Pith. sign in

REVIEW 3 major objections 5 minor 15 references

EMG-Driven Stiffness-Modulating Palpation for Telerehabilitation

T0 review · 3 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read A thumb-worn haptic device turns EMG signals into palpable stiffness for remote palpation.

desk verdict A clean servo demo with an unvalidated clinical link: the EMG-to-stiffness map is the load-bearing assumption the paper doesn't test. read the letter →

arxiv 2506.08303 v2 pith:DBAELY7X submitted 2025-06-10 cs.HC

Thomas M. Kwok , Hilary HY Cheng , Wai Tuck Chow This is my paper
classification cs.HC
keywords EMG-drivenhapticshoneycombjammingtelerehabilitationremotepalpationkinestheticfeedbackmusclestiffnesswearablehapticdevicevacuumpressuretracking
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper introduces HJ-Pal, a thumb-wearable haptic device that turns electromyography (EMG) signals into changes in physical stiffness through a honeycomb-jamming mechanism, so a therapist can 'feel' a remote patient's muscle activation during telerehabilitation. The authors report that the device tracks EMG-derived pressure commands with about 1 kPa root-mean-square error and very high correlation, for both the supinator and pronator muscles. A sympathetic reading is that this establishes the first stage of the argument: the haptic hardware can respond faithfully to noisy muscle signals, before any claim about clinically meaningful stiffness perception. The paper itself leaves the validation of perceived stiffness and clinical usefulness to future work.

What carries the argument

The load-bearing mechanism is the honeycomb jamming structure: a honeycomb core wrapped in membrane and jamming layers changes stiffness up to 0.85 N/mm as vacuum pressure varies, and a strain gauge captures finger indentation while a PID-controlled vacuum pump sets the pressure. The EMG side processes raw signals with a 4th-order Butterworth bandpass filter (10–500 Hz), detrending, rectification, a moving mean smoother, and normalization to maximum voluntary contraction, then maps the result to desired pressure within the device's stiffness range. The combination lets a local, low-latency stiffness change stand in for delayed force feedback from a remote robot.

What would settle it

A blinded study in which therapists press HJ-Pal at several EMG-derived stiffness levels and must rank the corresponding muscle activation levels: if their rankings are no better than chance, or if an independent stiffness measure (e.g. shear-wave elastography) does not order the rendered pressures in the same way, the central claim of clinically useful kinesthetic feedback is refuted.

Watch

Extended reading notes

Core claim

The central claim is that HJ-Pal can render EMG-measured muscle activation as kinesthetic feedback for remote palpation. In a single-participant experiment, processed EMG from the supinator and pronator muscles was mapped to a desired vacuum pressure, and a PID-controlled pump tracked that pressure with RMSE of 1.006 ± 0.031 kPa and 0.904 ± 0.056 kPa respectively, with Pearson correlations r = 0.994 ± 0.0004 and r = 0.966 ± 0.010. On these grounds the paper concludes that HJ-Pal delivers reliable kinesthetic feedback based on muscle activity, positioning it as a tool for assessing small or deep muscles where force-feedback teleoperation suffers from latency and size mismatch.

Load-bearing premise

The mapping from normalized EMG to desired vacuum pressure is assumed to produce a stiffness that feels like the patient's actual muscle stiffness, but that mapping is not defined, calibrated, or validated against any clinical reference.

Editorial extensions

If this is right

  • Therapists could perceive muscle engagement remotely without the desynchronization caused by teleoperation latency.
  • Small or deep muscles, such as the pronator and supinator, become assessable because EMG electrodes can target them directly.
  • The high correlation between EMG command and vacuum pressure suggests the device can track rapid, noisy activation changes in real time.
  • This creates a basis for comparing EMG-based haptic feedback against visual EMG displays and force-based haptics in future studies.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The paper's unstated EMG-to-pressure mapping is the fragile link: if that scaling is arbitrary, the device could track commands precisely while feeling nothing like the patient's actual muscle stiffness to a therapist.
  • A direct test would be to ask blinded therapists to rank stiffness levels produced by different EMG amplitudes and compare those rankings with manual palpation or an imaging-based stiffness reference such as shear-wave elastography.
  • If validated, the same EMG-to-stiffness channel could be used to drive assistive exoskeletons, letting therapy adjust in real time to measured muscle engagement.
  • The single-participant, one-session design means generalizability across anatomy, fatigue, and electrode placement remains open.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper introduces HJ-Pal, a thumb-wearable haptic device that uses a honeycomb jamming mechanism to modulate stiffness in response to EMG signals, with the goal of enabling remote palpation for telerehabilitation. A preliminary experiment with a single participant recorded EMG from the supinator and pronator muscles, mapped the processed EMG to a desired vacuum pressure, and used a PID controller to track that pressure. The reported RMSE values are approximately 1 kPa and Pearson correlations with the EMG-derived command exceed 0.96. The authors conclude that HJ-Pal can deliver reliable kinesthetic feedback based on muscle activity and position the device as a promising tool for remote muscle assessment.

Significance. The concept of using EMG to drive stiffness-modulating haptics for remote palpation is a plausible and potentially valuable contribution to telerehabilitation, where conventional teleoperated force feedback suffers from latency and poor access to small muscles. The device is lightweight and builds on prior honeycomb-jamming work. The paper's principal strength is the quantitative servo-tracking result, which demonstrates that the vacuum-pressure system can follow EMG-derived commands with low error on one participant. However, the central claim about kinesthetic palpation is not yet supported: no measurement of rendered stiffness, no perceptual discrimination test, and no clinical validation are reported. The manuscript is honest in deferring these to future work, but the current evidence supports only pressure tracking, not muscle-stiffness rendering.

major comments (3)
  1. [Section III-A] The central claim that "HJ-Pal can deliver reliable kinesthetic feedback based on muscle activity" is not established by the reported experiment. The measured output is internal vacuum pressure, which is the same variable used as the command: EMG is normalized, converted to a desired pressure p_d, and the PID system tracks p_d. The high correlation and low RMSE therefore quantify servo tracking of an EMG-derived command, not the rendering of clinically meaningful stiffness. To support the palpation claim, the paper would need to measure the force-displacement or stiffness behavior of the device as a function of commanded pressure, and show that the resulting stiffness differences are perceptually discriminable and correspond to relevant muscle-activation levels. The future-work statement ("Future work will explore the correlation between EMG activity and HJ-Pal's perceived stiffness") explicitly acknowledges that this link is missing.
  2. [Section III] The mapping from processed EMG to desired pressure is not specified. The text says only that the processed signal is "mapped to the desired pressure (p_d) within HJ-Pal's available stiffness range," but the functional form (e.g., linear, logarithmic, gain, offset, normalization details) is omitted. Without this mapping, the experiment is not reproducible, and the reported statistics impose no constraint on whether the mapping is clinically appropriate. Any injective transformation of the EMG trace would yield the same near-perfect tracking, so the correlation values do not validate the mapping's clinical meaning.
  3. [Section III-A] The experiment uses a single participant performing five cycles, with no repeated sessions, no comparison conditions, and no statistical treatment across participants or trials. This is a reasonable pilot, but it is insufficient to support the unqualified conclusions that HJ-Pal "can deliver reliable kinesthetic feedback" and is "a promising tool for remote muscle assessment." The conclusions should be tempered to describe the result as a feasibility demonstration of EMG-driven pressure tracking, not a validated palpation device.
minor comments (5)
  1. [Section III-A] The reported values contain typographical errors: "1.0060.031kPa" should be "1.006 ± 0.031 kPa" (likewise "0.9040.056kPa" and the correlation values).
  2. [Section III] The text "Maximum V oluntary Contraction" contains a stray space; it should read "Maximum Voluntary Contraction."
  3. [Section I and II] The manuscript refers to itself as "this abstract" and is formatted like a workshop submission. If this is intended for journal publication, the framing and contribution statement should be revised to match a full-paper format.
  4. [Fig. 2] Figure 2(a) is noted as generated with Sora by OpenAI. It would be helpful to clarify in the caption that this is an illustrative concept image rather than a photograph of the actual experimental setup.
  5. [Section III] The paper relies on reference [7] for the mechanical design and characterization of the honeycomb jamming mechanism, but it does not report the stiffness range or its measurement here. Since the argument depends on stiffness modulation, a brief summary of the device's force-displacement behavior would improve readability.

Circularity Check

1 steps flagged · score 6.0 of 10

The reported 'confirmation' that HJ-Pal renders muscle activity is by construction a servo tracking result: desired pressure is an EMG-derived command, so high correlation with measured pressure is definitional, and the clinical stiffness link is deferred to future work.

  1. self definitional [Section III-A, Results & Discussion; Fig. 2 framework; Conclusion]
    "The processed EMG signals were mapped to desired pressure values (p_d), which were tracked by the PID-controlled vacuum system of HJ-Pal. ... the measured internal pressure closely followed the EMG-derived pressure commands. ... the measured pressure correlated strongly with the EMG-derived commands (Pearson's correlation with r = 0.994 ± 0.0004 in supinator; r = 0.966 ± 0.010 in pronator, all p < 0.0001), confirming that the HJ-Pal is capable of responding to the high-frequency, noisy nature of EMG input."

    The validation metric is the correlation between measured internal pressure p_h and the desired pressure p_d, where p_d is by construction a filtered, normalized transformation of the EMG signal. A functioning PID loop is designed to make p_h track p_d, so the reported RMSE and near-unity Pearson r are properties of the servo controller, not evidence that the rendered stiffness corresponds to muscle activation or to a clinically useful palpation sensation. Any injective EMG-to-pressure mapping would produce the same tracking statistics.

full rationale

The paper's quantitative evidence reduces by construction to command tracking. Section III states that processed EMG is 'mapped to desired pressure values (p_d)' and that the PID vacuum system tracks p_d; Section III-A then reports that measured pressure 'closely followed the EMG-derived pressure commands' with high correlation. Since p_d is defined from EMG and the controller is designed to track p_d, the high correlation is a servo result, not an independent validation of EMG-to-stiffness rendering. The jump to 'reliable kinesthetic feedback based on muscle activity' requires an unstated and unvalidated link between internal pressure, rendered stiffness, and the therapist's perception of muscle stiffness. The paper's own future-work sentence explicitly defers 'the correlation between EMG activity and the device's rendered stiffness,' which is the missing link. The self-citation to prior work [7] for the honeycomb-jamming hardware is not circular: it is external device characterization that does not assume the target palpation result. Because the headline confirmation is by construction while the central clinical palpation claim is not yet tested, the partial circularity score is 6.

Assumptions & free parameters 1 free parameters · 3 assumptions · 0 invented entities

The paper's contribution rests on a small number of inherited and assumed ingredients: EMG as a proxy for muscle state, the jamming stiffness range from prior work, and an unstated EMG-to-pressure mapping. No new theoretical entities are introduced.

free parameters (1)
  • EMG-to-pressure mapping gain/scaling = not reported
    The paper says 'processed EMG signals were mapped to desired pressure values (p_d)' but gives no transfer function or calibration details. The mapping is central because it determines the intensity of the haptic feedback.
assumptions (3)
  • domain assumption EMG amplitude is a valid proxy for muscle activation and stiffness in small or deep muscles.
    The entire feedback concept relies on this. Section I cites [14], but the paper does not establish this relationship for the tested muscles.
  • domain assumption The honeycomb jamming stiffness range (0.85 N/mm under vacuum) is perceptually sufficient for fingertip palpation.
    The stiffness range is inherited from prior work [7]; the paper assumes it is discriminable and clinically useful.
  • domain assumption The PID-controlled vacuum system's closed-loop bandwidth covers the relevant EMG frequency content.
    The paper only tests slow supination cycles; it does not show tracking at higher EMG frequencies.

how reviews work

0 comments
Cite this review

Pith. "Pith review of EMG-Driven Stiffness-Modulating Palpation for Telerehabilitation." pith.science (2026). https://pith.science/paper/DBAELY7X

@misc{pith2026250608303,
  author       = {Pith},
  title        = {Pith review of: EMG-Driven Stiffness-Modulating Palpation for Telerehabilitation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DBAELY7X}},
  note         = {Machine review of arXiv:2506.08303}
}
read the original abstract

In this work, we introduce HJ-Pal, a lightweight wearable haptic device that leverages EMG-driven honeycomb jamming to render muscle activation as kinesthetic feedback, enabling remote palpation for small muscle assessment in telerehabilitation.

Figures

Figures reproduced from arXiv: 2506.08303 by the authors.

Figure 3
Figure 3. Vacuum pressure tracking of EMG-derived commands during forearm [PITH_FULL_IMAGE:figures/full_fig_p002_3.png] view at source ↗
Figure 2
Figure 2. (a) Conceptual illustration of a telerehabilitation scenario using HJ [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

15 extracted references · 12 canonical work pages

  1. [1]

    Emerging challenges in meeting physiotherapy needs during covid-19 through telerehabilitation.Bulletin of Faculty of Physical Ther- apy, 25:1–4, 2020

    Joseph Ayotunde Aderonmu. Emerging challenges in meeting physiotherapy needs during covid-19 through telerehabilitation.Bulletin of Faculty of Physical Ther- apy, 25:1–4, 2020

    work page 2020

  2. [2]

    beam-me-in strategy

    Kilian Baur, Nina Rohrbach, Joachim Hermsd ¨orfer, Robert Riener, and Verena Klamroth-Marganska. The “beam-me-in strategy”–remote haptic therapist-patient interaction with two exoskeletons for stroke therapy. Journal of neuroengineering and rehabilitation, 16:1–15, 2019

    work page 2019

  3. [3]

    Is it time to rethink us- ing digital palpation for assessment of muscle stiffness? Neurourology and urodynamics, 39(1):279–285, 2020

    Melissa J Davidson, Poul MF Nielsen, Andrew J Taberner, and Jennifer A Kruger. Is it time to rethink us- ing digital palpation for assessment of muscle stiffness? Neurourology and urodynamics, 39(1):279–285, 2020

    work page 2020

  4. [4]

    Framework for human haptic perception with delayed force feedback.IEEE Transactions on Human-Machine Systems, 49(2):171–182, 2018

    Wei Fu, Marinus M van Paassen, David A Abbink, and Max Mulder. Framework for human haptic perception with delayed force feedback.IEEE Transactions on Human-Machine Systems, 49(2):171–182, 2018

    work page 2018

  5. [5]

    Kwok and Haoyong Yu

    Thomas M. Kwok and Haoyong Yu. A novel bilateral underactuated upper limb exoskeleton for post-stroke bimanual adl training.IEEE Transactions on Neural Systems and Rehabilitation Engineering, 32:3299–3309,

    work page

  6. [6]

    Kwok and Haoyong Yu

    Thomas M. Kwok and Haoyong Yu. Asymmetric bi- manual adl training with underactuated exoskeleton using independent joint control and visual guidance.IEEE Access, 12:9277–9291, 2024. doi: 10.1109/ACCESS. 2024.3352911

    arXiv 2024

  7. [7]

    Kwok, Bohan Zhang, and Wai Tuck Chow

    Thomas M. Kwok, Bohan Zhang, and Wai Tuck Chow. A Wearable Stiffness-Rendering Haptic Device with a Hon- eycomb Jamming Mechanism for Bilateral Teleoperation. Machines, 13(1):27, January 2025. ISSN 2075-1702. doi: 10.3390/machines13010027

    work page doi:10.3390/machines13010027 2025

  8. [8]

    Teleoperation of two six-degree-of- freedom arm rehabilitation exoskeletons

    Jessica Lanini, Toshiaki Tsuji, Peter Wolf, Robert Riener, and Domen Novak. Teleoperation of two six-degree-of- freedom arm rehabilitation exoskeletons. In2015 IEEE International Conference on Rehabilitation Robotics (ICORR), pages 514–519, 2015. doi: 10.1109/ICORR. 2015.7281251

    arXiv 2015

Show all 15 references
  1. [9]

    Telerehabilitation for upper limb disabilities: a scoping review on functions, outcomes, and evaluation methods.Archives of Public Health, 80(1):196, 2022

    Khadijeh Moulaei, Abbas Sheikhtaheri, Mansour Sha- habi Nezhad, AliAkbar Haghdoost, Mohammad Gheysari, and Kambiz Bahaadinbeigy. Telerehabilitation for upper limb disabilities: a scoping review on functions, outcomes, and evaluation methods.Archives of Public Health, 80(1):196, 2022

    2022

  2. [10]

    Mª Teresa Mu ˜noz-Tom´as, Mario Burillo-Lafuente, Araceli Vicente-Parra, Mª Concepci ´on Sanz-Rubio, Carmen Suarez-Serrano, Yolanda Marc ´en-Rom´an, and Mª ´Angeles Franco-Sierra. Telerehabilitation as a thera- peutic exercise tool versus face-to-face physiotherapy: a systemat...

    2023

  3. [11]

    OpenAI. Sora. https://openai.com/sora, 2024. Video and image generation platform

    2024

  4. [12]

    Cutaneous force feedback as a sensory subtrac- tion technique in haptics.IEEE Transactions on Haptics, 5:289–300, 2011

    Domenico Prattichizzo, Claudio Pacchierotti, and Giulio Rosati. Cutaneous force feedback as a sensory subtrac- tion technique in haptics.IEEE Transactions on Haptics, 5:289–300, 2011

    2011

  5. [13]

    Touch, texture and haptic feedback: a review on how we feel the world around us.Applied Sciences, 12(9):4686, 2022

    Aaron Raymond See, Jose Antonio G Choco, and Kohila Chandramohan. Touch, texture and haptic feedback: a review on how we feel the world around us.Applied Sciences, 12(9):4686, 2022

    2022

  6. [14]

    Tian Xie, Yan Leng, Yihua Zhi, Chao Jiang, Na Tian, Zichong Luo, Hairong Yu, and Rong Song. Increased muscle activity accompanying with decreased complex- ity as spasticity appears: High-density emg-based case studies on stroke patients.Frontiers in Bioengineering and Biotechn...

    2020

  7. [2024]

    doi: 10.1109/TNSRE.2024.3407653

    2024

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.