Multimodal cognitive-load dataset

ADABase

51 subjects · 135 h of synchronized performance ratings, self-subjective questionnaires, physiological and behavioral signals recorded under n-back and simulated-driving workload.

Department Digital Health and Analytics, Fraunhofer IIS, Fraunhofer Institute for Integrated Circuits IIS · Department Artificial Intelligence in Biomedical Engineering, Friedrich-Alexander-University Erlangen Nuremberg

Request access Read paper 2023 Documentation notebook BibTeX
Dataset at a glance
Subjects
51 (30 public)
Recording
135 h
Modalities
7 signals
Format
HDF5
Biosignals
ECG, EDA, EMG, PPG, RSP, SKT · stored at 1 kHz
Eye / Video
Gaze & pupil (250 Hz), facial AUs & body landmarks (no raw video)
Labels
NASA-TLX, performance
License
Academic-only EULA
Overview

Two workload paradigms

The established n-back test and a novel simulator-based k-drive task induce single- and dual-task load of increasing intensity, captured on a common analog clock alongside gaze, facial action units and driving performance.

ADABase is continued by REVELIO, a gaming cohort recorded with the same setup. Together they comprise 96 sessions from 83 unique participants.

Physiological signals
Eye-tracking & gaze
Recording room
Driving simulator
Access · academic use only

Request the dataset

01
Check eligibility
Permanent position at an academic institution, institutional email address, non-commercial academic research only.
02
Request on Zenodo
Use Request access on the Zenodo record. State name, position, affiliation, postal address, institutional email, up to five co-researchers and the intended research use.
03
Accept the EULA
Submitting the request accepts the End User License Agreement. Access is granted once the request has been reviewed.
ADABase Public Set on Zenodo REVELIO Public Set on Zenodo Read the EULA

The Public Set holds recordings 0000–0029; the Private Set (0030–0050) is retained as a hidden test set for benchmarking. Access is free of charge. Questions and commercial inquiries: adabase@iis.fraunhofer.de

Documentation

Loading the public sets

A Jupyter notebook walks through the ADABase and REVELIO public sets end to end, from the HDF5 layout to model-ready windows with a subject-wise split. Driving and gaming files share one layout, so the same code reads both.

View in browser Notebook (.ipynb) on GitHub
  1. 01Files, tables and channels: SIGNALS, SUBJECTIVE, PERFORMANCE, META
  2. 02Participants and protocol, annotation by STUDY, PHASE, LEVEL
  3. 03NASA-TLX ratings and task performance per level
  4. 04Biosignals, downsampling to 100 Hz and normalization
  5. 05Window selection and building the dataset
  6. 06Suggested subject-wise split of the public sets
Cite

Cite this dataset

Please cite the Zenodo records you use and both papers: ADABase introduced the acquisition setup, and REVELIO extends it with the gaming cohort.

Oppelt, M.P.; Foltyn, A.; Deuschel, J.; Lang, N.R.; Holzer, N.; Eskofier, B.M.; Yang, S.H. ADABase: A Multimodal Dataset for Cognitive Load Estimation. Sensors 2023, 23(1), 340. doi:10.3390/s23010340

Oppelt, M.P.; Foltyn, A.; Lang-Richter, N.R.; Eskofier, B.M. REVELIO: Multimodal Task Load Estimation for Cross-Domain Generalization. International Journal of Human–Computer Interaction 2026, in press. Preprint: doi:10.48550/arXiv.2509.01642

BIBTEX
@article{oppelt2023adabase,
  author  = {Oppelt, Maximilian P. and Foltyn, Andreas and Deuschel, Jessica
             and Lang, Nadine R. and Holzer, Nina and Eskofier, Bjoern M.
             and Yang, Seung Hee},
  title   = {{ADABase}: A Multimodal Dataset for Cognitive Load Estimation},
  journal = {Sensors},
  year    = {2023},
  volume  = {23},
  number  = {1},
  pages   = {340},
  doi     = {10.3390/s23010340}
}

@article{oppelt2026revelio,
  author    = {Oppelt, Maximilian P. and Foltyn, Andreas
               and Lang-Richter, Nadine R. and Eskofier, Bjoern M.},
  title     = {{REVELIO}: Multimodal Task Load Estimation
               for Cross-Domain Generalization},
  journal   = {International Journal of Human--Computer Interaction},
  publisher = {Taylor \& Francis},
  year      = {2026},
  doi       = {10.48550/arXiv.2509.01642},
  note      = {In press. DOI refers to the arXiv preprint arXiv:2509.01642}
}
Zenodo records & identifiers
RecordSessionsDOI
ADABase: Cognitive Driving Public Setthis dataset3010.5281/zenodo.24051552
REVELIO: Cognitive Gaming Public Set3010.5281/zenodo.24060698
Paper DOI: 10.3390/s23010340 REVELIO preprint: arXiv:2509.01642