495","volume":"234","author":"M Junaid","year":"2023","unstructured":"Junaid, M., Ali, S., Eid, F., El-Sappagh, S. & Abuhmed, T. Explainable machine learning models based on multimodal time-series data for the early detection of parkinson\u2019s disease. Comput. Methods Programs Biomed. 234, 107495 (2023).","journal-title":"Comput. Methods Programs Biomed."},{"key":"31898_CR10","doi-asserted-by":"publisher","first-page":"4513","DOI":"10.3390\/app15084513","volume":"15","author":"J Cao","year":"2025","unstructured":"Cao, J. & Long, X. Racf: A multimodal deep learning framework for parkinson\u2019s disease diagnosis using snp and mri data. Appl. Sci. 15, 4513 (2025).","journal-title":"Appl. Sci."},{"key":"31898_CR11","doi-asserted-by":"crossref","unstructured":"Balaha, H.\u00a0M., Hassan, A. E.-S., Ahmed, R.\u00a0A. & Balaha, M.\u00a0H. Comprehensive multimodal approach for parkinson\u2019s disease classification using artificial intelligence: insights and model explainability. Soft Comput. 29 1845\u20131877 (2025).","DOI":"10.1007\/s00500-025-10463-9"},{"key":"31898_CR12","doi-asserted-by":"publisher","first-page":"229","DOI":"10.1038\/s41531-024-00832-w","volume":"10","author":"J Lian","year":"2024","unstructured":"Lian, J. et al. Personalized progression modelling and prediction in parkinson\u2019s disease with a novel multi-modal graph approach. npj Parkinson\u2019s Disease 10, 229 (2024).","journal-title":"npj Parkinson\u2019s Disease"},{"key":"31898_CR13","first-page":"441","volume":"12","author":"S Patil","year":"2024","unstructured":"Patil, S. et al. Parkinson\u2019s disease progression prediction using longitudinal imaging data and grey wolf optimizer-based feature selection. Int. J. Intell. Syst. Appl. Eng. 12, 441\u2013451 (2024).","journal-title":"Int. J. Intell. Syst. Appl. Eng."},{"key":"31898_CR14","doi-asserted-by":"publisher","unstructured":"Tsanas, A. & Little, M. Parkinsons Telemonitoring. UCI Machine Learning Repository https:\/\/doi.org\/10.24432\/C5ZS3N (2009).","DOI":"10.24432\/C5ZS3N"},{"key":"31898_CR15","doi-asserted-by":"publisher","first-page":"229","DOI":"10.1038\/s41531-024-00832-w","volume":"10","author":"J Lian","year":"2024","unstructured":"Lian, J. et al. Personalized progression modelling and prediction in parkinson\u2019s disease with a novel multi-modal graph approach. npj Parkinson\u2019s Disease 10, 229 (2024).","journal-title":"npj Parkinson\u2019s Disease"},{"key":"31898_CR16","doi-asserted-by":"publisher","DOI":"10.1016\/j.compbiomed.2022.106308","volume":"152","author":"L Huang","year":"2023","unstructured":"Huang, L., Ye, X., Yang, M., Pan, L. & hua Zheng, S. Mnc-net: Multi-task graph structure learning based on node clustering for early parkinson\u2019s disease diagnosis. Comput. Biol. Med. 152, 106308 (2023).","journal-title":"Comput. Biol. Med."},{"key":"31898_CR17","doi-asserted-by":"publisher","first-page":"1716","DOI":"10.1002\/mds.29514","volume":"38","author":"U Arshad","year":"2023","unstructured":"Arshad, U., Rahman, F., Hanan, N. & Chen, C. Longitudinal meta-analysis of historical parkinson\u2019s disease trials to inform future trial design. Mov. Disord. 38, 1716\u20131727 (2023).","journal-title":"Mov. Disord."},{"key":"31898_CR18","doi-asserted-by":"publisher","first-page":"2853","DOI":"10.1002\/mds.28730","volume":"36","author":"X Ren","year":"2021","unstructured":"Ren, X., Lin, J., Stebbins, G. T., Goetz, C. G. & Luo, S. Prognostic modeling of parkinson\u2019s disease progression using early longitudinal patterns of change. Mov. Disord. 36, 2853\u20132861 (2021).","journal-title":"Mov. Disord."},{"key":"31898_CR19","doi-asserted-by":"publisher","DOI":"10.1016\/j.cmpb.2023.107495","volume":"234","author":"M Junaid","year":"2023","unstructured":"Junaid, M., Ali, S., Eid, F., El-Sappagh, S. & Abuhmed, T. Explainable machine learning models based on multimodal time-series data for the early detection of parkinson\u2019s disease. Comput. Methods Programs Biomed. 234, 107495 (2023).","journal-title":"Comput. Methods Programs Biomed."},{"key":"31898_CR20","doi-asserted-by":"publisher","first-page":"4","DOI":"10.3390\/diagnostics15010004","volume":"15","author":"S Benredjem","year":"2024","unstructured":"Benredjem, S. et al. Parkinson\u2019s disease prediction: An attention-based multimodal fusion framework using handwriting and clinical data. Diagnostics 15, 4 (2024).","journal-title":"Diagnostics"},{"key":"31898_CR21","doi-asserted-by":"publisher","first-page":"bbaf088","DOI":"10.1093\/bib\/bbaf088","volume":"26","author":"X Zhou","year":"2025","unstructured":"Zhou, X. et al. A novel integrative multimodal classifier to enhance the diagnosis of parkinson\u2019s disease. Brief. Bioinform. 26, bbaf088 (2025).","journal-title":"Brief. Bioinform."},{"key":"31898_CR22","doi-asserted-by":"publisher","first-page":"111","DOI":"10.1038\/s41531-025-00953-w","volume":"11","author":"W-S Lim","year":"2025","unstructured":"Lim, W.-S. et al. Smartphone-derived multidomain features including voice, finger-tapping movement and gait aid early identification of parkinson\u2019s disease. npj Parkinson\u2019s Disease 11, 111 (2025).","journal-title":"npj Parkinson\u2019s Disease"},{"key":"31898_CR23","doi-asserted-by":"publisher","first-page":"20941","DOI":"10.1038\/s41598-024-70165-4","volume":"14","author":"V Dentamaro","year":"2024","unstructured":"Dentamaro, V., Impedovo, D., Musti, L., Pirlo, G. & Taurisano, P. Enhancing early parkinson\u2019s disease detection through multimodal deep learning and explainable ai: insights from the ppmi database. Sci. Rep. 14, 20941 (2024).","journal-title":"Sci. Rep."},{"key":"31898_CR24","doi-asserted-by":"publisher","first-page":"465","DOI":"10.59796\/jcst.V13N2.2023.694","volume":"13","author":"S Pechprasarn","year":"2023","unstructured":"Pechprasarn, S., Manavibool, L., Supmool, N., Vechpanich, N. & Meepadung, P. Predicting parkinson\u2019s disease severity using telemonitoring data and machine learning models: A principal component analysis-based approach for remote healthcare services during covid-19 pandemic. J. Curr. Sci. Technol. 13, 465\u2013485 (2023).","journal-title":"J. Curr. Sci. Technol."},{"key":"31898_CR25","doi-asserted-by":"publisher","first-page":"227","DOI":"10.1007\/s13534-020-00156-7","volume":"10","author":"AH Shahid","year":"2020","unstructured":"Shahid, A. H. & Singh, M. P. A deep learning approach for prediction of parkinson\u2019s disease progression. Biomed. Eng. Lett. 10, 227\u2013239 (2020).","journal-title":"Biomed. Eng. Lett."},{"key":"31898_CR26","doi-asserted-by":"crossref","unstructured":"Li, Y. Predicting updrs scores in parkinson\u2019s disease based on deep learning. In 2023 International Conference on Data Science, Advanced Algorithm and Intelligent Computing (DAI 2023), 705\u2013715 (Atlantis Press, 2023).","DOI":"10.2991\/978-94-6463-370-2_71"}],"container-title":["Scientific Reports"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.nature.com\/articles\/s41598-025-31898-y.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/www.nature.com\/articles\/s41598-025-31898-y","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/www.nature.com\/articles\/s41598-025-31898-y.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,1,15]],"date-time":"2026-01-15T12:40:27Z","timestamp":1768480827000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.nature.com\/articles\/s41598-025-31898-y"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,12,11]]},"references-count":26,"journal-issue":{"issue":"1","published-online":{"date-parts":[[2026,12]]}},"alternative-id":["31898"],"URL":"https:\/\/doi.org\/10.1038\/s41598-025-31898-y","relation":{},"ISSN":["2045-2322"],"issn-type":[{"value":"2045-2322","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,12,11]]},"assertion":[{"value":"12 August 2025","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"5 December 2025","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"11 December 2025","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"order":1,"name":"Ethics","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"The authors declare no competing interests.","order":2,"name":"Ethics","group":{"name":"EthicsHeading","label":"Competing interests"}}],"article-number":"2104"}}