Pith. sign in

REVIEW 4 major objections 4 minor 107 references

Wirelessly transmitted subthalamic nucleus signals predict endogenous pain levels in Parkinson's disease patients

T0 review · 4 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read Wireless deep-brain signals can distinguish higher from lower pain in Parkinson's patients, the authors report.

desk verdict A plausible, honest small study showing STN LFP can decode pain fluctuations in PD, but the permutation test ignores session-level structure and the group-level significance rests on the minimum possible n=6. read the letter →

arxiv 2506.21439 v2 pith:VOVRARDD submitted 2025-06-26 q-bio.NC

Abdi Reza , Takufumi Yanagisawa , Naoki Tani , Ryohei Fukuma , Takuto Emura , Satoru Oshino , Ben Seymour , Haruhiko Kishima This is my paper
classification q-bio.NC
keywords Parkinson'sdiseasesubthalamicnucleuslocalfieldpotentialspainfluctuationsendogenousmachinelearningrandomforestbiomarker
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

The authors set out to test whether pain fluctuations in Parkinson's disease leave a readable trace in the subthalamic nucleus (STN), the brain region targeted by deep brain stimulation. Using wirelessly transmitted local field potentials from both STNs, they trained per-patient random forest classifiers to separate higher from lower self-reported pain after median-splitting each patient's VAS ratings. In four of six pain reports with PD-related characteristics, the LFP features decoded pain level significantly above chance, and the group-level average balanced accuracy was 67.99% ± 12.83% (P = 0.0156). If the result holds, STN spectral power is an objective biomarker of endogenous pain fluctuations that could eventually guide closed-loop stimulation and pain management.

What carries the argument

The load-bearing machinery is a per-patient, per-pain-report binary classifier whose inputs are fourteen non-normalized feature values: the average power-spectral-density amplitudes in seven canonical frequency bands (delta 1–4 Hz, theta 5–8 Hz, alpha 9–12 Hz, low beta 13–20 Hz, high beta 21–30 Hz, low gamma 31–60 Hz, high gamma 61–90 Hz) from the first principal component of three bipolar electrode pairs on each STN side. A fifth-order Butterworth filter (1–120 Hz) and Welch's method produce the spectra from roughly twenty-second wireless recordings. Pain ratings from a 0–100 visual analogue scale are binarized by each report's median, and reports with a class above 60% of observations are discarded as non-fluctuating. A random forest with nested five-fold stratified cross-validation performs the decoding, while permutation tests, Gini importance, and SHAP values determine significance and feature influence.

What would settle it

Record STN signals over a longer ambulatory period after peri-lead edema resolves and timestamp medication doses; if a day-blocked permutation test yields P >= 0.05 for the PDRP group, or if balanced accuracy collapses when medication state is added as a covariate, the pain-specific decoding claim is refuted.

Watch

Extended reading notes

Core claim

On its own terms, the paper's central claim is that the binary state of endogenous pain fluctuation—higher versus lower pain—can be read out of power in the subthalamic nucleus local field potential, using features from both hemispheres. Across six PD-related pain reports from six patients, a report-specific random forest reached a mean balanced accuracy of 67.99% ± 12.83%, significantly above chance (P = 0.0156, one-sided Wilcoxon signed-rank test; Bonferroni-adjusted α = 0.0167). Four of the six reports passed individual permutation tests (#B 68.6%, #C 86.3%, #E 74.6%, #F 71.7%; P = 0.018, 0.001, 0.012, 0.016). The same pipeline did not decode the two non-PD-related pain reports. The authors take this to show that STN activity stores an accessible representation of PD-related pain fluctuations, that both sides of the STN contribute, and that the influence is spectrally structured: beta and gamma bands dominate on the contralateral side for lateral pain, delta and theta on the ipsilateral side, and the right STN is more influential for pain felt in the body midline.

Load-bearing premise

The result stands or falls on the assumption that within-patient repeated observations are independent and that the median split into 'higher' and 'lower' pain captures pain fluctuations rather than day-to-day medication timing, peri-lead edema, or motor state changes.

Editorial extensions

If this is right

  • If the finding holds, STN local field potentials become an objective, continuously available readout of pain state in PD, usable when self-report is unavailable or unreliable.
  • Because the same wireless telemetry is already built into modern adaptive deep brain stimulation systems, a pain decoder could be implemented without new hardware, and stimulation could in principle be adjusted when the predicted pain level rises.
  • The specificity to PD-related pain annotations suggests that pain signatures in the STN are tied to PD pathophysiology rather than to pain in general, since non-PD-related pain reports did not decode above chance in this cohort.
  • Bilateral recording appears necessary: the contralateral STN (beta and gamma bands) and the ipsilateral STN (delta and theta bands) both carried predictive weight, so models using only one side may be missing information.
  • The STN-LFP model performed about as well as models built on self-reported motor symptoms, but it has the advantage of being electrophysiological and therefore not dependent on the patient's subjective report of motor state.

Reading between the lines

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

  • Editorial extension: a longer ambulatory recording would test whether these pain-specific spectral signatures persist after peri-lead edema resolves; the present data were collected within days of implantation, so the biomarker's stability over time is unknown.
  • Editorial extension: the significance tests assume exchangeable observations; a stricter day-level permutation test that resamples blocks of same-day sessions rather than individual ratings would show how much of the 68% group accuracy survives within-day autocorrelation.
  • Editorial extension: if the contralateral beta/gamma versus ipsilateral delta/theta split reflects a genuine somatotopic or hemispheric organization, then pain decoding should generalize to new patients with lateralized pain in the same body region; this is testable in a validation cohort.
  • Editorial extension: the comparison with mood/fatigue and motor-symptom models hints that a multimodal model fusing all three feature sets could outperform any single modality, but that fusion was not tested here.
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

4 major / 4 minor

Summary. The paper reports an observational study of Parkinson's disease patients with subthalamic nucleus (STN) deep brain stimulation, testing whether wirelessly transmitted STN local field potential (LFP) signals can predict endogenous pain fluctuations. From 31 pain-annotation reports collected from 11 patients, 8 reports from 6 patients were retained after exclusions for class imbalance and unilateral recordings. For each report, VAS pain ratings were median-split into binary labels, and a random forest with nested cross-validation was trained on bilateral canonical-band LFP features. Among six reports classified as PD-related pain (PDRP), four showed individually significant balanced accuracy in permutation tests, and the group-level one-sided Wilcoxon test yielded a mean balanced accuracy of 67.99% ± 12.83% (P = 0.0156, n = 6, Bonferroni-adjusted α = 0.0167). The authors also report bilateral STN contributions via Gini importance, with contralateral beta/gamma and ipsilateral delta/theta bands most influential, and compare the LFP model with alternative models based on mood, fatigue, and motor symptoms.

Significance. If the reported association is genuine, this is a useful step toward an objective, electrophysiological biomarker for pain fluctuations in PD and has plausible implications for adaptive DBS. The study has real strengths: personalized per-report modeling, nested cross-validation with hyperparameter selection kept inside the training loop, permutation tests with conservative P-value estimation, Bonferroni adjustment for the three feature-model families, and a transparent limitations section. The Gini/SHAP analysis of bilateral STN contributions is also a reasonable data-driven exploration. However, the central statistical claim currently rests on two fragile pillars: per-report permutation tests that assume exchangeability of observations collected up to four times per day over a short hospitalization, and a group-level Wilcoxon P that is the minimum attainable for n = 6. The manuscript does not provide public code or data, which limits independent verification, though the methods are described in sufficient detail to be replicable in principle.

major comments (4)
  1. [Statistics and Data collection (pp. 25–27, 34–35)] The per-report permutation test shuffles individual pain labels across all sessions, which is valid only if observations within each annotated pain report are exchangeable. The protocol collected up to four sessions per day over a short hospitalization, with medication, stimulation, rehabilitation, and peri-lead edema varying over days; the paper itself acknowledges these confounds in the Limitations paragraph (p. 23). The one-way ANOVA of time-of-day quartiles (e.g., Fig. 3c and Supplementary Figs. S2–S8) does not test serial autocorrelation or day-level confounding, and no medication or session covariate is modeled. If pain ratings and STN-LFP power both track slow day- or session-level states, the label-shuffling null destroys temporal structure retained in the true labels, making the null distribution too narrow and inflating the per-report P values (e.g., 0.018, 0.001, 0.012, 0.016 for reports #B, #C, #E, #F) and the group-level P. I request a block-respecting permutation test (e.g., permuting labels within days or session blocks) and/or a day/session-covariate analysis, with results reported for both per-report and group-level tests.
  2. [Preprocessing (p. 31) and STN-LFP features (p. 16)] PCA is applied to the three bipolar electrode pairs of each side before computing PSD features and before the nested cross-validation. Because the PC1 loadings are estimated from the full annotated pain report, including observations that later form test folds, this is an unsupervised preprocessing step with potential information leakage into the cross-validation estimate. The paper explicitly justifies non-normalized features by avoiding 'information leakage' (p. 16), but the PCA step is not embedded in the training folds. Please either re-estimate PC1 within each training fold, or show that the loadings are stable enough that the leak is negligible (e.g., by comparing CV accuracy with PCA fit on the training set only).
  3. [Results, Fig. 5] The group-level Wilcoxon signed-rank test gives P = 0.0156 for n = 6, which is the smallest attainable value for that sample size and clears the Bonferroni-adjusted α = 0.0167 by a small margin. The test pairs each report's balanced accuracy with its own shuffled mean and reaches significance only because all six PDRP reports have mean balanced accuracy above the shuffled average; two of the six reports (#A and #D) are not significant individually (P = 0.253 and 0.401). Any one report moving to chance or negative under a block-respecting null would make the group-level P non-significant. The conclusion that 'STN-LFP features effectively classified' PDRP pain should therefore be tempered, and the analysis should report how robust the group result is to removal of each report (a leave-one-report-out analysis).
  4. [Fig. 2 and Supplementary Table S1] The PDRP/non-PDRP classification is operationalized with a modified, unvalidated criterion set (Fig. 2), and the group-level analysis is restricted to the six reports classified as PDRP. The reliability of this classification is therefore load-bearing. Please report inter-rater reliability if two clinicians scored the reports, or at least provide a sensitivity analysis showing that the main results are stable under plausible alternative assignments of borderline reports (e.g., report #D, which lacks a temporal relationship with disease course in Supplementary Table S1).
minor comments (4)
  1. [Abstract and Introduction] The phrase 'annotated pain reports' is used repeatedly as a noun phrase; consider simplifying to 'pain annotations' or 'annotated pain series' for readability.
  2. [Figure 1] In the STROBE flowchart, the text 'Oneparticipanthadnoconsistentpaincomplaints...' is missing spaces and punctuation; please correct this formatting error.
  3. [Results, first paragraph] The sentence 'Twenty-threepain-annotationreportsmustbeexcluded' should read 'had to be excluded' or 'were excluded'.
  4. [Methods, Statistics subsection] The sentence 'a scatterplot that visualizes the relationship between pain rating and the ANOVA procedure is used' is unclear; it should describe the one-way ANOVA comparing pain ratings across four time quartiles, with the scatterplot as a visualization aid.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: STN-LFP decoding of pain levels is a genuine out-of-sample prediction task.

full rationale

The central claim is that balanced accuracy for classifying higher vs. lower pain (median-split VAS) from bilateral STN-LFP canonical-band PSD features is above chance, with significance assessed by label-permutation tests and a group-level Wilcoxon test. The labels are patient-reported VAS ratings; the features are wireless STN-LFP spectral amplitudes. Nested cross-validation keeps test folds out of training, and the permutation null shuffles labels while retraining the same random-forest pipeline, so the reported accuracies are not fitted values or re-statements of the training objective. The median-split and 60% balance inclusion rule define the cohort, but they do not constrain the classifier's out-of-sample accuracy; indeed two of six PDRP reports failed to decode significantly. The two self-citations (ref. 17, prior ML-on-STN-LFP work; ref. 49, prior surgical procedure description) appear only as background or methods detail and are not load-bearing for the present decoding result. Concerns about serial dependence, day-level medication/edema confounds, and the minimal n=6 Wilcoxon P are statistical validity/generalizability issues, not circularity: they do not make the 'prediction' equivalent to its inputs by construction.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

The analysis rests on several hand-chosen thresholds and domain assumptions, most importantly the definition of a fluctuating report, the median-split labels, and the independence assumption. No new physical entities are postulated.

free parameters (5)
  • class imbalance exclusion threshold = 60%
    Reports in which any binary class exceeds 60% of observations are excluded as 'non-fluctuation'. This threshold determines which 8 of 31 reports are analyzed, and a different threshold would change the set of reports and results.
  • minimum observations per class = 10
    Only reports with at least 10 observations per class after median split are retained for random forest training; this further reduces the analyzed set.
  • canonical frequency band boundaries = delta 1-4, theta 5-8, alpha 9-12, low-beta 13-20, high-beta 21-30, low-gamma 31-60, high-gamma 61-90 Hz
    PSD features are averages over these hand-defined bands; the reported band-side contributions (gamma/beta contralateral, delta/theta ipsilateral) depend directly on these boundaries.
  • binary label threshold = per-report median VAS
    Each report's VAS ratings are split at the series median into higher/lower pain classes; the labels are therefore defined from the same data used to score predictions, and the classes are relative rather than absolute pain states.
  • PCA first principal component = PC1 per side
    Signals from three bipolar electrode pairs per side are reduced to the first principal component before spectral analysis; this discards potential electrode-specific contributions and influences the band PSD features.
assumptions (5)
  • domain assumption Patient VAS ratings are a valid measure of endogenous pain intensity.
    The entire label structure is built on self-reported VAS scores collected up to four times per day; if these ratings reflect mood, fatigue, or demand characteristics rather than pain, the model is not decoding pain. Location: Methods, Self-reported symptoms.
  • domain assumption STN LFP recorded during resting, stimulation-OFF periods reflects pain-relevant neural activity.
    The BrainSense Survey recordings last about 20 seconds with stimulation OFF; movement artifacts, peri-lead edema, and medication state could confound the signal. Location: Methods, Wireless STN signal recording; Discussion limitations.
  • domain assumption Repeated observations within a patient are statistically independent for permutation and Wilcoxon tests.
    Up to four sessions per day over a short hospitalization are treated as independent samples; temporal autocorrelation would inflate significance. The paper checks time-of-day quartile balance but not serial dependence. Location: Methods, Statistics.
  • ad hoc to paper The modified PDRP criteria correctly classify PD-related versus non-PD-related pain.
    A six-item clinical algorithm (Fig. 2 and Table S1) is introduced for this study; misclassification could explain why only PDRP reports decode significantly. Location: Methods, Pain annotation and categorization.
  • domain assumption Gini importance from the random forest reflects the neural contribution to pain.
    Feature importance measures are descriptive and can be biased by correlated features; the paper interprets them as biomarkers without a causal or statistical test. Location: Methods, Gini importance.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Wirelessly transmitted subthalamic nucleus signals predict endogenous pain levels in Parkinson's disease patients." pith.science (2026). https://pith.science/paper/VOVRARDD

@misc{pith2026250621439,
  author       = {Pith},
  title        = {Pith review of: Wirelessly transmitted subthalamic nucleus signals predict endogenous pain levels in Parkinson's disease patients},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VOVRARDD}},
  note         = {Machine review of arXiv:2506.21439}
}
read the original abstract

Parkinson disease (PD) patients experience pain fluctuations that significantly reduce their quality of life. Despite the vast knowledge of the subthalamic nucleus (STN) role in PD, the STN biomarkers for pain fluctuations and the relationship between bilateral subthalamic nucleus (STN) activities and pain occurrence are still less understood. This observational study used data-driven methods by collecting annotated pain followed by a series of corresponding binary pain ratings and wirelessly transmitted STN signals, then leveraging the explainable machine learning algorithm to predict binary pain levels and sort the feature influence. The binary pain levels could be predicted among annotated pain reports corresponding to PD-related pain characteristics. The STN activity from both sides could impact pain prediction, with gamma and beta bands in the contralateral STN and delta and theta bands in the ipsilateral STN showing a prominent role. This study emphasizes the role of bilateral STN biomarkers on endogenous pain fluctuations.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

107 extracted references · 76 canonical work pages

  1. [1]

    & Deuschl, G

    Wasner, G. & Deuschl, G. Pains in Parkinson disease—many syndromes under one umbrella. Nat Rev Neurol 8, 284–294 (2012)

    2012

  2. [2]

    & Brefel-Courbon, C

    Marques, A. & Brefel-Courbon, C. Chronic pain in Parkinson’s disease: Clinical and pathophysiological aspects. Revue Neurologique 177, 394–399 (2021)

    2021

  3. [3]

    J., Brown, P

    Mostofi, A., Morgante, F., Edwards, M. J., Brown, P. & Pereira, E. A. C. Pain in Parkinson’s disease and the role of the subthalamic nucleus. Brain 144, 1342–1350 (2021)

    2021

  4. [4]

    Mylius, V. et al. Pain in Parkinson’s Disease: Current Concepts and a New Diagnostic Algorithm. Mov Disord Clin Pract 2, 357–364 (2015)

    2015

  5. [5]

    A., Walker, R

    Lee, M. A., Walker, R. W., Hildreth, T. J. & Prentice, W. M. A Survey of Pain in Idiopathic Parkinson’s Disease. Journal of Pain and Symptom Management 32, 462–469 (2006). 38

    2006

  6. [6]

    Broen, M. P. G., Braaksma, M. M., Patijn, J. & Weber, W. E. J. Prevalence of pain in Parkinson’s disease: A systematic review using the modified QUADAS tool. Movement Disorders 27, 480–484 (2012)

    2012

  7. [7]

    Kurihara, K. et al. Fluctuating pain in Parkinson’s disease: Its prevalence and impact on quality of life. eNeurologicalSci 25, 100371 (2021)

    2021

  8. [8]

    Quinn, N. P. Classification of fluctuations in patients with Parkinson’s disease. Neurology 51, (1998)

    1998

Show all 107 references
  1. [9]

    Samura, K. et al. Intractable facial pain in advanced Parkinson’s disease alleviated by subthalamic nucleus stimulation. Journal of Neurology, Neurosurgery & Psychiatry 79, 1410–1411 (2008)

    2008

  2. [10]

    Storch, A. et al. Pain Fluctuations in Parkinson’s Disease and Their Association with Motor and Non- Motor Fluctuations. JPD 1–18 (2024) doi:10.3233/JPD-240026

    2024 doi

  3. [11]

    Storch, A. et al. Nonmotor fluctuations in Parkinson disease: Severity and correlation with motor complications. Neurology 80, 800–809 (2013)

    2013

  4. [12]

    & Becker, S

    Desch, S., Schweinhardt, P., Seymour, B., Flor, H. & Becker, S. Evidence for dopaminergic involvement in endogenous modulation of pain relief. eLife 12, e81436 (2023)

    2023

  5. [13]

    Bannister, K., Bee, L. A. & Dickenson, A. H. Preclinical and Early Clinical Investigations Related to Monoaminergic Pain Modulation. Neurotherapeutics 6, 703–712 (2009)

    2009

  6. [14]

    Giannicola, G. et al. The effects of levodopa and ongoing deep brain stimulation on subthalamic beta oscillations in Parkinson’s disease. Experimental Neurology 226, 120–127 (2010)

    2010

  7. [15]

    Mathiopoulou, V. et al. Modulation of subthalamic beta oscillations by movement, dopamine, and deep brain stimulation in Parkinson’s disease. npj Parkinsons Dis. 10, 1–7 (2024)

    2024

  8. [16]

    Lofredi, R. et al. Dopamine-dependent scaling of subthalamic gamma bursts with movement velocity in patients with Parkinson’s disease. eLife 7, e31895 (2018)

    2018

  9. [17]

    Sand, D. et al. Machine learning-based personalized subthalamic biomarkers predict ON-OFF levodopa states in Parkinson patients. J. Neural Eng. 18, 046058 (2021). 39

    2021

  10. [18]

    & Summers, R

    Morelli, N. & Summers, R. L. S. Association of subthalamic beta frequency sub-bands to symptom severity in patients with Parkinson’s disease: A systematic review. Parkinsonism & Related Disorders 110, 105364 (2023)

    2023

  11. [19]

    Belasen, A. et al. The Effects of Mechanical and Thermal Stimuli on Local Field Potentials and Single Unit Activity in Parkinson’s Disease Patients. Neuromodulation: Technology at the Neural Interface 19, 698–707 (2016)

    2016

  12. [20]

    Parker, T. et al. Pain-Induced Beta Activity in the Subthalamic Nucleus of Parkinson’s Disease. Stereotact Funct Neurosurg 98, 193–199 (2020)

    2020

  13. [21]

    Diao, Y. et al. A Meta-Analysis of the Effect of Subthalamic Nucleus-Deep Brain Stimulation in Parkinson’s Disease-Related Pain. Frontiers in Human Neuroscience 15, (2021)

    2021

  14. [22]

    & Brown, P

    Little, S. & Brown, P. Debugging Adaptive Deep Brain Stimulation for Parkinson’s Disease. Movement Disorders 35, 555–561 (2020)

    2020

  15. [23]

    Device profile of the percept PC deep brain stimulation system for the treatment of Parkinson’s disease and related disorders

    Jimenez-Shahed, J. Device profile of the percept PC deep brain stimulation system for the treatment of Parkinson’s disease and related disorders. Expert Review of Medical Devices 18, 319–332 (2021)

    2021

  16. [24]

    R., Mendez, I

    Constantoyannis, C., Berk, C., Honey, C. R., Mendez, I. & Brownstone, R. M. Reducing Hardware- Related Complications of Deep Brain Stimulation. Canadian Journal of Neurological Sciences 32, 194–200 (2005)

    2005

  17. [25]

    Feldmann, L. K. et al. Toward therapeutic electrophysiology: beta-band suppression as a biomarker in chronic local field potential recordings. npj Parkinsons Dis. 8, 44 (2022)

    2022

  18. [26]

    Shirvalkar, P. et al. First-in-human prediction of chronic pain state using intracranial neural biomarkers. Nat Neurosci 26, 1090–1099 (2023)

    2023

  19. [27]

    Naeini, E. K. et al. Pain Recognition With Electrocardiographic Features in Postoperative Patients: Method Validation Study. Journal of Medical Internet Research 23, e25079 (2021). 40

    2021

  20. [28]

    A., Lanctin, D., Ince, N

    Thompson, J. A., Lanctin, D., Ince, N. F. & Abosch, A. Clinical Implications of Local Field Potentials for Understanding and Treating Movement Disorders. Stereotact Funct Neurosurg 92, 251–263 (2014)

    2014

  21. [29]

    Archer, K. J. & Kimes, R. V. Empirical characterization of random forest variable importance measures. Computational Statistics & Data Analysis 52, 2249–2260 (2008)

    2008

  22. [30]

    & Oneto, L

    Chicco, D. & Oneto, L. Computational intelligence identifies alkaline phosphatase (ALP), alpha- fetoprotein (AFP), and hemoglobin levels as most predictive survival factors for hepatocellular carcinoma. Health Informatics J 27, 1460458220984205 (2021)

    2021

  23. [31]

    Mayr, A. et al. Patients with chronic pain exhibit individually unique cortical signatures of pain encoding. Human Brain Mapping 43, 1676–1693 (2022)

    2022

  24. [32]

    & Dunbrack Jr, R

    Wei, Q. & Dunbrack Jr, R. L. The role of balanced training and testing data sets for binary classifiers in bioinformatics. PloS one 8, e67863 (2013)

    2013

  25. [34]

    Benerradi, J., Clos, J., Landowska, A., Valstar, M. F. & Wilson, M. L. Benchmarking framework for machine learning classification from fNIRS data. Frontiers in Neuroergonomics 4, (2023)

    2023

  26. [35]

    Sürücü, O., Baumann-Vogel, H., Uhl, M., Imbach, L. L. & Baumann, C. R. Subthalamic deep brain stimulation versus best medical therapy for l-dopa responsive pain in Parkinson’s disease. Pain 154, 1477–1479 (2013)

    2013

  27. [36]

    Cury, R. G. et al. Subthalamic deep brain stimulation modulates conscious perception of sensory function in Parkinson’s disease. Pain 157, 2758–2765 (2016)

    2016

  28. [37]

    Marques, A. et al. Central pain modulation after subthalamic nucleus stimulation: A crossover randomized trial. Neurology 81, 633–640 (2013)

    2013

  29. [38]

    & Okuyucu, E

    Dogru Huzmeli, E., Yilmaz, A. & Okuyucu, E. Analysis of the effects of subthalamic nucleus deep brain stimulation on somatosensation in Parkinson’s disease patients. Neurol Sci 41, 925–931 (2020). 41

    2020

  30. [39]

    Redding, A. et al. Medically unexplained pain and suicidal ideation among US adults. Journal of Affective Disorders 351, 425–429 (2024)

    2024

  31. [40]

    Somatotopic Organization of Human Secondary Somatosensory Cortex

    Ruben, J. Somatotopic Organization of Human Secondary Somatosensory Cortex. Cerebral Cortex 11, 463–473 (2001)

    2001

  32. [41]

    Shreve, L. A. et al. Subthalamic oscillations and phase amplitude coupling are greater in the more affected hemisphere in Parkinson’s disease. Clinical Neurophysiology 128, 128–137 (2017)

    2017

  33. [42]

    & Betancourt, J

    Tajerian, M., Amrami, M. & Betancourt, J. M. Is there hemispheric specialization in the chronic pain brain? Experimental Neurology 355, 114137 (2022)

    2022

  34. [43]

    & Hebb, A

    Darvas, F. & Hebb, A. O. Task specific inter-hemispheric coupling in human subthalamic nuclei. Front. Hum. Neurosci. 8, (2014)

    2014

  35. [44]

    Arnold Anteraper, S. et al. Resting-State Functional Connectivity of the Subthalamic Nucleus to Limbic, Associative, and Motor Networks. Brain Connectivity 8, 22–32 (2018)

    2018

  36. [45]

    Prenassi, M. et al. Peri-lead edema and local field potential correlation in post-surgery subthalamic nucleus deep brain stimulation patients. Front. Hum. Neurosci. 16, (2022)

    2022

  37. [46]

    von Elm, E. et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. The Lancet 370, 1453–1457 (2007)

    2007

  38. [47]

    Maruo T. et al. Translation and reliability and validity of a Japanese version of the revised Short-Form McGill Pain Questionnaire (SF-MPQ-2). PAIN RESEARCH 28, 43–53 (2013)

    2013

  39. [48]

    Okuma, Y. et al. Fatigue in Japanese patients with Parkinson’s disease: A study using Parkinson fatigue scale. Movement Disorders 24, 1977–1983 (2009)

    2009

  40. [49]

    & Kishima, H

    Tani, N., Yaegaki, T. & Kishima, H. A Case Report: Hemorrhagic Venous Infarction after Deep Brain Stimulation Surgery Probably Due to Coagulation of Intradural Veins. NMC Case Report Journal 8, 315–318 (2021). 42

    2021

  41. [50]

    Mulders, A. E. P. et al. Usability of the Experience Sampling Method in Parkinson’s Disease on a Group and Individual Level. Movement Disorders 35, 1145–1152 (2020)

    2020

  42. [51]

    Jason, L. A. et al. Contrasting Case Definitions for Chronic Fatigue Syndrome, Myalgic Encephalomyelitis/Chronic Fatigue Syndrome and Myalgic Encephalomyelitis. Eval Health Prof 35, 280–304 (2012)

    2012

  43. [52]

    Interpretation of visual analog scale ratings and change scores: a reanalysis of two clinical trials of postoperative pain

    Jensen, M. Interpretation of visual analog scale ratings and change scores: a reanalysis of two clinical trials of postoperative pain. The Journal of Pain 4, 407–414 (2003)

    2003

  44. [53]

    Does the Clinically Significant Difference in Visual Analog Scale Pain Scores Vary with Gender, Age, or Cause of Pain? Academic Emergency Medicine 5, 1086–1090 (1998)

    Kelly, A. Does the Clinically Significant Difference in Visual Analog Scale Pain Scores Vary with Gender, Age, or Cause of Pain? Academic Emergency Medicine 5, 1086–1090 (1998)

    1998

  45. [54]

    & Drake, F

    Van Rossum, G. & Drake, F. Python 3 Reference Manual. Scotts Valley, CA: CreateSpace; 2009. (2020)

    2020

  46. [55]

    Data Structures for Statistical Computing in Python

    McKinney, W. Data Structures for Statistical Computing in Python. in 56–61 (Austin, Texas, 2010). doi:10.25080/Majora-92bf1922-00a

    2010 doi

  47. [56]

    Harris, C. R. et al. Array programming with NumPy. Nature 585, 357–362 (2020)

    2020

  48. [57]

    Virtanen, P. et al. SciPy 1.0: fundamental algorithms for scientific computing in Python. Nat Methods 17, 261–272 (2020)

    2020

  49. [58]

    Pingouin: statistics in Python

    Vallat, R. Pingouin: statistics in Python. Journal of Open Source Software 3, 1026 (2018)

    2018

  50. [59]

    & Perktold, J

    Seabold, S. & Perktold, J. Statsmodels: Econometric and Statistical Modeling with Python. in 92–96 (Austin, Texas, 2010). doi:10.25080/Majora-92bf1922-011

    2010 doi

  51. [60]

    MLxtend: Providing machine learning and data science utilities and extensions to Python’s scientific computing stack

    Raschka, S. MLxtend: Providing machine learning and data science utilities and extensions to Python’s scientific computing stack. Journal of Open Source Software 3, 638 (2018)

    2018

  52. [61]

    Pedregosa, F. et al. Scikit-learn: Machine learning in Python. the Journal of machine Learning research 12, 2825–2830 (2011)

    2011

  53. [62]

    Lundberg, S. M. & Lee, S.-I. A Unified Approach to Interpreting Model Predictions. in Advances in Neural Information Processing Systems vol. 30 (Curran Associates, Inc., 2017). 43

    2017

  54. [63]

    Hunter, J. D. Matplotlib: A 2D Graphics Environment. Computing in Science & Engineering 9, 90–95 (2007)

    2007

  55. [64]

    Waskom, M. L. seaborn: statistical data visualization. Journal of Open Source Software 6, 3021 (2021)

    2021

  56. [65]

    & Ahmmed, R

    Asadur Rahman, Md., Foisal Hossain, Md., Hossain, M. & Ahmmed, R. Employing PCA and t- statistical approach for feature extraction and classification of emotion from multichannel EEG signal. Egyptian Informatics Journal 21, 23–35 (2020)

    2020

  57. [66]

    Mecikalski, J. R. et al. A Random-Forest Model to Assess Predictor Importance and Nowcast Severe Storms Using High-Resolution Radar–GOES Satellite–Lightning Observations. Monthly Weather Review 149, 1725–1746 (2021)

    2021

  58. [67]

    Unbiased variable importance for random forests

    Loecher, M. Unbiased variable importance for random forests. Communications in Statistics - Theory and Methods 51, 1413–1425 (2022)

    2022

  59. [68]

    & Jauhiainen, S

    Saarela, M. & Jauhiainen, S. Comparison of feature importance measures as explanations for classification models. SN Appl. Sci. 3, 272 (2021)

    2021

  60. [69]

    Ward, I. R. et al. Explainable artificial intelligence for pharmacovigilance: What features are important when predicting adverse outcomes? Computer Methods and Programs in Biomedicine 212, 106415 (2021)

    2021

  61. [70]

    Mitra, J. et al. Statistical machine learning to identify traumatic brain injury (TBI) from structural disconnections of white matter networks. NeuroImage 129, 247–259 (2016)

    2016

  62. [71]

    Hanuschkin, A. et al. Investigation of cycle-to-cycle variations in a spark-ignition engine based on a machine learning analysis of the early flame kernel. Proceedings of the Combustion Institute 38, 5751–5759 (2021)

    2021

  63. [72]

    Lundberg, S. M. et al. From local explanations to global understanding with explainable AI for trees. Nat Mach Intell 2, 56–67 (2020). 44

    2020

  64. [73]

    Obaido, G. et al. An Interpretable Machine Learning Approach for Hepatitis B Diagnosis. Applied Sciences 12, 11127 (2022)

    2022

  65. [74]

    Model Evaluation, Model Selection, and Algorithm Selection in Machine Learning

    Raschka, S. Model Evaluation, Model Selection, and Algorithm Selection in Machine Learning. Preprint at http://arxiv.org/abs/1811.12808 (2020)

    2020 arXiv

  66. [75]

    & Fery, A

    Schletz, D., Breidung, M. & Fery, A. Validating and Utilizing Machine Learning Methods to Investigate the Impacts of Synthesis Parameters in Gold Nanoparticle Synthesis. J. Phys. Chem. C 127, 1117– 1125 (2023)

    2023

  67. [76]

    R., Chen, K.-F., Lin, Y

    Martha, S. R., Chen, K.-F., Lin, Y. & Thompson, H. J. Plasma Phospholipid Metabolites Associate With Functional Outcomes Following Mild Traumatic Brain Injury in Older Adults. Biological Research For Nursing 23, 127–135 (2021)

    2021

  68. [77]

    fatigue” and the

    Vieira, S., Garcia-Dias, R. & Lopez Pinaya, W. H. Chapter 19 - A step-by-step tutorial on how to build a machine learning model. in Machine Learning (eds. Mechelli, A. & Vieira, S.) 343–370 (Academic Press, Cambridge, United Kingdom, 2020). doi:10.1016/B978-0-12-815739-8.00019...

    2020 doi

  69. [78]

    Feldmann, L. K. et al. Toward therapeutic electrophysiology: beta-band suppression as a biomarker in chronic local field potential recordings. Npj Park. Dis. 8, 44 (2022)

    2022

  70. [79]

    Mamun, K. A. et al. Movement decoding using neural synchronization and inter-hemispheric connectivity from deep brain local field potentials. J. Neural Eng. 12, 056011 (2015)

    2015

  71. [80]

    Thompson, J. A. et al. Sleep patterns in Parkinson’s disease: direct recordings from the subthalamic nucleus. J. Neurol. Neurosurg. Psychiatry 89, 95–104 (2018)

    2018

  72. [81]

    Verma, A. K. et al. Parkinsonian daytime sleep-wake classification using deep brain stimulation lead recordings. Neurobiol. Dis. 176, 105963 (2023)

    2023

  73. [82]

    Sand, D. et al. Machine learning-based personalized subthalamic biomarkers predict ON-OFF levodopa states in Parkinson patients. J. Neural Eng. 18, 046058 (2021)

    2021

  74. [83]

    & Durso, R

    McNamara, P., Stavitsky, K., Harris, E., Szent-Imrey, O. & Durso, R. Mood, side of motor symptom onset and pain complaints in Parkinson’s disease. Int. J. Geriatr. Psychiatry 25, 519–524 (2010)

    2010

  75. [84]

    Berardelli, A. et al. Pathophysiology of pain and fatigue in Parkinson’s disease. Parkinsonism Relat. Disord. 18, S226–S228 (2012)

    2012

  76. [85]

    & Brundin, L

    Hagell, P. & Brundin, L. Towards an understanding of fatigue in Parkinson disease. J. Neurol. Neurosurg. Psychiatry 80, 489–492 (2009)

    2009

  77. [86]

    Cattaneo, C. et al. Long-Term Effects of Safinamide on Mood Fluctuations in Parkinson’s Disease. J. Park. Dis. 7, 629–634 (2017)

    2017

  78. [87]

    Pauletti, C. et al. Fatigue in fluctuating Parkinson’s disease patients: possible impact of safinamide. J. Neural Transm. 130, 915–923 (2023)

    2023

  79. [88]

    Jasti, S., Dudley, W. N. & Goldwater, E. SAS Macros for Testing Statistical Mediation in Data With Binary Mediators or Outcomes. Nurs. Res. 57, 118–122 (2008). 83

    2008

  80. [89]

    Jacobson, C. J. et al. Disclosure and Self-Report of Emotional, Social, and Physical Health in Children and Adolescents With Chronic Pain—A Qualitative Study of PROMIS Pediatric Measures. J. Pediatr. Psychol. 38, 82–93 (2013)

    2013

  81. [90]

    & Bushnell, M

    Villemure, C. & Bushnell, M. C. Mood Influences Supraspinal Pain Processing Separately from Attention. J. Neurosci. 29, 705–715 (2009)

    2009

  82. [91]

    Rana, A. Q. et al. Association of restless legs syndrome, pain, and mood disorders in parkinson’s disease. Int. J. Neurosci. 126, 116–120 (2016)

    2016

  83. [92]

    Lim, S.-Y. et al. Do dyskinesia and pain share common pathophysiological mechanisms in Parkinson’s disease? Mov. Disord. 23, 1689–1695 (2008)

    2008

  84. [93]

    Pain in Parkinson’s disease

    Ford, B. Pain in Parkinson’s disease. Mov. Disord. 25, S98–S103 (2010)

    2010

  85. [94]

    Tinazzi, M. et al. Pain and motor complications in Parkinson’s disease. J. Neurol. Neurosurg. Psychiatry 77, 822–825 (2006)

    2006

  86. [95]

    Yin, Z. et al. Local field potentials in Parkinson’s disease: A frequency-based review. Neurobiol. Dis. 155, 105372 (2021)

    2021

  87. [96]

    Valkovic, P. et al. Pain in Parkinson´s Disease: A Cross-Sectional Study of Its Prevalence, Types, and Relationship to Depression and Quality of Life. PLOS ONE 10, e0136541 (2015)

    2015

  88. [97]

    & Steinbach, J

    Broetz, D., Eichner, M., Gasser, T., Weller, M. & Steinbach, J. P. Radicular and nonradicular back pain in Parkinson’s disease: A controlled study. Mov. Disord. 22, 853–856 (2007)

    2007

  89. [98]

    Zambito Marsala, S. et al. Spontaneous pain, pain threshold, and pain tolerance in Parkinson’s disease. J. Neurol. 258, 627–633 (2011)

    2011

  90. [99]

    M., Figueiras-Mendez, R., Riva-Meana, C

    Magariños-Ascone, C. M., Figueiras-Mendez, R., Riva-Meana, C. & Córdoba-Fernández, A. Subthalamic neuron activity related to tremor and movement in Parkinson’s disease. Eur. J. Neurosci. 12, 2597–2607 (2000). 84

    2000

  91. [100]

    & Oshima, T

    Narabayashi, Y. & Oshima, T. Analysis of parkinsonian tremor and rigidity relying on peripheral and central intensifying mechanisms through thalamic activities. Neurol. Clin. Neurosci. 2, 28–37 (2014)

    2014

  92. [101]

    A., Walker, R

    Lee, M. A., Walker, R. W., Hildreth, T. J. & Prentice, W. M. A Survey of Pain in Idiopathic Parkinson’s Disease. J. Pain Symptom Manage. 32, 462–469 (2006)

    2006

  93. [102]

    C., Bohlhalter, S., Ciampi de Andrade, D

    Mylius, V., Möller, J. C., Bohlhalter, S., Ciampi de Andrade, D. & Perez Lloret, S. Diagnosis and Management of Pain in Parkinson’s Disease: A New Approach. Drugs Aging 38, 559–577 (2021)

    2021

  94. [103]

    & Rascol, O

    Nègre-Pagès, L., Regragui, W., Bouhassira, D., Grandjean, H. & Rascol, O. Chronic pain in Parkinson’s disease: The cross-sectional French DoPaMiP survey. Mov. Disord. 23, 1361–1369 (2008)

    2008

  95. [104]

    Mylius, V. et al. Pain in Parkinson’s Disease: Current Concepts and a New Diagnostic Algorithm. Mov. Disord. Clin. Pract. 2, 357–364 (2015)

    2015

  96. [105]

    Bruno, A. E. & Sethares, K. A. Fatigue in Parkinson Disease: An Integrative Review. J. Neurosci. Nurs. 47, 146–153 (2015)

    2015

  97. [106]

    Lin, I. et al. Triggers and alleviating factors for fatigue in Parkinson’s disease. PLOS ONE 16, e0245285 (2021)

    2021

  98. [107]

    Zarotti, N. et al. Psychological interventions for people with Parkinson’s disease in the early 2020s: Where do we stand? Psychol. Psychother. Theory Res. Pract. 94, 760–797 (2021)

    2021

  99. [108]

    B., Tait, R

    Margolis, R. B., Tait, R. C. & Krause, S. J. A rating system for use with patient pain drawings: Pain 24, 57–65 (1986)

    1986

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.