Mental workload / stress · WESAD · Area under ROC
uses extra training data
Leaderboard row
Results explorer
One row per reported result from wearable PPG, ECG, motion and sleep papers. Filter by task, dataset, metric, sensor and year, or show only SOTA rows. Each value links to its source: a paper table or figure, or the leaderboard row it came from.
Last updated
433
reported results
224
with a table or figure reference
92
papers and models
134
task × dataset × metric pairs
38
benchmarks with a SOTA star
433 results
Mental workload / stress · WESAD · exact_match
subject-disjoint
Leaderboard row
NormWear
In NormWear: Toward Foundation Model for Multivariate Wearable Sensing of Physiological Signals
84.381%↑
Multi-task benchmark · NormWear 18-task benchmark (macro avg.) · AUROC (%)
held-out test setwithin-datasetlinear probe, full-shot
TF-C
Baseline in NormWear: Toward Foundation Model for Multivariate Wearable Sensing of Physiological Signals
81.230%↑
Multi-task benchmark · NormWear 18-task benchmark (macro avg.) · AUROC (%)
held-out test setwithin-datasetlinear probe, full-shot
Statistical
Baseline in NormWear: Toward Foundation Model for Multivariate Wearable Sensing of Physiological Signals
80.247%↑
Multi-task benchmark · NormWear 18-task benchmark (macro avg.) · AUROC (%)
held-out test setwithin-datasetlinear probe, full-shot
CLAP
Baseline in NormWear: Toward Foundation Model for Multivariate Wearable Sensing of Physiological Signals
80.103%↑
Multi-task benchmark · NormWear 18-task benchmark (macro avg.) · AUROC (%)
held-out test setwithin-datasetlinear probe, full-shot
Chronos
Baseline in NormWear: Toward Foundation Model for Multivariate Wearable Sensing of Physiological Signals
79.488%↑
Multi-task benchmark · NormWear 18-task benchmark (macro avg.) · AUROC (%)
held-out test setwithin-datasetlinear probe, full-shot
Potassium estimation · mc-med · Pearson correlation (labs / continuous)
uses extra training data
Leaderboard row
Qwen2.5-7B-Instruct + PulsePPG encoder
In PulseLM: A Foundation Dataset and Benchmark for PPG-Text Learning
72.8%↑
PPG question answering · PulseLM (16 datasets, in-domain) · Exact-match accuracy (%)
dataset train/test splitsin-domainLoRA fine-tuning, frozen PPG encoder12 closed-form QA categories
LLaMA3.1-8B + PulsePPG
Baseline in PulseLM: A Foundation Dataset and Benchmark for PPG-Text Learning
70.6%↑
PPG question answering · PulseLM (16 datasets, in-domain) · Exact-match accuracy (%)
dataset train/test splitsin-domainLoRA fine-tuning, frozen PPG encoder12 closed-form QA categories
LLaMA3.2-3B + PulsePPG
Baseline in PulseLM: A Foundation Dataset and Benchmark for PPG-Text Learning
69.9%↑
PPG question answering · PulseLM (16 datasets, in-domain) · Exact-match accuracy (%)
dataset train/test splitsin-domainLoRA fine-tuning, frozen PPG encoder12 closed-form QA categories
Qwen3-4B + PulsePPG
Baseline in PulseLM: A Foundation Dataset and Benchmark for PPG-Text Learning
69.3%↑
PPG question answering · PulseLM (16 datasets, in-domain) · Exact-match accuracy (%)
dataset train/test splitsin-domainLoRA fine-tuning, frozen PPG encoder12 closed-form QA categories
Fusion PPG (PulsePPG, no LLM)
Baseline in PulseLM: A Foundation Dataset and Benchmark for PPG-Text Learning
63.9%↑
PPG question answering · PulseLM (16 datasets, in-domain) · Exact-match accuracy (%)
dataset train/test splitsin-domainLoRA fine-tuning, frozen PPG encoder12 closed-form QA categories
Blind (text only)
Baseline in PulseLM: A Foundation Dataset and Benchmark for PPG-Text Learning
56.7%↑
PPG question answering · PulseLM (16 datasets, in-domain) · Exact-match accuracy (%)
dataset train/test splitsin-domainLoRA fine-tuning, frozen PPG encoder12 closed-form QA categories
Respiratory rate · BIDMC · MAE (breaths/min)
5-fold cross-validationwithin-datasetfine-tuned
Respiratory rate · BIDMC · MAE (breaths/min)
5-fold cross-validationwithin-datasetfine-tuned
0.76 [0.66, 0.82]↑
Sleep apnea ahi · MESA Sleep · AHI estimation R² (fraction of variance)
Protocol not structured
Leaderboard row
2026
0.71 [0.66, 0.76]↑
Sleep apnea severity · MESA Sleep · Macro-averaged F1
Protocol not structured
Leaderboard row
2026
0.977↑
Sleep staging · Cleveland Family Study (CFS) · Area under ROC
REM vs all other epochs (binary, 30-s epochs); not a 5-class staging score
Leaderboard row
71.2↑
Sleep staging · Cleveland Family Study (CFS) · Macro-averaged F1
uses extra training data5 classes (W, N1, N2, N3, REM)
Leaderboard row
Sleep staging · DREAMT · Area under ROC
Protocol not structured
Leaderboard row
2026
Multimodal self-supervised PPG+ACC model
0.30↑
Sleep staging · DREAMT · Cohen's κ
not stated in the abstractwithin-datasetself-supervised pretraining, then fine-tuning4-class
Fully supervised baseline
0.06↑
Sleep staging · DREAMT · Cohen's κ
not stated in the abstractwithin-datasetself-supervised pretraining, then fine-tuning4-class
Sleep staging · DREAMT · Cohen’s κ (e.g. sleep staging)
Protocol not structured
Leaderboard row
2026
0.7425↑
Sleep staging · HMC · Balanced accuracy
subject split 100 / 25 / 26 by sorted subject IDwithin-datasetfrozen features, strict linear probe; mean of 3 pretraining seeds
0.7425↑
Sleep staging · hmc · Balanced accuracy (HAR / class-imbalance)
uses extra training data
Leaderboard row
2026
0.7916↑
Sleep staging · ISRUC (subgroup I) · Balanced accuracy
subject split 1–80 / 81–90 / 91–100 (train / validation / test)within-datasetfrozen features, strict linear probe; mean of 3 pretraining seeds
0.7916↑
Sleep staging · isruc-s1 · Balanced accuracy (HAR / class-imbalance)
uses extra training data
Leaderboard row
2026
wav2sleep
In wav2sleep: A Unified Multi-Modal Approach to Sleep Stage Classification from Physiological Signals
0.78↑
Sleep staging · MESA · Cohen's κ
distinct train/validation/test setswithin-datasetend-to-end training30 s epochs, whole nightPPG + thoracic respiratory effort4-class (Wake, Light, Deep, REM)
wav2sleep (trained on all signals)
In wav2sleep: A Unified Multi-Modal Approach to Sleep Stage Classification from Physiological Signals
0.742↑
Sleep staging · MESA · Cohen's κ
held-out test participantswithin-datasetend-to-end training30 s epochs, whole nightPPG only at test time4-class (Wake, Light, Deep, REM)
wav2sleep (ECG→PPG transfer)
Baseline in wav2sleep: A Unified Multi-Modal Approach to Sleep Stage Classification from Physiological Signals
0.732↑
Sleep staging · MESA · Cohen's κ
held-out test participantswithin-datasetend-to-end training30 s epochs, whole nightPPG only at test time4-class (Wake, Light, Deep, REM)
wav2sleep (PPG only)
Baseline in wav2sleep: A Unified Multi-Modal Approach to Sleep Stage Classification from Physiological Signals
0.728↑
Sleep staging · MESA · Cohen's κ
held-out test participantswithin-datasetend-to-end training30 s epochs, whole nightPPG only at test time4-class (Wake, Light, Deep, REM)
SleepPPG-Net (ECG→PPG transfer)
Baseline in wav2sleep: A Unified Multi-Modal Approach to Sleep Stage Classification from Physiological Signals
0.724↑
Sleep staging · MESA · Cohen's κ
held-out test participantswithin-datasetend-to-end training30 s epochs, whole nightPPG only at test time4-class (Wake, Light, Deep, REM)
SleepPPG-Net (re-implemented, PPG)
Baseline in wav2sleep: A Unified Multi-Modal Approach to Sleep Stage Classification from Physiological Signals
0.713↑
Sleep staging · MESA · Cohen's κ
held-out test participantswithin-datasetend-to-end training30 s epochs, whole nightPPG only at test time4-class (Wake, Light, Deep, REM)
Bakker et al. (2021)
Baseline in wav2sleep: A Unified Multi-Modal Approach to Sleep Stage Classification from Physiological Signals
0.68↑
Sleep staging · MESA · Cohen's κ
distinct train/validation/test setswithin-datasetend-to-end training30 s epochs, whole nightPPG + thoracic respiratory effort4-class (Wake, Light, Deep, REM)
Random Forest (PPG features)
0.65↑
Sleep staging · MESA · Cohen's κ
not stated in the abstractwithin-datasetclassical machine learning on PPG features5-class
Sleep staging · opentslm-sleep-cot · Macro-averaged F1
Protocol not structured
Leaderboard row
wav2sleep
In wav2sleep: A Unified Multi-Modal Approach to Sleep Stage Classification from Physiological Signals
0.78↑
Sleep staging · SHHS · Cohen's κ
distinct train/validation/test setswithin-datasetend-to-end training30 s epochs, whole nightECG + thoracic respiratory effort4-class (Wake, Light, Deep, REM)
wav2sleep
In wav2sleep: A Unified Multi-Modal Approach to Sleep Stage Classification from Physiological Signals
0.74↑
Sleep staging · SHHS · Cohen's κ
distinct train/validation/test setswithin-datasetend-to-end training30 s epochs, whole nightECG only4-class (Wake, Light, Deep, REM)
Carter et al. (2024)
Baseline in wav2sleep: A Unified Multi-Modal Approach to Sleep Stage Classification from Physiological Signals
0.75↑
Sleep staging · SHHS · Cohen's κ
distinct train/validation/test setswithin-datasetend-to-end training30 s epochs, whole nightECG + thoracic respiratory effort4-class (Wake, Light, Deep, REM)
Sridhar et al. (2020)
Baseline in wav2sleep: A Unified Multi-Modal Approach to Sleep Stage Classification from Physiological Signals
0.66↑
Sleep staging · SHHS · Cohen's κ
distinct train/validation/test setswithin-datasetend-to-end training30 s epochs, whole nightECG only4-class (Wake, Light, Deep, REM)
Bakker et al. (2021)
Baseline in wav2sleep: A Unified Multi-Modal Approach to Sleep Stage Classification from Physiological Signals
0.64↑
Sleep staging · SHHS · Cohen's κ
distinct train/validation/test setswithin-datasetend-to-end training30 s epochs, whole nightECG + thoracic respiratory effort4-class (Wake, Light, Deep, REM)
0.908↑
Sleep staging · Sleep Heart Health Study (SHHS, NSRR) · Area under ROC
REM vs all other epochs (binary, 30-s epochs); not a 5-class staging score
Leaderboard row
74.1↑
Sleep staging · Sleep Heart Health Study (SHHS, NSRR) · Macro-averaged F1
uses extra training data5 classes (W, N1, N2, N3, REM)
Leaderboard row
Dense temporal apnea profiling (PPG, SpO2-assisted)
0.71↑
Sleep-disordered breathing · MESA · Macro F1
not stated in the abstractwithin-datasetend-to-endsingle-channel PPG plus SpO2 for severitysleep apnea severity classes
Sleep-disordered breathing · SDB (Garde et al.) · AUROC
subject-level train/val/testwithin-datasetlinear probe on frozen embeddings
SimCLR
Baseline in PaPaGei: Open Foundation Models for Optical Physiological Signals
0.61↑
Sleep-disordered breathing · SDB (Garde et al.) · AUROC
subject-level train/val/testwithin-datasetlinear probe on frozen embeddings
0.59↑
Sleep-disordered breathing · SDB (Garde et al.) · AUROC
subject-level train/val/testwithin-datasetlinear probe on frozen embeddings
Chronos
Baseline in PaPaGei: Open Foundation Models for Optical Physiological Signals
0.58↑
Sleep-disordered breathing · SDB (Garde et al.) · AUROC
subject-level train/val/testwithin-datasetlinear probe on frozen embeddings
0.58↑
Sleep-disordered breathing · SDB (Garde et al.) · AUROC
subject-level train/val/testwithin-datasetlinear probe on frozen embeddings
Showing 351–400 of 433 results
Recent changes
New entry: A cross-subject stress recognition method…, 0.9158
Stress detection (acute / prolonged / affect): WESAD: accuracy
New entry: PC-AVCT wearable-state model, 7.697
From the hub's dated sync history: new results as they are added, and SOTA stars as they are awarded or move.
How SOTA is decided
- marks the best value for one task, dataset and metric, using the metric's direction (higher or lower is better).
- Results are compared only under the same evaluation protocol: split type, within- or cross-dataset testing, filtering, inputs and label set. Different protocols stay separate and are never ranked together.
- Only results with a table, figure or section reference compete, and a star needs at least two competing methods. Ties go to the earlier publication.
- Stars are recomputed whenever the research hub syncs, and every change is logged.
Downloads of every result (JSON and CSV) are free for members. Sign in to download.
Results as reported by the authors; not independently reproduced. Leaderboard rows without a structured protocol are listed but don't receive stars. See also the leaderboards.
Cite or reuse this
CC BY 4.0ChatPPG compiled this data. You can copy, adapt and republish it, including in articles, slides, papers and AI tools, as long as you credit ChatPPG and link to this page. Please also cite the original studies for any number you use.
Attribution
Source: ChatPPG, Wearable AI results explorer, https://chatppg.com/research/results (CC BY 4.0)
HTML link
Source: <a href="https://chatppg.com/research/results">ChatPPG, Wearable AI results explorer</a> (<a href="https://creativecommons.org/licenses/by/4.0/" rel="license">CC BY 4.0</a>)