Foundation models challenge spanning EEG and EMG neural interfaces.
2026
Toward OpenEEG-Bench: A Live Community-Driven Benchmark for EEG Foundation Models
34th European Signal Processing Conference (EUSIPCO 2026), Bruges, Belgium , pp. 1-5
BCI · EEG · DEEP LEARNING
Pierre Guetschel
Research
I build foundation models and deep-learning algorithms for neurophysiological signals to advance Brain–Computer Interfaces (BCI): reducing calibration times, increasing robustness, and improving transfer learning. I am interested in whether deep learning can push the limits of what can be decoded from EEG and other neurophysiological signals.
- Affiliation
- Radboud University · Donders Institute · NeuroTechnology Lab
- Supervisor
- Michael Tangermann
- Co-supervisor
- Thomas Moreau
- Based in
- Nijmegen, the Netherlands
Publications
Lead author
2025
Identifying Good Donor Datasets for Transfer Learning Scenarios in Motor Imagery BCI
Artificial Intelligence and Machine Learning (Springer) , pp. 209-235
2024
Review of Deep Representation Learning Techniques for Brain-Computer Interfaces
Journal of Neural Engineering 21(6) , pp. 1-20
2024
2022
2021
Contributed
2026
2025
2025
Learning from Small Datasets — Review of Workshop 6 of the 10th International BCI Meeting 2023
Journal of Neural Engineering 22(3)
2025
2024
Approximate UMAP Allows for High-Rate Online Visualization of High-Dimensional Data Streams
9th Graz BCI Conference
2024
Synthesizing EEG Signals from Event-Related Potential Paradigms with Conditional Diffusion Models
9th Graz BCI Conference
2024
2022
WOLF: A Modular Estimation Framework for Robotics Based on Factor Graphs
IEEE Robotics and Automation Letters 7(2) , pp. 4710-4717
2021
Tools for Convulsive Seizures and Interictal Spikes Detection
Open Source
Braindecode
core maintainer · 60+ PRs
Deep-learning tools for neurophysiological signal decoding.
- Standardized model architectures
- New dataloaders, including generic BIDS
- Hugging Face Hub support and static typing
MOABB
core maintainer · 70+ PRs
Mother of All BCI Benchmarks — a library for reliably benchmarking BCI algorithms.
- Caching mechanism and MNE-BIDS integration
- Dataset visualization and new datasets (c-VEP, SSVEP, ERP)
- PapersWithCode leaderboard automation
OpenEEGBench
lead author
Community-driven live benchmark for evaluating EEG foundation models.
Other contributions
contributor
Upstream contributions across the scientific Python ecosystem.
- EEGDash — open-source platform for ML on public neurophysiological data
- PyTorch — data-loading mechanism
- MNE — BIDS integration
- Nemar — converted and shared several datasets in BIDS format
Organisation
Competitions
From cross-task to cross-subject EEG decoding · NeurIPS 2025 competition.
Workshops
Workshop on foundation models for brain and body signals.
Experience
Feb. 2022 – Jul. 2026
Ph.D. Researcher
Radboud University · Donders Institute · NeuroTechnology Lab · Nijmegen, NL
with Michael Tangermann, Thomas Moreau
Foundation models for neurophysiological signals; reducing BCI calibration times; improving robustness and transfer learning.
- OpenEEGBench — community-driven live benchmark for EEG foundation models (lead author)
- Co-organizer, NeurIPS 2025 EEG Foundation Challenge
- Co-organizer, NeurIPS 2025 Brain & Body Foundation Models workshop
Aug. 2024 – Feb. 2025
Research Intern — EMG Foundations Team
Meta · Paris, FR
with Alexandre Gramfort, Zaccharie Ramzi
Deep learning for surface-EMG signal decoding for wearable BCI (details under NDA).
Sept. 2020 – Aug. 2021
Year of Research (ARPE)
Donders Institute · INRIA Sophia Antipolis · University of Freiburg (BSD Lab)
with Michael Tangermann, Théodore Papadopoulo
Novel representations of EEG signals using deep learning (continued in the Ph.D.).
Mar. 2020 – Aug. 2020
M2 Research Intern
INRIA Sophia Antipolis · ATHENA team · Valbonne, FR
with Théodore Papadopoulo
Dictionary learning for EEG signals; seizure detection and prediction for epileptic mice.
Mar. 2019 – Aug. 2019
M1 Research Intern
Institut de Robòtica i Informàtica Industrial (IRI) · Barcelona, ES
with Joan Solà
CNN-based visual place recognition for loop closure in SLAM; work included in the WOLF library.
Mar. 2018 – Jul. 2018
Bachelor Research Intern
Aix-Marseille University · LIS · Marseille, FR
with Alexis Nasr
Machine learning for coreference resolution in French.
Education
Feb. 2022 – Jul. 2026
Ph.D. — A.I. & Neurotechnology
Radboud University · Donders Institute · Nijmegen, NL
Foundation models for neurophysiological signals; reducing BCI calibration times; improving robustness and transfer learning.
2020
M.Sc. — Learning & Algorithms (M2A)
Sorbonne University · Paris, FR
Theory and applications of machine learning and deep learning. Thesis: epileptic seizure detection in mice EEG.
2019
M.Sc. — Parisian Master of Research in Computer Science (MPRI)
ENS Paris-Saclay · Paris, FR
Fundamental computer science. Thesis: place recognition for robust loop closure in SLAM graphs.
2018
B.Sc. — Fundamental Computer Science
ENS Paris-Saclay · Paris, FR
Theoretical and applied computer science. Thesis: coreference resolution in natural language.
Talks
2026
EEG foundation models
Donders Day
2025
Braindecode library ▶ video
IEEE NER workshop
2025
Braindecode library
Practical MEEG 2025
2024
MOABB
9th Graz BCI Conference
2024
SignalJEPA
9th Graz BCI Conference
2023
The MOABB library for reliable benchmarking of BCI algorithms
Cutting Gardens 2023
2023
Transfer Learning between Heterogeneous EEG Motor Imagery Datasets
BCI Meeting · Brussels
2022
Transfer Learning between Heterogeneous EEG Motor Imagery Datasets
IEEE MetroXRAINE · Rome
Teaching
Bachelor-level BCI course
- Group-work guidance
- Lecture: Deep Learning for BCI decoding
- Exam creation & grading
Master-level advanced BCI course
- Lectures: Deep Learning for BCI decoding; Technologies for BCI
- Group-work guidance
- Exam creation & grading
Supervision
Guido Klein
Diffusion models for EEG generation — Conditional diffusion models for EEG-based BCI example synthesis.
Peter Wassenaar
Online visualization of embedded EEG signals — Hybrid projection combining UMAP and PCA for online visualization.
Thyra Hogervorst
Foundation models for online BCI decoding — Applying foundation models to real-time (online) decoding of brain-computer interfaces.
Theses
Embeddings for EEG Signals
University of Freiburg & INRIA Sophia Antipolis
EEG Signal Analysis for Epileptic Seizure Genesis Study
INRIA Sophia Antipolis
Visual Place Recognition for Robust Loop Closure in SLAM
IRI, Barcelona
Résolution de Coréférences
LIS, Marseille
Awards & Grants
Best Poster Award
Theme: Natural Computing & Neurotechnology.
Normalien status
Competitive scholarship — €1,500/month for four years.
Reviewing
Conferences & workshops
ICLR
3 — 2025 main track
ICML
6 papers — 2026 (Gold Reviewer award)
MLSP
3 — 2025 workshop
NeurIPS
6 (2025 main) · 2 (2025 BERT2s workshop) · 3 (2024 main) · 4 (2026 main) · 3 (2026 competitions)