Two adults recorded while building an adaptive learning platform. Failures were found after the fact by analysing the files (except the deliberate, timed dropouts). The README gives per-failure file, time and description, plus pandas loaders.
- Type
- Dataset
- Licence
- Not reported
- Access
- Not reported
- Subjects
- 2
- PPG sites
- Other
- Setting
- Not reported
- Wrist
- No
- Raw waveforms
- Yes
- Updated
- Oct 27, 2026
Original description
Recordings from a Muse S Athena headband (4-channel EEG at 256 Hz; optical heart sensor, PPG, at about 64 Hz with 4 or 16 channels), a webcam (mean face colour at about 30 Hz, no video) and a Galaxy Watch7 single-lead ECG at 500 Hz. Two adults made them while developing an adaptive learning platform.
The included README documents real device failures found in the recordings and where each occurs:
six whole-device dropouts, caused by walking out of Bluetooth range and by switching the headband off, with EEG and PPG recorded together in one file; poor electrode contact; mains interference that the headband’s own contact grade missed; a noisy EEG channel; bursty Bluetooth delivery and timestamp jitter; a heart-rate estimate that locked onto step rhythm during exercise; webcam heart-rate errors against ECG.
For each dropout, the README gives the outage length (10.8 to 38.4 s), whether samples were lost or delivered late (lost), and how quickly electrode contact recovered (3.4 to 4.0 s). Two recordings at 16 optical channels show a strained link that held without dropping.
Contents: 3 dropout recordings (7 to 15 min), 3 raw EEG recordings (4 to 9 min), 2 EEG score recordings, 6 PPG recordings (1 to 8 min), 2 webcam recordings, and 12 ECG readings of 30 s each. Files are gzipped JSON Lines and CSV, and the README gives a pandas loader for each file type.
Released under CC BY 4.0. Please cite the dataset when using it. Do not attempt to identify the wearers.
Verbatim fromZenodo· Licence: cc-by-4.0

Key numbers
How it works
No model in the hub reports results on this dataset yet.
- Sampling
- Muse S Athena optical (PPG) ~64 Hz, 4 ch (730 nm L/R, 850 nm L/R) or 16 ch; EEG 256 Hz (TP9/AF7/AF8/TP10); webcam mean face RGB ~30 Hz; Galaxy Watch7 lead-I ECG 500 Hz (30 s readings)
- Reference labels
- hr, ecg, sqi
Devices· 2
Icons show the form factor: we have no openly licensed or manufacturer-cleared photo of these exact devices yet.
Models trained or evaluated on it
FAQ
How many subjects are in Muse S Athena sensor-failure recordings?
Muse S Athena sensor-failure recordings has data from 2 subjects, as listed in its research record.
What signals does Muse S Athena sensor-failure recordings include?
Muse S Athena sensor-failure recordings includes PPG, EEG and ECG data. Sampling rate as recorded: Muse S Athena optical (PPG) ~64 Hz, 4 ch (730 nm L/R, 850 nm L/R) or 16 ch; EEG 256 Hz (TP9/AF7/AF8/TP10); webcam mean face RGB ~30 Hz; Galaxy Watch7 lead-I ECG 500 Hz (30 s readings). Raw waveforms are included.
What can Muse S Athena sensor-failure recordings be used for?
Research tasks it supports: heart rate estimation and motion and signal quality. Reference labels: heart rate, ECG and signal quality.
How do I get Muse S Athena sensor-failure recordings, and what is its licence?
It is hosted on Zenodo at https://zenodo.org/records/23651565. No licence is recorded for it here, so check the terms on the source page before you use it.
Cite
APA
Muse S Athena sensor-failure recordings [Data set]. (n.d.). zenodo. https://doi.org/10.5281/zenodo.23651565
BibTeX
@misc{musesathenasensomuse,
title = {{Muse S Athena sensor-failure recordings}},
howpublished = {zenodo},
doi = {10.5281/zenodo.23651565}
}Built from the hub record. Check it against the publisher before submitting.
Cite the ChatPPG record
Cite the dated snapshot 2026-10 for a fixed, reproducible version. Please also cite the original publications listed above.
APA
ChatPPG. (2026). Muse S Athena sensor-failure recordings: dataset record (Version 2026-10) [Data set]. ChatPPG Research Hub. https://chatppg.com/research/datasets/muse-s-athena-sensor-failures
BibTeX
@misc{chatppg_dataset_muse_s_athena_sensor_failures,
author = {{ChatPPG}},
title = {{Muse S Athena sensor-failure recordings: dataset record}},
year = {2026},
month = oct,
version = {2026-10},
howpublished = {\url{https://chatppg.com/research/datasets/muse-s-athena-sensor-failures}},
note = {ChatPPG Research Hub, Snapshot 2026-10, https://chatppg.com/research/snapshots/2026-10}
}CITATION.cff
cff-version: 1.2.0 message: "If you use this resource, please cite it as below." type: dataset title: "Muse S Athena sensor-failure recordings: dataset record" authors: - name: "ChatPPG" url: "https://chatppg.com/research/datasets/muse-s-athena-sensor-failures" repository-artifact: "https://chatppg.com/research/snapshots/2026-10" version: "2026-10" date-released: "2026-10-27"
History
Record updated
Papers
MoWaveQFormer: A Motion-Conditioned Quality-Gated Transformer for Smartphone-Based PPG Heart Rate EstimationSame task: Signal qualityarXiv preprint · 2026- Interpretable Reliability Control for Wearable PPG Heart Rate EstimationSame task: Signal qualityIEEE Access · 2026
- Reliability-Evidence Fusion for Selective Heart-Rate Estimation from Wearable PPG: A Transformation-Aware Cross-Cohort StudySame task: Signal qualitySSRN preprint · 2026
Datasets
WildPPGSame task: Signal quality16 subjects · PPG · ECG · IMU- Finger-contact ECG + dual-wavelength (660/940 nm) PPG, rest + post-exercise recovery (n=148, 444 records)Same task: Signal quality148 subjects · PPG · ECG
IEEE Signal Processing Cup 2015 (TROIKA)Same task: Signal qualitywrist PPG during intensive exercise12 subjects · PPG · ECG · IMU
