Historical record · 2021 to 2025
Track record.
Previous competitions explain the platform behind the current 2026 challenge. This page is historical context. Use the current routes to choose and prepare a track.
StatePreview
CompetitionSep 21 - Nov 21, 2026
BEETL Motor Imagery.
First in the lineage. Cross-subject motor-imagery EEG decoding across heterogeneous BCI datasets. Established the team's pattern of code-submission scoring with a held-out evaluation set, and the use of Braindecode + MOABB as the canonical stack.
What it set
First time cross-subject EEG transfer ran as a NeurIPS competition, with code submissions on a held-out set and an open-source stack from day one.
Brain Age Prediction.
EEG-based age regression on a multi-cohort benchmark. Stress-tested cross-site generalisation and introduced the dataset-aggregator pattern that NeuralBench inherits today.
What it set
Multi-cohort aggregation as a benchmark axis. The pattern that became NeuralBench's dataset registry.
Sleep States.
Sleep-stage classification across PhysioNet, Sleep-EDF, and HMC sources established an earlier wearable-EEG evaluation approach. The 2026 Muse Sleep-Onset track builds on that experience with repeated home recordings per participant and its own labeling protocol.
What it set
An earlier precedent for evaluating sleep algorithms on wearable EEG.
EEG Foundation Challenge 2025.
The immediate predecessor. Cross-task EEG decoding with single-task and foundation-transfer arms. The 2026 challenge is its operational continuation: same lead, same Codabench infrastructure, same open-source stack, plus EMG as a parallel modality.
Continuity
Same lead. Same platform. Same stack. EMG added as a parallel modality.
What 2026 changes from 2025.
Three additions over the 2025 edition. Each is a deliberate extension of the same competition design, not a rewrite.
New modality
ADD
EMG as a parallel modality.
For the first time, surface EMG hand-pose regression sits inside the same sealed benchmark as EEG. EMG2Pose pairs 16-channel wrist signals with 20 joint-angle trajectories and evaluates generalization to held-out users. The harness, splits, and audit are identical across modalities, so EEG and EMG results are directly comparable - a setup the 2025 edition did not include.
Track separation
ADD
Fixed evaluation across four tracks.
The four tracks evaluate task-specific decoders on the same data and metrics, so a per-track score is attributable to the model rather than the protocol.
Controlled shifts
ADD
Per-track primary shift.
Each task-specific track isolates a primary generalization problem: cross-stimulus for Track 1, cross-session for Track 2, within- and cross-user variability for Track 3, and cross-user for Track 4. That keeps post-competition analysis shift-specific rather than leaderboard-only.
What stays the same.
The 2026 edition keeps every load-bearing piece of the 2025 round. Lead, platform, and open-source stack are unchanged.
Lead
KEEP
Same coordinator.
Bruno Aristimunha leads both the 2025 and 2026 editions. Yneuro hosts both.
Platform
KEEP
Same Codabench infrastructure.
Codabench (Université Paris-Saclay, Sylvain Chevallier) hosts both. The 2026 challenge funds and contributes upstream a scaled AWS-backed deployment with Deloitte.
Stack
KEEP
Same open-source stack.
EEGLAB, MNE-Python, Braindecode, MOABB, EEGDash, NeuralSet, NeuralBench. The 2026 baselines reproduce in NeuralBench out of the box.
Current challenge · 2026
Ready to participate?
Compare the four current tracks, run a baseline, and register for the one you want.
Register your team.
Five years of EEG decoding competitions, condensed into one four-track benchmark. Open-source, sealed test, reproducibility audit for the Brain and Body Workshop at NeurIPS.