NeuroAtlas: Benchmarking Foundation Models for Clinical EEG and Brain-Computer Interfaces
- 1MIT ,
- 2KU Leuven ,
- 3TUSZ Siena SeizeIT1 CHB - MIT SeizeIT2 Helsinki Epilepsiae
Abstract
Foundation models (FMs) promise to extract unified representations that generalize across downstream tasks. They have emerged across fields, including electroencephalography (EEG), but it is less clear how effective they are in this particular field. Published evaluations differ in datasets, in the EEG-specific preprocessing that might influence reported results, and in the reported metrics, frequently obscuring the clinical relevance in EEG. We introduce NeuroAtlas, the largest EEG benchmark to date: 42 datasets and 260k hours covering clinical EEG (epilepsy, sleep medicine, brain age estimation) and brain-computer interfaces, and include multiple datasets per task along with bespoke clinical evaluation metrics. Besides evaluating EEG-FMs with respect to supervised baselines, we present results from generic time-series FMs. We report three findings. First, EEG-specific FMs do not consistently outperform time-series FMs, which have neither EEG-focused architectures nor been pretrained on EEG. Second, standard machine learning metrics are insufficient to assess clinical utility: thus, we thoroughly evaluate more appropriate measures such as the quality of event-level decision-making, hypnogram-derived features, and the brain-age gap in the domains of epilepsy, sleep, and brain age, respectively. Third, model rankings and performance can vary substantially within domains. We conclude that pretrained models perform largely on par, with only narrow advantages for a few, and that current models do not yet deliver on the promise of an out-of-the-box unified EEG model. NeuroAtlas exposes this gap and provides the datasets and metrics for the next generation of unified EEG FMs.
Related
- Clinical safety assessment of implanted brain-computer interfaces: current evidence and a pathway to standardization
- Human learning of noninvasive brain–computer interfaces via manifold geometry
- Real-time noise cancellation for EEG monitoring and brain-computer interfaces
- Large language models integrated into brain–computer interfaces for communication and control: a systematic review
- Regulating Next-Generation Implantable Brain-Computer Interfaces: Recommendations for Ethical Development and Implementation
BibTeX
@misc{kontras2026neuroatlas,
title = {{NeuroAtlas: Benchmarking Foundation Models for Clinical EEG and Brain-Computer Interfaces}},
author = {Konstantinos Kontras and Trui Osselaer and Stylianos G. Mouslech and Angeliki-Ιlektra Karaiskou and Guido Gagliardi and Thomas Strypsteen and Mohammad Hossein Badiei and Anku Rani and Maarten Vanmarcke and Miguel Bhagubai and Chanakya Ekbote and Jaedong Hwang and Christos Chatzichristos and Paul Pu Liang and Maarten De Vos},
howpublished = {arXiv (Cornell University)},
year = {2026},
month = may,
doi = {10.48550/arxiv.2605.14698},
url = {https://arxiv.org/abs/2605.14698},
note = {CC BY-NC-ND},
}
RIS
TY - UNPB
TI - NeuroAtlas: Benchmarking Foundation Models for Clinical EEG and Brain-Computer Interfaces
AU - Konstantinos Kontras
AU - Trui Osselaer
AU - Stylianos G. Mouslech
AU - Angeliki-Ιlektra Karaiskou
AU - Guido Gagliardi
AU - Thomas Strypsteen
AU - Mohammad Hossein Badiei
AU - Anku Rani
AU - Maarten Vanmarcke
AU - Miguel Bhagubai
AU - Chanakya Ekbote
AU - Jaedong Hwang
AU - Christos Chatzichristos
AU - Paul Pu Liang
AU - Maarten De Vos
PY - 2026
DA - 2026/05/14
PB - arXiv (Cornell University)
DO - 10.48550/arxiv.2605.14698
UR - https://arxiv.org/abs/2605.14698
AB - Foundation models (FMs) promise to extract unified representations that generalize across downstream tasks. They have emerged across fields, including electroencephalography (EEG), but it is less clear how effective they are in this particular field. Published evaluations differ in datasets, in the EEG-specific preprocessing that might influence reported results, and in the reported metrics, frequently obscuring the clinical relevance in EEG. We introduce NeuroAtlas, the largest EEG benchmark to date: 42 datasets and 260k hours covering clinical EEG (epilepsy, sleep medicine, brain age estimation) and brain-computer interfaces, and include multiple datasets per task along with bespoke clinical evaluation metrics. Besides evaluating EEG-FMs with respect to supervised baselines, we present results from generic time-series FMs. We report three findings. First, EEG-specific FMs do not consistently outperform time-series FMs, which have neither EEG-focused architectures nor been pretrained on EEG. Second, standard machine learning metrics are insufficient to assess clinical utility: thus, we thoroughly evaluate more appropriate measures such as the quality of event-level decision-making, hypnogram-derived features, and the brain-age gap in the domains of epilepsy, sleep, and brain age, respectively. Third, model rankings and performance can vary substantially within domains. We conclude that pretrained models perform largely on par, with only narrow advantages for a few, and that current models do not yet deliver on the promise of an out-of-the-box unified EEG model. NeuroAtlas exposes this gap and provides the datasets and metrics for the next generation of unified EEG FMs.
SN - 2331-8422
ER -