Pith. sign in

REVIEW 5 major objections 5 minor 26 references

Multimodal Cardiovascular Risk Profiling Using Self-Supervised Learning of Polysomnography

T0 review · 5 major / 5 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 Projection scores from self-supervised PSG embeddings add predictive value to the Framingham Risk Score for cardiovascular outcomes.

desk verdict Reasonable first application of SSL PSG embeddings to CVD risk, with a genuinely external cohort, but the external validation has a provenance gap and the claimed gains lack confidence intervals. read the letter →

arxiv 2507.09009 v1 pith:DUBOXJAD submitted 2025-07-11 cs.LG cs.AI

Zhengxiao He , Huayu Li , Geng Yuan , William D.S. Killgore , Stuart F. Quan , Chen X. Chen , Ao Li This is my paper
classification cs.LGcs.AI
keywords self-supervisedlearningcardiovasculardiseaseriskpolysomnographylatentrepresentationvectorprojectionEEGECGrespiratorysignals
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 argues that an overnight polysomnography recording contains cardiovascular risk information that standard summary metrics and the Framingham Risk Score leave untapped. The authors train a self-supervised residual-transformer model on raw EEG, ECG, and respiratory signals, then derive a disease vector for each outcome as the difference between average embeddings of affected and unaffected people. Each person's projection score onto that vector measures how closely their sleep physiology aligns with the disease phenotype. Combining these projection scores with the Framingham Risk Score improved discrimination, with AUCs from 0.607 to 0.965 across outcomes, and the improvement persisted in an independent external cohort. If the finding holds, routine sleep studies could double as individualized cardiovascular risk profiles.

What carries the argument

The machinery is the disease-vector projection. After self-supervised pretraining that reconstructs randomly masked 30-second PSG segments under a total-coding-rate penalty, the model maps each modality (EEG, ECG, respiratory) to a 256-dimensional embedding. For each outcome, the disease vector is defined as $\vec{v}_{\mathrm{disease}} = \mu_{\mathrm{positive}} - \mu_{\mathrm{negative}}$, the normalized difference between group centroids; a subject's projection score is the normalized dot product of their segment embeddings with this vector, averaged over their top three segments. This turns an unsupervised representation into one interpretable scalar per modality and outcome, which logistic regression can then combine with age, sex, BMI, and the Framingham Risk Score. The same vectors are applied unchanged to the external cohort, which is why the transferability of $\vec{v}_{\mathrm{disease}}$ is the load-bearing step.

What would settle it

Compute disease vectors from the external cohort's own embeddings and compare the resulting projection-score AUCs with those obtained using the training cohort's vectors; if the transfer assumption holds the two sets of AUCs should be close, whereas a large drop or a reversal in score distributions would show the external results depend on domain-specific artifacts. Alternatively, permute outcome labels when estimating the disease vector: if projection scores derived from permuted centroids achieve similar AUCs to the reported ones, the signal is not disease-specific.

Watch

Extended reading notes

Core claim

The paper's central discovery is that a single direction in a learned embedding space, namely the vector between average representations of people with and without a given outcome, carries enough disease signal to predict both prevalent and incident cardiovascular conditions from raw sleep recordings. ECG-derived projections discriminate right bundle branch block (AUC 0.997) and atrial fibrillation (AUC 0.961) nearly on their own; EEG-derived projections associate with hypertension and CVD mortality; respiratory projections add incremental value when combined with ECG. Adding all modality projection scores to the Framingham Risk Score consistently improved discrimination in the held-out test set (AUCs 0.607–0.965), and the additive benefit persisted in an independent external cohort, where combined models reached AUCs of 0.753 for incident CVD and 0.807 for coronary artery disease.

Load-bearing premise

The load-bearing premise is that the disease vector computed from the training cohort's embeddings, the average difference between people with and without each outcome, points in a direction that generalizes to an independent cohort recorded with different equipment and in a different population, without any adaptation.

Editorial extensions

If this is right

  • ECG-derived projection scores can act as near-standalone markers for rhythm and conduction abnormalities, with AUCs above 0.96 for atrial fibrillation and right bundle branch block.
  • EEG-derived scores carry independent risk information for hypertension and cardiovascular mortality, supporting EEG as a digital biomarker for vascular risk.
  • Respiratory signals, weak alone, add complementary value when combined with ECG, improving discrimination of congestive heart failure and myocardial infarction.
  • Adding three-modality projection scores to the Framingham Risk Score, rather than replacing it, gives the best or near-best AUC for most outcomes in both internal and external cohorts.
  • Because the encoder needs no manual sleep staging, the method can produce risk scores from raw PSG signals at an inference cost of under five seconds per patient.

Reading between the lines

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

  • Beyond the paper, the same disease-vector construction could be applied to other outcomes encoded in sleep physiology, such as diabetes or cognitive decline, since the pretraining phase needs no labels.
  • Beyond the paper, a testable extension is to recompute disease vectors from the external cohort's embeddings and compare per-outcome AUCs with the transferred vectors; close agreement would support the vectors as biomarkers, while divergence would implicate acquisition-protocol artifacts.
  • Beyond the paper, if the vectors are stable across cohorts, PSG-derived projection scores could be issued on every routine sleep study as a low-cost, repeatable cardiovascular risk screen without changing the clinical protocol.
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

5 major / 5 minor

Summary. The paper proposes a self-supervised Residual-Transformer model that encodes 30-second EEG, ECG, and respiratory PSG segments into a shared 256-dimensional latent space, trained with masked-segment reconstruction and a total coding rate penalty. For each cardiovascular outcome, a disease vector is computed as the difference between the mean embeddings of training-set cases and controls (Eq. 5), and each subject receives per-modality projection scores obtained by averaging the top three segment-level dot products onto the disease vector. These scores are then entered into logistic regression models, alone or combined with the Framingham Risk Score, and evaluated on a held-out SHHS test set and an external Wisconsin Sleep Cohort (WSC) sample. The central claims are that the projection scores are interpretable, capture complementary risk information, and that adding them to the FRS consistently improves AUC across prevalent and incident CVD outcomes in both internal and external cohorts.

Significance. If the results hold, the framework would be a useful step toward leveraging routine PSG data for cardiovascular risk stratification without requiring manual sleep staging, and the interpretable projection scores could aid clinical adoption. The study uses two community-based cohorts, reports multimodal comparisons, and makes source code available, which are genuine strengths. However, the current support is weakened by the very small number of events in the internal test set, the absence of confidence intervals for the AUC comparisons, the unexplained provenance of the disease vectors in the external cohort, and an unexamined selection statistic (top-three segments). These issues are load-bearing for the abstract's claim of robust external validation and consistent incremental value.

major comments (5)
  1. [§2.2.2 and Table 4] The manuscript does not state whether the disease vectors used to score WSC subjects were derived from SHHS training labels or recomputed within WSC. This matters because the WSC outcomes in Table 4 (incident CVD, prevalent CVD, CAD) do not appear in the SHHS outcome set in Table 3, so it is unclear what outcome labels would have been available to construct corresponding SHHS-trained vectors. If any WSC disease vector was derived from WSC labels, the external validation is no longer independent and the abstract's claim of robust external replication is unsupported. Please specify the exact provenance of each WSC vector; if SHHS-derived, describe the label mapping; if WSC-derived, re-run the external analysis with frozen SHHS vectors or present it as a second internal validation.
  2. [§3.3, Tables 3 and 4] No confidence intervals or significance tests are reported for any AUC or for the increments in AUC when projection scores are added to the Framingham Risk Score. This is a serious issue because the SHHS test set contains only 13 AF, 13 CHF, and 11 CVD-mortality cases (Table 1), so the rank-based AUCs are highly unstable; a single reordering of a few subjects can move the AUC by several hundredths. The WSC increments are also small (e.g., hypertension +0.007, CVD +0.012, Table 4) and are within plausible sampling noise. Please provide bootstrap or DeLong confidence intervals for the AUCs and, in particular, for the FRS-composite versus FRS-only differences; otherwise the 'consistently improved predictive performance' claim is not statistically grounded.
  3. [§2.2.3] The subject-level projection score is defined as the average of the top three highest segment-level projections, but this choice is presented without any ablation or justification. Because this statistic is the sole input to the downstream logistic regression, the results could be highly sensitive to the arbitrary choice k=3. The manuscript should report results for alternative choices (e.g., top-1, top-5, mean over all segments) to show that the conclusions are not an artifact of this selection. Additionally, Eq. (6) uses an undefined quantity V_p and a nonstandard product notation; the authors should define all terms and write the projection as a dot product of normalized vectors.
  4. [§2.1, §2.2.2 and §3.3] The external validation assumes that the SHHS-trained disease vector v_disease = μ_positive − μ_negative transfers to the WSC embedding space, but the two cohorts differ in acquisition protocol (in-home versus in-laboratory PSG), age composition, and BMI distribution (Tables 1 and 2), and the disease vector is an unadjusted centroid difference that can encode demographic and protocol confounds rather than disease-specific physiology. No diagnostic is provided for distribution shift between the cohorts, so the WSC projection scores could be predictive through demographic transfer even if the disease-related physiological signal does not transfer. Please include embedding-space distribution comparisons (e.g., MMD or per-modal PCA overlays) and/or a permutation test that evaluates WSC predictions using a randomly reoriented or demographic-only disease vector.
  5. [Abstract, §1, and §2.2.2] The paper repeatedly describes the framework as not relying on labels (e.g., 'without relying on labels' in §1), but Eq. (5) constructs disease vectors directly from the training-set outcome labels. While the embedding itself is self-supervised, the projection scores that are actually evaluated are supervised quantities. This overstates the label-free nature of the pipeline and should be corrected to 'without manual sleep-stage annotations' or 'self-supervised embeddings with supervised disease-vector construction.'
minor comments (5)
  1. [Eq. (6) and surrounding text] The notation 'Signal * V_disease / V_p' is ambiguous; V_p is not defined and the product symbol conflates element-wise and vector operations. Please rewrite using an explicit dot product and define the normalization steps.
  2. [Table 4 footnotes] The footnotes for Table 4 are misnumbered: 'FRS Score1' is followed by footnote 1 that says 'Baseline includes age, sex, and BMI,' and 'Baseline2' is followed by a footnote that says 'FRS Score.' The mapping should be corrected.
  3. [§3.2, Figure 5 caption] The sentence 'individuals in the negative outcome groups consistently demonstrated elevated projection scores compared to controls' is confusing because the negative outcome group and the controls appear to be the same group; please clarify the comparison.
  4. [§2.2.4 and §3.4] The significance threshold is stated as p < 0.05 in Methods but Figure 8 is described with p < 0.005; the two thresholds should be reconciled, and the issue of multiple testing across many modality-outcome pairs should be addressed.
  5. [§2.2.2] The manuscript does not specify whether the centroids μ_positive and μ_negative are computed over all segments pooled from the subjects, or over per-subject averaged embeddings; this affects the interpretation of the disease vector and should be stated.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the disease-vector derivation is evaluated out-of-sample on a held-out test set and an external cohort.

full rationale

The derivation chain is not circular. The SSL encoder is trained without labels on SHHS training/validation data. The disease vector v_disease = mu_positive - mu_negative (Eq. 5) is computed from labeled training embeddings, and projection scores are then computed for held-out SHHS test subjects by projecting their embeddings onto these fixed vectors; logistic-regression combinations with FRS are likewise fitted in the training phase and applied to the held-out test set (Sections 2.2.2-2.2.4, 3.3). This is a standard supervised linear-probe evaluation, not a prediction of the same data used to define the probe. The WSC external test is a separate cohort with different PSG acquisition protocols and demographics, so the AUCs in Table 4 are out-of-sample unless disease vectors for WSC-only outcomes were recomputed using WSC labels; the paper does not state that, and the provenance gap is a correctness/external-validity concern rather than a demonstrated circularity. The only overlapping self-citation (ref. [12], used as architectural motivation) is not load-bearing for the central claim. The domain-shift worry about Eq. 5 transferring to WSC is a validity threat to be tested, not an equivalence of output to input.

Assumptions & free parameters 3 free parameters · 4 assumptions · 2 invented entities

The central claim rests on several hand-chosen parameters (top-3 selection, 256-dim embeddings, mask count), a linear separability assumption for disease directions, and an assumption of cross-cohort transfer without adaptation. The projection score and disease vector are derived constructs without independent biological validation.

free parameters (3)
  • top_k_segments = 3
    Subject-level scores average the top three highest segment-level projections (Section 2.2.3). This value is chosen by hand without ablation and is a data-dependent summary statistic.
  • embedding_dim = 256
    The encoder maps 30-second segments to 256-dimensional latent representations (Section 2.2.1). This architectural choice is not varied or justified.
  • mask_count = 20-30
    Random masking is performed 20-30 times per segment for the similarity loss (Section 2.2.1). The exact value and its effect are not reported.
assumptions (4)
  • domain assumption The SSL encoder produces embeddings in which disease-relevant physiological variation is linearly separable, captured by the centroid difference vector.
    Section 2.2.2 and Eq. 5 define the disease vector as the difference of group centroids, assuming a linear direction separates disease-positive and disease-negative embeddings.
  • domain assumption The SHHS-trained encoder and disease vectors transfer to the WSC cohort without adaptation.
    Section 3.3 applies the framework to WSC, but no domain adaptation or recomputation of disease vectors is described.
  • ad hoc to paper The top-three highest segment projections are a valid subject-level summary of disease alignment.
    Section 2.2.3 selects the top three projections to 'account for the possibility that only a few segments may exhibit disease-relevant patterns', with no ablation study.
  • domain assumption Outcome labels are correctly adjudicated in SHHS and WSC.
    Incident and prevalent outcomes are defined using cohort follow-up adjudication, but the paper relies on these labels without verification.
invented entities (2)
  • projection_score
    purpose: A scalar risk score summarizing a subject's alignment with a disease-specific latent direction, used as input to logistic regression.
    The score is a derived feature, not a measured biomarker. Its only evidence is the paper's own AUC/OR analyses; no external biological validation is provided.
  • disease_vector
    purpose: A latent-space direction representing disease-related variation, used to compute projection scores.
    The vector is a centroid difference computed from training labels; it has no independent falsifiable handle outside the paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Multimodal Cardiovascular Risk Profiling Using Self-Supervised Learning of Polysomnography." pith.science (2026). https://pith.science/paper/DUBOXJAD

@misc{pith2026250709009,
  author       = {Pith},
  title        = {Pith review of: Multimodal Cardiovascular Risk Profiling Using Self-Supervised Learning of Polysomnography},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DUBOXJAD}},
  note         = {Machine review of arXiv:2507.09009}
}
read the original abstract

Methods: We developed a self-supervised deep learning model that extracts meaningful patterns from multi-modal signals (Electroencephalography (EEG), Electrocardiography (ECG), and respiratory signals). The model was trained on data from 4,398 participants. Projection scores were derived by contrasting embeddings from individuals with and without CVD outcomes. External validation was conducted in an independent cohort with 1,093 participants. The source code is available on https://github.com/miraclehetech/sleep-ssl. Results: The projection scores revealed distinct and clinically meaningful patterns across modalities. ECG-derived features were predictive of both prevalent and incident cardiac conditions, particularly CVD mortality. EEG-derived features were predictive of incident hypertension and CVD mortality. Respiratory signals added complementary predictive value. Combining these projection scores with the Framingham Risk Score consistently improved predictive performance, achieving area under the curve values ranging from 0.607 to 0.965 across different outcomes. Findings were robustly replicated and validated in the external testing cohort. Conclusion: Our findings demonstrate that the proposed framework can generate individualized CVD risk scores directly from PSG data. The resulting projection scores have the potential to be integrated into clinical practice, enhancing risk assessment and supporting personalized care.

Figures

Figures reproduced from arXiv: 2507.09009 by the authors.

Figure 3
Figure 3. Visualization of disease vector construction in the embedding space. Each dot represents [PITH_FULL_IMAGE:figures/full_fig_p024_3.png] view at source ↗
Figure 4
Figure 4. Overview of the proposed backbone architecture, comprising residual and transformer [PITH_FULL_IMAGE:figures/full_fig_p024_4.png] view at source ↗
Figure 6
Figure 6. Violin plots of disease-specific projection scores derived from EEG embeddings, stratified by disease status [PITH_FULL_IMAGE:figures/full_fig_p024_6.png] view at source ↗
Figures from the paper (1 more)
Figure 9
Figure 9. Figure 9: Example of a patient-specific cardiovascular risk card based on projection scores. The patient’s ECG-derived scores are compared with population-level averages from the SHHS cohort across multiple disease outcomes, enabling individualized physiological interpretation …

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

26 extracted references · 26 canonical work pages

  1. [1]

    Prospective study of the association between sleep-disordered breathing and hypertension

    Peppard PE, Young T, Palta M, Skatrud J. Prospective study of the association between sleep-disordered breathing and hypertension. N Engl J Med. 2000;342(19):1378-1384

    work page 2000

  2. [2]

    Sleep duration and cardiovascular disease risk: epidemiologic and experimental evidence

    Covassin N, Singh P. Sleep duration and cardiovascular disease risk: epidemiologic and experimental evidence. Sleep Health. 2016;2(2):64-68

    work page 2016

  3. [3]

    Insomnia and incident cardiovascular disease: the Penn State Cohort

    Fernandez-Mendoza J, Vgontzas AN, Liao D, et al. Insomnia and incident cardiovascular disease: the Penn State Cohort. Sleep. 2012;35(6):709-716

    work page 2012

  4. [4]

    Berry RB, Budhiraja R, Gottlieb DJ, et al. Rules for scoring respiratory events in sleep: update of the 2007 AASM manual for the scoring of sleep and associated events: deliberations of the sleep apnea definitions task force of the American Academy of Sle ep Medicine. J Clin Sleep Med. 2012;8(5):597-619

    work page 2007

  5. [5]

    Association of objectively measured sleep characteristics and hypertension: analysis from the Multi -Ethnic Study of Atherosclerosis (MESA)

    Luo J, Lutsey PL, et al. Association of objectively measured sleep characteristics and hypertension: analysis from the Multi -Ethnic Study of Atherosclerosis (MESA). Hypertension. 2023;80(6):1352-1360

    work page 2023

  6. [6]

    Associations of sleep duration, sleep architecture and blood pressure dipping in middle-aged adults

    Zhang X, et al. Associations of sleep duration, sleep architecture and blood pressure dipping in middle-aged adults. Sleep Health. 2022;8(2):166-173

    work page 2022

  7. [7]

    Sleep-disordered breathing and cardiovascular disease: cross- sectional results of the Sleep Heart Health Study

    Redline S, Yenokyan G, et al. Sleep-disordered breathing and cardiovascular disease: cross- sectional results of the Sleep Heart Health Study. Am J Respir Crit Care Med . 2010;182(3):277-283

    work page 2010

  8. [8]

    Slow-wave sleep is associated with incident hypertension: the sleep heart health study

    Javaheri S, Zhao YY , Punjabi NM, Quan SF, Gottlieb DJ, Redline S. Slow-wave sleep is associated with incident hypertension: the sleep heart health study. Sleep. 2018;41(1):zsx179. 22

    work page 2018

Show all 26 references
  1. [9]

    Prediction of Incident Atrial Fibrillation Using Handheld Single -Lead Electrocardiograms from the VITAL -AF Trial

    Khurshid S, Friedman S, Al-Alusi M, et al. Prediction of Incident Atrial Fibrillation Using Handheld Single -Lead Electrocardiograms from the VITAL -AF Trial. Circulation. 2024;150(Suppl_1):A4135754–A4135754

    2024

  2. [10]

    A deep learning digital biomarker to detect hypertension and stratify cardiovascular risk from the electrocardiogram

    Al-Alusi MA, Friedman SF, Kany S, et al. A deep learning digital biomarker to detect hypertension and stratify cardiovascular risk from the electrocardiogram. NPJ Digit Med . 2025;8(1):120

    2025

  3. [11]

    Electrocardiogram -based artificial intelligence to identify prevalent coronary artery disease

    Kany S, Friedman SF, Al-Alusi M, et al. Electrocardiogram -based artificial intelligence to identify prevalent coronary artery disease. Eur Heart J. 2024;45(Supplement_1):ehae666– 1348

    2024

  4. [12]

    MTS -LOF: medical time -series representation learning via occlusion -invariant features

    Li H, Carreon -Rascon AS, Chen X, Yuan G, Li A. MTS -LOF: medical time -series representation learning via occlusion -invariant features. IEEE J Biomed Health Inform . Published online 2024

    2024

  5. [13]

    Self-supervised contrastive learning for medical time series: A systematic review

    Liu Z, Alavi A, Li M, Zhang X. Self-supervised contrastive learning for medical time series: A systematic review. Sensors. 2023;23(9):4221

    2023

  6. [14]

    SleepFM: multi -modal representation learning for sleep across brain activity, ECG and respiratory signals

    Thapa R, He B, Kjær MR, et al. SleepFM: multi -modal representation learning for sleep across brain activity, ECG and respiratory signals. In: Proceedings of the 41st International Conference on Machine Learning. ; 2024:48019-48037

    2024

  7. [15]

    A foundational transformer leveraging full night, multichannel sleep study data accurately classifies sleep stages

    Fox B, Jiang J, Wickramaratne S, et al. A foundational transformer leveraging full night, multichannel sleep study data accurately classifies sleep stages. Sleep. Published online 2025:zsaf061

    2025

  8. [16]

    A novel electroencephalogram-derived measure of disrupted delta wave activity during sleep predicts all-cause mortality risk

    Lechat B, Hansen KL, Melaku YA, et al. A novel electroencephalogram-derived measure of disrupted delta wave activity during sleep predicts all-cause mortality risk. Ann Am Thorac Soc. 2022;19(4):649-658

    2022

  9. [17]

    The hypoxic burden of sleep apnoea predicts cardiovascular disease-related mortality: the Osteoporotic Fractures in Men Study and the Sleep Heart Health Study

    Azarbarzin A, Sands SA, Stone KL, et al. The hypoxic burden of sleep apnoea predicts cardiovascular disease-related mortality: the Osteoporotic Fractures in Men Study and the Sleep Heart Health Study. Eur Heart J. 2019;40(14):1149-1157

    2019

  10. [18]

    Unsupervised deep learning of electrocardiograms enables scalable human disease profiling

    Friedman SF, Khurshid S, Venn RA, et al. Unsupervised deep learning of electrocardiograms enables scalable human disease profiling. NPJ Digit Med. 2025;8(1):23

    2025

  11. [19]

    Burden of sleep apnea: rationale, design, and major findings of the Wisconsin Sleep Cohort study

    Young T, Palta M, Dempsey J, Peppard PE, Nieto FJ, Hla KM. Burden of sleep apnea: rationale, design, and major findings of the Wisconsin Sleep Cohort study. WMJ. 2009;108(5):246

    2009

  12. [20]

    Segmentation of multivariate mixed data via lossy data coding and compression

    Ma Y , Derksen H, Hong W, Wright J. Segmentation of multivariate mixed data via lossy data coding and compression. IEEE Trans Pattern Anal Mach Intell. 2007;29(9):1546-1562

    2007

  13. [21]

    Graham I, Atar D, Borch -Johnsen K, et al. European guidelines on cardiovascular disease prevention in clinical practice: Executive summary - Fourth Joint Task Force of the European Society of Cardiology and Other Societies on Cardiovascular Disease Prevention in Clinical Prac...

    2007 doi

  14. [22]

    General cardiovascular risk profile for use in primary care: The Framingham heart study

    D’Agostino RB, Vasan RS, Pencina MJ, et al. General cardiovascular risk profile for use in primary care: The Framingham heart study. Circulation. 2008;117(6):743 -753. doi:10.1161/CIRCULATIONAHA.107.699579

    2008 doi

  15. [23]

    Understanding brain function in vascular cognitive impairment and dementia with EEG and MEG: A systematic review

    Torres-Simón L, Doval S, Nebreda A, Llinas SJ, Marsh EB, Maestú F. Understanding brain function in vascular cognitive impairment and dementia with EEG and MEG: A systematic review. Neuroimage Clin. 2022;35(May). doi:10.1016/j.nicl.2022.103040 23

    2022

  16. [24]

    Qualitative electroencephalogram and its predictors in the diagnosis of stroke

    Ag Lamat MSN, Abd Rahman MSH, Wan Zaidi WA, et al. Qualitative electroencephalogram and its predictors in the diagnosis of stroke. Front Neurol . 2023;14(June). doi:10.3389/fneur.2023.1118903

    2023

  17. [25]

    Explainable machine learning model based on EEG, ECG, and clinical features for predicting neurological outcomes in cardiac arrest patient

    Niu Y , Chen X, Fan J, et al. Explainable machine learning model based on EEG, ECG, and clinical features for predicting neurological outcomes in cardiac arrest patient. Sci Rep . 2025;15(1):11498. doi:10.1038/s41598-025-93579-0

    2025 doi

  18. [26]

    Baseline

    Zhou S, Song G, Sun H, et al. Continuous sleep depth index annotation with deep learning yields novel digital biomarkers for sleep health. NPJ Digit Med . 2025;8(1). doi:10.1038/s41746-025-01607-0 24 Figure Legends Figure 1: Flowchart of Polysomnographic (PSG) Recordings Selec...

    2025 doi

Pith tools

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