Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

  • Article
  • Published:

Neural network-assisted personalized handwriting analysis for Parkinson’s disease diagnostics

Nature Chemical Engineering volume 2, pages 358–368 (2025)Cite this article

Subjects

Abstract

Diagnosing Parkinson’s disease (PD) promptly, accessibly and effectively is crucial for improving patient outcomes, yet reaching this goal remains a challenge. Here we developed a diagnostic pen featuring a soft magnetoelastic tip and ferrofluid ink, capable of sensitively and quantitatively converting both on-surface and in-air writing motions into high-fidelity, analyzable signals for self-powered PD diagnostics. The diagnostic pen’s working mechanism is based on the magnetoelastic effect in its magnetoelastic tip and the dynamic movement of the ferrofluid ink. To validate the clinical potential, a pilot human study was conducted, incorporating both patients with PD and healthy participants. The diagnostic pen accurately recorded handwriting signals, and a one-dimensional convolutional neural network-assisted analysis successfully distinguished patients with PD with an average accuracy of 96.22%. Our development of the diagnostic pen represents a low-cost, widely disseminable and reliable technology with the potential to improve PD diagnostics across large populations and resource-limited areas.

Access through your institution
Buy or subscribe

This is a preview of subscription content, access via your institution

Access options

Access through your institution

Buy this article

  • Purchase on SpringerLink
  • Instant access to full article PDF
Buy now

Prices may be subject to local taxes which are calculated during checkout

Fig. 1: Design and working mechanism of the magnetoelastic diagnostic pen.
Fig. 2: Fabrication and characterization.
Fig. 3: Converting handwriting into high-fidelity sensing signals.
Fig. 4: Neural network-assisted personalized handwriting analysis for PD diagnostics with pilot human studies.

Similar content being viewed by others

Data availability

Data supporting the results in this study are available within the Article and its Supplementary Information. Human study data are not publicly available because they contain information that could compromise research participant privacy. Source data are provided with this paper.

Code availability

The machine leaning code in this study is available via GitHub at https://github.com/JCLABShare/PD-PEN.

References

  1. Poewe, W. et al. Parkinson disease. Nat. Rev. Dis. Primers 3, 17013 (2017).

    Article  PubMed  Google Scholar 

  2. Bloem, B. R., Okun, M. S. & Klein, C. Parkinson’s disease. Lancet 397, 2284–2303 (2021).

    Article  CAS  PubMed  Google Scholar 

  3. Weintraub, D. et al. The neuropsychiatry of Parkinson’s disease: advances and challenges. Lancet Neurol. 21, 89–102 (2022).

    Article  PubMed  PubMed Central  Google Scholar 

  4. Kowal, S. L., Dall, T. M., Chakrabarti, R., Storm, M. V. & Jain, A. The current and projected economic burden of Parkinson’s disease in the United States. Mov. Disord. 28, 311–318 (2013).

    Article  PubMed  Google Scholar 

  5. Armstrong, M. J. & Okun, M. S. Diagnosis and treatment of Parkinson disease: a review. JAMA 323, 548–560 (2020).

    Article  PubMed  Google Scholar 

  6. Dorsey, E. R. et al. Global, regional, and national burden of Parkinson’s disease, 1990–2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet Neurol. 17, 939–953 (2018).

    Article  Google Scholar 

  7. Aarsland, D. et al. Parkinson disease-associated cognitive impairment. Nat. Rev. Dis. Primers 7, 47 (2021).

    Article  PubMed  Google Scholar 

  8. Milekovic, T. et al. A spinal cord neuroprosthesis for locomotor deficits due to Parkinson’s disease. Nat. Med. 29, 2854–2865 (2023).

    Article  CAS  PubMed  Google Scholar 

  9. Pavy-Le Traon, A. et al. Clinical rating scales for urinary symptoms in Parkinson disease: critique and recommendations. Mov. Disord. Clin. Pract. 5, 479–491 (2018).

    Article  PubMed  PubMed Central  Google Scholar 

  10. Fox, S. H. et al. International Parkinson and movement disorder society evidence-based medicine review: update on treatments for the motor symptoms of Parkinson’s disease. Mov. Disord. 33, 1248–1266 (2018).

    Article  CAS  PubMed  Google Scholar 

  11. Moustafa, A. A. et al. Motor symptoms in Parkinson’s disease: a unified framework. Neurosci. Biobehav. Rev. 68, 727–740 (2016).

    Article  PubMed  Google Scholar 

  12. Tolosa, E., Garrido, A., Scholz, S. W. & Poewe, W. Challenges in the diagnosis of Parkinson’s disease. Lancet Neurol. 20, 385–397 (2021).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  13. Musubire, A. K. An international exchange observership at Yale University: a Ugandan physician experience. Neurology 92, 582–584 (2019).

    Article  PubMed  PubMed Central  Google Scholar 

  14. Höglinger, G. U. et al. A biological classification of Parkinson’s disease: the SynNeurGe research diagnostic criteria. Lancet Neurol. 23, 191–204 (2024).

    Article  PubMed  Google Scholar 

  15. Sulzer, D. & Edwards, R. H. The physiological role of α-synuclein and its relationship to Parkinson’s disease. J. Neurochem. 150, 475–486 (2019).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  16. Pereira, J. B. et al. DOPA decarboxylase is an emerging biomarker for Parkinsonian disorders including preclinical Lewy body disease. Nat. Aging 3, 1201–1209 (2023).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  17. Wagner, S. K. et al. Retinal optical coherence tomography features associated with incident and prevalent Parkinson disease. Neurology 101, e1581–e1593 (2023).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  18. Lotankar, S., Prabhavalkar, K. S. & Bhatt, L. K. Biomarkers for Parkinson’s disease: recent advancement. Neurosci. Bull. 33, 585–597 (2017).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  19. Schiess, N. et al. Six action steps to address global disparities in Parkinson disease: a World Health Organization priority. JAMA Neurol. 79, 929–936 (2022).

    Article  PubMed  Google Scholar 

  20. Impedovo, D. & Pirlo, G. Dynamic handwriting analysis for the assessment of neurodegenerative diseases: a pattern recognition perspective. IEEE Rev. Biomed. Eng. 12, 209–220 (2019).

    Article  PubMed  Google Scholar 

  21. Broeder, S. et al. The effects of dual tasking on handwriting in patients with Parkinson’s disease. Neurosci. 263, 193–202 (2014).

    Article  CAS  Google Scholar 

  22. Barrett, M. J., Wylie, S. A., Harrison, M. B. & Wooten, G. F. Handedness and motor symptom asymmetry in Parkinson’s disease. J. Neurol. Neurosurg. Psychiatry 82, 1122–1124 (2011).

    Article  PubMed  Google Scholar 

  23. van der Hoorn, A., Burger, H., Leenders, K. L. & de Jong, B. M. Handedness correlates with the dominant Parkinson side: a systematic review and meta-analysis. Mov. Disord. 27, 206–210 (2012).

    Article  PubMed  Google Scholar 

  24. van der Hoorn, A., Bartels, A. L., Leenders, K. L. & de Jong, B. M. Handedness and dominant side of symptoms in Parkinson’s disease. Parkinsonism Relat. Disord. 17, 58–60 (2011).

    Article  PubMed  Google Scholar 

  25. Rosenblum, S., Samuel, M., Zlotnik, S., Erikh, I. & Schlesinger, I. Handwriting as an objective tool for Parkinson’s disease diagnosis. J. Neurol. 260, 2357–2361 (2013).

    Article  PubMed  Google Scholar 

  26. Thomas, M., Lenka, A. & Kumar Pal, P. Handwriting analysis in Parkinson’s disease: current status and future directions. Mov. Disord. Clin. Pract. 4, 806–818 (2017).

    Article  PubMed  PubMed Central  Google Scholar 

  27. Masud, M. K. et al. Superparamagnetic nanoarchitectures for disease-specific biomarker detection. Chem. Soc. Rev. 48, 5717–5751 (2019).

    Article  CAS  PubMed  Google Scholar 

  28. Lu, W., Shen, Y., Xie, A. & Zhang, W. Green synthesis and characterization of superparamagnetic Fe3O4 nanoparticles. J. Magn. Magn. Mater. 322, 1828–1833 (2010).

    Article  CAS  Google Scholar 

  29. Jeong, U., Teng, X., Wang, Y., Yang, H. & Xia, Y. Superparamagnetic colloids: controlled synthesis and niche applications. Adv. Mater. 19, 33–60 (2007).

    Article  CAS  Google Scholar 

  30. Cowley, M. D. & Rosensweig, R. E. The interfacial stability of a ferromagnetic fluid. J. Fluid Mech. 30, 671–688 (1967).

    Article  CAS  Google Scholar 

  31. Neuringer, J. L. & Rosensweig, R. E. Ferrohydrodynamics. Phys. Fluids 7, 1927–1937 (1964).

    Article  Google Scholar 

  32. Rosensweig, R. E. Ferrohydrodynamics (Cambridge Univ. Press, 1985).

  33. Zhou, Y. et al. Giant magnetoelastic effect in soft systems for bioelectronics. Nat. Mater. 20, 1670–1676 (2021).

    Article  CAS  PubMed  Google Scholar 

  34. Qian, L. & Li, D. Use of magnetic fluid in accelerometers. J. Sens. 2014, 375623 (2014).

    Article  Google Scholar 

  35. Bailey, R. L. Lesser known applications of ferrofluids. J. Magn. Magn. Mater. 39, 178–182 (1983).

    Article  Google Scholar 

  36. Kole, M. & Khandekar, S. Engineering applications of ferrofluids: a review. J. Magn. Magn. Mater. 537, 168222 (2021).

    Article  CAS  Google Scholar 

  37. Impedovo, D., Pirlo, G., Vessio, G. & Angelillo, M. T. A handwriting-based protocol for assessing neurodegenerative dementia. Cognit. Comput. 11, 576–586 (2019).

    Article  Google Scholar 

  38. Balestrino, R. & Schapira, A. H. V. Parkinson disease. Eur. J. Neurol. 27, 27–42 (2020).

    Article  CAS  PubMed  Google Scholar 

  39. Kamran, I., Naz, S., Razzak, I. & Imran, M. Handwriting dynamics assessment using deep neural network for early identification of Parkinson’s disease. Future Gener. Comput. Syst. 117, 234–244 (2021).

    Article  Google Scholar 

  40. Fereshtehnejad, S.-M. et al. Evolution of prodromal Parkinson’s disease and dementia with Lewy bodies: a prospective study. Brain 142, 2051–2067 (2019).

    Article  PubMed  Google Scholar 

  41. Zerveas, G., Jayaraman, S., Patel, D., Bhamidipaty, A. & Eickhoff, C. A transformer-based framework for multivariate time series representation learning. In Proc. 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining 2114–2124 (Association for Computing Machinery, 2021).

  42. Wen, Q. et al. Transformers in time series: a survey. In Proc. Thirty-Second International Joint Conference on Artificial Intelligence 6778–6786 (Association for Computing Machinery, 2023).

  43. Ahmed, S. et al. Transformers in time-series analysis: a tutorial. Circ. Syst. Signal Process. 42, 7433–7466 (2023).

    Article  Google Scholar 

  44. Classify Time Series Using Wavelet Analysis and Deep Learning (MATLAB, 2024).

  45. Szegedy, C. et al. Going deeper with convolutions. In Proc. IEEE Conference on Computer Vision and Pattern Recognition 1–9 (CVPR, 2015).

  46. Zhuang, F. et al. A comprehensive survey on transfer learning. Proc. IEEE 109, 43–76 (2021).

    Article  Google Scholar 

  47. Eklund, M. et al. Diagnostic value of micrographia in Parkinson’s disease: a study with [123I]FP-CIT SPECT. J. Neural Transm. 129, 895–904 (2022).

    Article  CAS  PubMed  Google Scholar 

  48. Urso, D., van Wamelen, D. J., Trivedi, D., Ray Chaudhuri, K. & Falup-Pecurariu, C. in International Review of Movement Disorders, Vol. 5 (eds Ferro, A. S. & Monje, M. H. G.) 49–70 (Academic Press, 2023).

  49. Mirelman, A., Rochester, L., Simuni, T. & Hausdoff, J. M. Digital mobility measures to predict Parkinson’s disease. Lancet Neurol. 22, 1098–1100 (2023).

    Article  PubMed  Google Scholar 

  50. He, T., Chen, J., Xu, X. & Wang, W. Exploiting smartphone voice recording as a digital biomarker for Parkinson’s disease diagnosis. IEEE Trans. Instrum. Meas. 73, 1–12 (2024).

    Google Scholar 

  51. Yang, Y. et al. Artificial intelligence-enabled detection and assessment of Parkinson’s disease using nocturnal breathing signals. Nat. Med. 28, 2207–2215 (2022).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  52. Liu, Y. et al. Monitoring gait at home with radio waves in Parkinson’s disease: a marker of severity, progression, and medication response. Sci. Transl. Med. 14, eadc9669 (2022).

    Article  PubMed  Google Scholar 

  53. Fereshtehnejad, S.-M., Zeighami, Y., Dagher, A. & Postuma, R. B. Clinical criteria for subtyping Parkinson’s disease: biomarkers and longitudinal progression. Brain 140, 1959–1976 (2017).

    Article  PubMed  Google Scholar 

  54. Lin, C.-H. et al. Blood NfL. Neurology 93, e1104–e1111 (2019).

    Article  CAS  PubMed  Google Scholar 

  55. Mitchell, T. et al. Emerging neuroimaging biomarkers across disease stage in Parkinson disease: a review. JAMA Neurol. 78, 1262–1272 (2021).

    Article  PubMed  PubMed Central  Google Scholar 

  56. Ates, H. C. et al. End-to-end design of wearable sensors. Nat. Rev. Mater. 7, 887–907 (2022).

    Article  PubMed  PubMed Central  Google Scholar 

  57. Qamar, S. Healthcare data analysis by feature extraction and classification using deep learning with cloud based cyber security. Comput. Electr. Eng. 104, 108406 (2022).

    Article  Google Scholar 

  58. Yang, Y. et al. In-sensor dynamic computing for intelligent machine vision. Nat. Electron. 7, 225–233 (2024).

    Article  CAS  Google Scholar 

  59. Ni, Y. et al. Visualized in-sensor computing. Nat. Commun. 15, 3454 (2024).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  60. Chikwetu, L. et al. Does deidentification of data from wearable devices give us a false sense of security? A systematic review. Lancet Digit. Health 5, e239–e247 (2023).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

J.C. acknowledges the Vernroy Makoto Watanabe Excellence in Research Award at the UCLA Samueli School of Engineering, the Office of Naval Research Young Investigator Award (award ID N00014-24-1-2065), National Institutes of Health Grant (award ID R01 CA287326), National Science Foundation Grant (award number 2425858), the American Heart Association Innovative Project Award (award ID 23IPA1054908), the American Heart Association Transformational Project Award (award ID 23TPA1141360) and the American Heart Association’s Second Century Early Faculty Independence Award (award ID 23SCEFIA1157587). S.L. acknowledges the National Institute of Health (NS126918) and the Broad Stem Cell Research Center, the Jonsson Comprehensive Cancer Center and California NanoSystems Institute at UCLA. G.C. acknowledges the Amazon Doctoral Student Fellowship from Amazon AWS and the UCLA Science Hub for Humanity and Artificial Intelligence. G.C. also acknowledges the Predoctoral Fellowship from the American Heart Association and The VIVA Foundation (award ID 24PRE1193744). T.T. and J.C. acknowledge the Caltech/UCLA joint NIH T32 Training Grant (award ID T32EB027629). We also acknowledge the careful editing from the UCLA Writing Center for a one-on-one personalized writing consultation.

Author information

Authors and Affiliations

  1. Department of Bioengineering, University of California, Los Angeles, Los Angeles, CA, USA

    Guorui Chen, Trinny Tat, Yihao Zhou, Zhaoqi Duan, Kamryn Scott, Xun Zhao, Zeyang Liu, Song Li & Jun Chen

  2. Department of Computer Science, University of California, Los Angeles, Los Angeles, CA, USA

    Junkai Zhang & Wei Wang

  3. Department of Medicine, University of California, Los Angeles, Los Angeles, CA, USA

    Song Li

  4. Department of Neurology, University of California, Los Angeles, Los Angeles, CA, USA

    Katy A. Cross

Authors
  1. Guorui Chen
    View author publications

    Search author on:PubMed Google Scholar

  2. Trinny Tat
    View author publications

    Search author on:PubMed Google Scholar

  3. Yihao Zhou
    View author publications

    Search author on:PubMed Google Scholar

  4. Zhaoqi Duan
    View author publications

    Search author on:PubMed Google Scholar

  5. Junkai Zhang
    View author publications

    Search author on:PubMed Google Scholar

  6. Kamryn Scott
    View author publications

    Search author on:PubMed Google Scholar

  7. Xun Zhao
    View author publications

    Search author on:PubMed Google Scholar

  8. Zeyang Liu
    View author publications

    Search author on:PubMed Google Scholar

  9. Wei Wang
    View author publications

    Search author on:PubMed Google Scholar

  10. Song Li
    View author publications

    Search author on:PubMed Google Scholar

  11. Katy A. Cross
    View author publications

    Search author on:PubMed Google Scholar

  12. Jun Chen
    View author publications

    Search author on:PubMed Google Scholar

Contributions

J.C. conceived the idea and guided the entire project. J.C., G.C., Y.Z. and X.Z. designed the experiments, analyzed the data, drew the figures and composed the paper. G.C., T.T. and K.A.C. contributed to the human studies. G.C., X.Z., J.C., Z.L., Z.D. and Y.Z. contributed to the device design, fabrication and characterization. Z.D. and Y.Z. contributed to the theory study. J.Z., W.W. and G.C. contributed to the machine learning study. J.C. and S.L. contributed to the funding acquisition. G.C., K.A.C., T.T., K.S., S.L. and J.C. revised the paper. All authors have read the paper, agreed to its content and approved the final submission.

Corresponding author

Correspondence to Jun Chen.

Ethics declarations

Competing interests

J.C. and G.C. are inventors on a provisional patent application (UCLA case no. 2025-283) related to the development and application of the diagnostic pen, filed by the University of California, Los Angeles. The other authors declare no competing interests.

Peer review

Peer review information

Nature Chemical Engineering thanks Sibo Cheng, Martin J. McKeown, Huiliang Wang and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data

Extended Data Fig. 1 Magnetic properties and normal field instability of the ferrofluid ink.

a, The ferromagnetic material is composed of tiny magnetic domains with magnetic spins. b, The superparamagnetic ferrofluid ink has a quick response to the external magnetic field and minimal hysteresis. c, A transmission electron microscopy image of the single domain nanomagnets inside the ferrofluid ink. Scale bar, 40 nm. Inset is the enlarged view. Scale bar, 10 nm. d, These single domain nanomagnets have a particle diameter of 7.13 ± 1.40 nm. e-g, Magnetic hysteresis loops of the ferrofluid ink. (e) Under a low external magnetic field, the single domain nanomagnets exhibit a near-random distribution due to minimal magnetic interactions. (f) As the external magnetic field increases, the nanomagnets begin to partially align along the field direction. (g) With a further increase in the magnetic field, the single domain nanomagnets become predominantly aligned with the external magnetic field. h-j, Surface topography of the ferrofluid ink under increasing external magnetic fields, resulting from the balance between surface tension, gravity, and electromagnetic stress. (h) Initial state of the substrate containing the ferrofluid ink. (i) A high-speed camera captures the frame where the spike pattern is observed. (j) Closer proximity of the external magnetic field leads to larger and more pronounced spike patterns. Scale bars, 1.5 mm.

Source data

Extended Data Fig. 2 Block diagram showing the design of the pilot human study.

To recruit the participants, study flyers are distributed, and physician referrals are requested. Telephone screenings and information sessions ensure that participants qualify according to specific inclusion and exclusion criteria. With informed consent, participants write specific tasks using a diagnostic pen. Collected clinical data is securely stored with authorized access restricted to essential team members. Finally, a convolutional neural network analyzes the handwriting signals for PD diagnostics. Figure partially created with BioRender.com.

Extended Data Fig. 3 Personalized handwriting analysis.

a, System-level design of using the diagnostic pen for personalized handwriting analysis. b, Current signals recorded while using the diagnostic pen to draw continuous wavy lines (Task 1) for three cycles on the surface. c, Current signals recorded while using the diagnostic pen to draw spirals (Task 2) for three cycles on the surface. d, Current signals recorded while using the diagnostic pen to write letters (Task 3) for three cycles on the surface. A.U., arbitrary units. Panel a partially created with BioRender.com.

Extended Data Fig. 4 Handwriting signal analysis for in-air handwriting tasks.

a, Representative current signals from a recruited participant using the diagnostic pen to draw continuous wavy lines (Task 1) for three cycles in the air. b, Representative current signals from a recruited participant using the diagnostic pen to draw spirals (Task 2) for three cycles in the air. c, Representative current signals from a recruited participant using the diagnostic pen to write letters (Task 3) for three cycles in the air.

Supplementary information

Supplementary Information

Supplementary Figs. 1–24, Tables 1–5 and Notes 1–15.

Reporting Summary

Supplementary Video 1

Normal field instability of ferrofluid ink.

Supplementary Video 2

Spike pattern formation of ferrofluid ink under magnetization.

Supplementary Video 3

A patient with PD using the diagnostic pen for a writing task.

Source data

Source Data Fig. 1

Statistical source data for Fig. 1.

Source Data Fig. 2

Statistical source data for Fig. 2.

Source Data Fig. 3

Statistical source data for Fig. 3.

Source Data Fig. 4

Statistical source data for Fig. 4.

Source Data Extended Data Fig. 1

Statistical source data for Extended Data Fig. 1.

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Chen, G., Tat, T., Zhou, Y. et al. Neural network-assisted personalized handwriting analysis for Parkinson’s disease diagnostics. Nat Chem Eng 2, 358–368 (2025). https://doi.org/10.1038/s44286-025-00219-5

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue date:

  • DOI: https://doi.org/10.1038/s44286-025-00219-5

Search

Advanced search

Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing