pith. sign in

arxiv: 2604.22763 · v1 · submitted 2026-03-19 · 💻 cs.HC

A learning health system in Neurorehabilitation as a foundation for multimodal patient representation

Thomas Weikert , Eljas Roellin , Lukas Heumos , Fabian J. Theis , Diego Paez-Granados , Chris Easthope Awai This is my paper

Pith reviewed 2026-05-15 08:43 UTC · model grok-4.3

classification 💻 cs.HC
keywords learning health systemneurorehabilitationstroke rehabilitationmultimodal datacomputational modelsclinical visualizationclinician-ML collaborationpatient trajectories
0
0 comments X

The pith

Embedding a learning health system in neurorehabilitation integrates multimodal data collection, model computation, and visualization to support clinician-machine learning collaboration during stroke rehabilitation.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper establishes that a learning health system framework can be embedded in neurorehabilitation by combining multimodal data collection with computational models and clinical visualizations. This setup enables structured data capture, secure processing, and interoperable patient trajectory views that bring research models into daily clinical workflows. A real-world deployment in stroke rehabilitation demonstrates the system bridging gaps between fragmented data systems and practical use. The approach addresses barriers like poor interoperability and low clinician engagement with data-driven tools, offering a pathway for computational methods to personalize long-term neurological care.

Core claim

We embed the learning health system framework in neurorehabilitation through integration of multimodal data collection, model computation, and clinical visualization that enables clinician-ML collaboration in everyday neurorehabilitation practice. The system facilitates structured digital data capture, secure computational processing, and interoperable visualization of patient trajectories. Through a real-world deployment in stroke rehabilitation, we demonstrate how such an infrastructure bridges the gap between research models and clinical use, showcasing one approach to a translational pathway for computational neurorehabilitation.

What carries the argument

The learning health system framework, which integrates multimodal data collection, model computation, and clinical visualization to enable ongoing clinician-ML collaboration in daily practice.

Load-bearing premise

Fragmented data systems, poor interoperability, and low clinician engagement can be overcome through integration of multimodal data collection, model computation, and visualization into everyday neurorehabilitation workflows.

What would settle it

If deployment in stroke rehabilitation leaves data still fragmented or shows no increase in clinicians using model outputs and visualizations for treatment decisions, the integration would fail to bridge research and clinical use.

Figures

Figures reproduced from arXiv: 2604.22763 by Chris Easthope Awai, Diego Paez-Granados, Eljas Roellin, Fabian J. Theis, Lukas Heumos, Thomas Weikert.

Figure 1
Figure 1. Figure 1: (a) Technology platform: Schematic overview depicting three layers. Data collection from [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
read the original abstract

Neurological disorders represent a growing global health burden requiring long-term, interdisciplinary rehabilitation. Computational neurorehabilitation (compNR) - the use of data-driven and model-based approaches to personalize treatment - offers new opportunities for precision rehabilitation. However, its clinical deployment is limited by fragmented data systems, poor interoperability, and low clinician engagement in model development. We embed the learning health system (LHS) framework in Neurorehabilitation through integration of multimodal data collection, model computation, and clinical visualization that enables clinician-ML collaboration in everyday neurorehabilitation practice. The system facilitates structured digital data capture, secure computational processing, and interoperable visualization of patient trajectories. Through a real-world deployment in stroke rehabilitation, we demonstrate how such an infrastructure bridges the gap between research models and clinical use, showcasing one approach to a translational pathway for compNR.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 1 minor

Summary. The manuscript describes the design of a learning health system (LHS) for neurorehabilitation that integrates multimodal data collection, secure model computation, and interoperable clinical visualization. It claims that a real-world deployment in stroke rehabilitation demonstrates how this infrastructure bridges research models to clinical use by overcoming fragmented data systems, poor interoperability, and low clinician engagement.

Significance. If the deployment were shown to produce measurable improvements in data completeness, interoperability, and clinician-model interaction, the work would supply a concrete translational template for computational neurorehabilitation, moving data-driven personalization from research prototypes into routine practice.

major comments (1)
  1. [Deployment description] The section describing the real-world deployment supplies only an architectural narrative and does not report any quantitative indicators (pre/post data-capture completeness, interoperability latency or error rates, clinician session logs, model-view frequency, or changes in treatment decisions) that would substantiate the claim that the system overcomes the three named barriers.
minor comments (1)
  1. [Abstract] The abstract asserts that the deployment 'demonstrates' bridging of the research-clinical gap; this phrasing should be softened to 'illustrates a possible pathway' until supporting metrics are added.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their constructive feedback on our manuscript describing a learning health system for neurorehabilitation. We address the major comment below and will revise the manuscript to strengthen the evidence from the deployment.

read point-by-point responses

  1. Referee: The section describing the real-world deployment supplies only an architectural narrative and does not report any quantitative indicators (pre/post data-capture completeness, interoperability latency or error rates, clinician session logs, model-view frequency, or changes in treatment decisions) that would substantiate the claim that the system overcomes the three named barriers.

    Authors: We agree that the deployment section is primarily architectural and lacks the requested quantitative indicators. In the revised manuscript we will add a new subsection that reports pre/post-deployment metrics on data-capture completeness, interoperability latency and error rates, clinician session logs, model-view frequency, and observed changes in treatment decisions drawn from the stroke rehabilitation deployment. These additions will directly substantiate how the infrastructure addresses fragmented data systems, poor interoperability, and low clinician engagement. revision: yes

Circularity Check

0 steps flagged

No significant circularity; paper is a descriptive systems report without derivations or fitted predictions

full rationale

The manuscript describes an integrated learning health system architecture for neurorehabilitation, emphasizing multimodal data capture, secure computation, and visualization components deployed in stroke rehabilitation. No equations, parameter fits, predictions, or first-principles derivations appear in the provided text. The central claim rests on a narrative account of the deployment rather than any reduction of outputs to inputs by construction. No self-citation load-bearing steps, uniqueness theorems, or ansatz smuggling are present. The work is therefore self-contained as an implementation description and receives the default non-circularity finding.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The approach rests on the domain assumption that structured digital data capture and interoperable visualization will improve clinician engagement and model use; no free parameters or invented entities are introduced in the abstract.

axioms (1)
  • domain assumption Multimodal data integration and visualization can overcome fragmented data systems and low clinician engagement in neurorehabilitation
    Invoked as the basis for embedding the LHS framework and enabling clinician-ML collaboration.

pith-pipeline@v0.9.0 · 5462 in / 1142 out tokens · 33993 ms · 2026-05-15T08:43:51.241447+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

31 extracted references · 31 canonical work pages

  1. [1]

    rehabilitation 2030: a call for action

    Francesca Gimigliano and Stefano Negrini. The world health organization “rehabilitation 2030: a call for action”.European Journal of Physical and Rehabilitation Medicine, 53:155–168, 4

    work page 2030

  2. [2]

    doi: 10.23736/S1973-9087.17.04746-3

    work page doi:10.23736/s1973-9087.17.04746-3

  3. [3]

    How will the main risk factors contribute to the burden of non-communicable diseases under different scenarios by 2050? a modelling study.PLoS One, 15(4):e0231725, 2020

    Marion Devaux, Aliénor Lerouge, Giovanna Giuffre, Susanne Giesecke, Sara Baiocco, Andrea Ricci, Francisco Reyes, David Cantarero, Bruno Ventelou, and Michele Cecchini. How will the main risk factors contribute to the burden of non-communicable diseases under different scenarios by 2050? a modelling study.PLoS One, 15(4):e0231725, 2020

    work page 2050

  4. [4]

    Global burden of cardiovascular diseases: projections from 2025 to 2050.European Journal of Preventive Cardiology, page zwae281, 2024

    Bryan Chong, Jayanth Jayabaskaran, Silingga Metta Jauhari, Siew Pang Chan, Rachel Goh, Martin Tze Wah Kueh, Henry Li, Yip Han Chin, Gwyneth Kong, Vickram Vijay Anand, et al. Global burden of cardiovascular diseases: projections from 2025 to 2050.European Journal of Preventive Cardiology, page zwae281, 2024

    work page 2025

  5. [5]

    Yiran Cheng, Yongqi Lin, Hujuan Shi, Mingmei Cheng, Baoliang Zhang, Xiaofeng Liu, Chuan Shi, Yanzhong Wang, Chunhua Xia, and Wanqing Xie. Projections of the stroke burden at the global, regional, and national levels up to 2050 based on the global burden of disease study 2021.Journal of the American Heart Association, 13(23):e036142, 2024

    work page 2050

  6. [6]

    Burnout in healthcare workers: prevalence, impact and preventative strategies

    Stefan De Hert. Burnout in healthcare workers: prevalence, impact and preventative strategies. Local and regional anesthesia, pages 171–183, 2020

    work page 2020

  7. [7]

    Artificial intelligence, machine learning and health systems.Journal of global health, 8(2):020303, 2018

    Trishan Panch, Peter Szolovits, and Rifat Atun. Artificial intelligence, machine learning and health systems.Journal of global health, 8(2):020303, 2018

    work page 2018

  8. [8]

    Computa- tional neurorehabilitation: modeling plasticity and learning to predict recovery.Journal of NeuroEngineering and Rehabilitation, 13(1):1–25, 2016

    David J Reinkensmeyer, Etienne Burdet, Maura Casadio, John W Krakauer, Gert Kwakkel, Catherine E Lang, Stephan P Swinnen, Nick S Ward, and Nicolas Schweighofer. Computa- tional neurorehabilitation: modeling plasticity and learning to predict recovery.Journal of NeuroEngineering and Rehabilitation, 13(1):1–25, 2016

    work page 2016

  9. [9]

    Springer, 2022

    Mark Johnson and Emily Smith.Computational Neurorehabilitation: AI and Robotics in Therapy. Springer, 2022. ISBN 978-3-031-08995-4

    work page 2022

  10. [10]

    Towards ai-based precision rehabilitation via contextual model-based rein- forcement learning.medRxiv, pages 2025–01, 2025

    Dongze Ye et al. Towards ai-based precision rehabilitation via contextual model-based rein- forcement learning.medRxiv, pages 2025–01, 2025

    work page 2025

  11. [11]

    A causal framework for precision rehabilitation

    R James Cotton, Bryant A Seamon, Richard L Segal, Randal D Davis, Amrita Sahu, Michelle M McLeod, Pablo Celnik, and Sharon L Ramey. A causal framework for precision rehabilitation. arXiv preprint arXiv:2411.03919, 2024

    work page arXiv 2024

  12. [12]

    Digital twins for health: a scoping review.NPJ Digital Medicine, 7 (1):77, 2024

    Evangelia Katsoulakis et al. Digital twins for health: a scoping review.NPJ Digital Medicine, 7 (1):77, 2024

    work page 2024

  13. [13]

    Enabling precision rehabilitation interventions using wearable sensors and machine learning to track motor recovery

    C Adans-Dester, N Hankov, A O’Brien, G Vergara-Diaz, R Black-Schaffer, R Zafonte, et al. Enabling precision rehabilitation interventions using wearable sensors and machine learning to track motor recovery. npj digital medicine. 2020; 3 (1): 1–10.PM&R, 13:535–9, 2021

    work page 2020

  14. [14]

    Johannes Pohl, Geert Verheyden, Jeremia Philipp Oskar Held, Andreas Ruediger Luft, Chris Easthope Awai, and Janne Marieke Veerbeek. Construct validity and responsiveness of clinical upper limb measures and sensor-based arm use within the first year after stroke: a longitudinal cohort study.Journal of NeuroEngineering and Rehabilitation, 22(1):14, 2025

    work page 2025

  15. [15]

    Validation of real-world actigraphy to capture post-stroke motor recovery.medRxiv, pages 2024–11, 2024

    Keith R Lohse, Allison E Miller, Marghuretta D Bland, Jin-Moo Lee, and Catherine E Lang. Validation of real-world actigraphy to capture post-stroke motor recovery.medRxiv, pages 2024–11, 2024

    work page 2024

  16. [16]

    Accuracy of gait and posture classification using movement sensors in individuals with mobility impairment after stroke.Frontiers in Physiology, 13:933987, 2022

    Johannes Pohl, Alain Ryser, Janne Marieke Veerbeek, Geert Verheyden, Julia Elisabeth V ogt, Andreas Rüdiger Luft, and Chris Awai Easthope. Accuracy of gait and posture classification using movement sensors in individuals with mobility impairment after stroke.Frontiers in Physiology, 13:933987, 2022

    work page 2022

  17. [17]

    Classification of functional and non-functional arm use by inertial measurement units in individuals with upper limb impairment after stroke

    Johannes Pohl, Alain Ryser, Janne Marieke Veerbeek, Geert Verheyden, Julia Elisabeth V ogt, Andreas Rüdiger Luft, and Chris Awai Easthope. Classification of functional and non-functional arm use by inertial measurement units in individuals with upper limb impairment after stroke. Frontiers in Physiology, 13:952757, 2022. 5

    work page 2022

  18. [18]

    Opencap: Human movement dynamics from smartphone videos.PLoS computational biology, 19(10):e1011462, 2023

    Scott D Uhlrich, Antoine Falisse, Łukasz Kidzi ´nski, Julie Muccini, Michael Ko, Akshay S Chaudhari, Jennifer L Hicks, and Scott L Delp. Opencap: Human movement dynamics from smartphone videos.PLoS computational biology, 19(10):e1011462, 2023

    work page 2023

  19. [19]

    Differentiable biomechanics for markerless motion capture in upper limb stroke rehabilitation: A comparison with optical motion capture.arXiv preprint, 2024

    Tim Unger et al. Differentiable biomechanics for markerless motion capture in upper limb stroke rehabilitation: A comparison with optical motion capture.arXiv preprint, 2024

    work page 2024

  20. [20]

    Towards modeling evolving longitudinal health trajectories with a transformer- based deep learning model.Annals of Epidemiology, 2025

    Hans Moen, Vishnu Raj, Andrius Vabalas, Markus Perola, Samuel Kaski, Andrea Ganna, and Pekka Marttinen. Towards modeling evolving longitudinal health trajectories with a transformer- based deep learning model.Annals of Epidemiology, 2025

    work page 2025

  21. [21]

    Learning the natural his- tory of human disease with generative transformers.Nature, pages 1–9, 2025

    Artem Shmatko, Alexander Wolfgang Jung, Kumar Gaurav, Søren Brunak, Laust Hvas Mortensen, Ewan Birney, Tom Fitzgerald, and Moritz Gerstung. Learning the natural his- tory of human disease with generative transformers.Nature, pages 1–9, 2025

    work page 2025

  22. [22]

    Device-and analytics-agnostic infrastructure for continuous inpatient moni- toring: a technical note.Digital biomarkers, 4(2):62–68, 2020

    Noé Brasier, Lukas Geissmann, Miro Käch, Markus Mutke, Bianca Hoelz, Fiorangelo De Ieso, and Jens Eckstein. Device-and analytics-agnostic infrastructure for continuous inpatient moni- toring: a technical note.Digital biomarkers, 4(2):62–68, 2020

    work page 2020

  23. [23]

    doi: 10.1109/IEEESTD.2022.9762855

    Ieee standard for wearable consumer electronic devices–overview and architecture.IEEE Std 360-2022, pages 1–35, 2022. doi: 10.1109/IEEESTD.2022.9762855

    work page doi:10.1109/ieeestd.2022.9762855 2022

  24. [24]

    A frame- work for value-creating learning health systems.Health research policy and systems, 17(1):79, 2019

    Matthew Menear, Marc-André Blanchette, Olivier Demers-Payette, and Denis Roy. A frame- work for value-creating learning health systems.Health research policy and systems, 17(1):79, 2019

    work page 2019

  25. [25]

    Learning health system strategies in the ai era.npj Health Systems, 2(1):21, 2025

    Peter AD Steel, Gabriel Wardi, Robert A Harrington, and Christopher A Longhurst. Learning health system strategies in the ai era.npj Health Systems, 2(1):21, 2025

    work page 2025

  26. [26]

    An open- source framework for end-to-end analysis of electronic health record data.Nature medicine, 30 (11):3369–3380, 2024

    Lukas Heumos, Philipp Ehmele, Tim Treis, Julius Upmeier zu Belzen, Eljas Roellin, Lilly May, Altana Namsaraeva, Nastassya Horlava, Vladimir A Shitov, Xinyue Zhang, et al. An open- source framework for end-to-end analysis of electronic health record data.Nature medicine, 30 (11):3369–3380, 2024

    work page 2024

  27. [27]

    Ros 2 healthcare: Closing the loop between human sensing and robotic interventions

    Jorge Peña-Queralta, Ricardo Javier Manríquez Cisterna, Heba Hussein, Oriella Gnarra, Mehdi Ejtehadi, Ankit A Ravankar, Jose Victorio Salazar Luces, Yasuhisa Hirata, and Diego Paez- Granados. Ros 2 healthcare: Closing the loop between human sensing and robotic interventions. InAssistive Systems: Lab to Patient Care-Full Day Workshop (ICRA-24). Assistive S...

    work page 2024

  28. [28]

    Automated prediction of item-level arat scores from wearable sensors

    Thomas Weikert, Yanke Li, Diego Paez-Granados, and Chris Awai Easthope. Automated prediction of item-level arat scores from wearable sensors. In2025 International Conference On Rehabilitation Robotics (ICORR), pages 1239–1244. IEEE, 2025

    work page 2025

  29. [29]

    Identiarat: Toward automated identification of individual arat items from wearable sensors.arXiv preprint arXiv:2504.12921, 2025

    Daniel Homm, Patrick Carqueville, Christian Eichhorn, Thomas Weikert, Thomas Menard, David A Plecher, and Chris Awai Easthope. Identiarat: Toward automated identification of individual arat items from wearable sensors.arXiv preprint arXiv:2504.12921, 2025

    work page arXiv 2025

  30. [30]

    Upper limb movement quality measures: comparing imus and optical motion capture in stroke patients performing a drinking task.Frontiers in Digital Health, 6:1359776, 2024

    Tim Unger, Ricardo de Sousa Ribeiro, Marwen Mokni, T Weikert, J Pohl, A Schwarz, JPO Held, L Sauerzopf, B Kühnis, E Gavagnin, et al. Upper limb movement quality measures: comparing imus and optical motion capture in stroke patients performing a drinking task.Frontiers in Digital Health, 6:1359776, 2024

    work page 2024

  31. [31]

    Usability of capturing upper limb movement kinematics in clinical routine: Do wearable sensors beat markerless motion capture?Gait & Posture, 121:66–67, 2025

    Tim Unger, Xin Yao, Anna Schmitt, Charlotte Saillard, Elena Ruz, Alma Gaite Flueck, and Chris Easthope Awai. Usability of capturing upper limb movement kinematics in clinical routine: Do wearable sensors beat markerless motion capture?Gait & Posture, 121:66–67, 2025. 6

    work page 2025