Motor unit identification from high-density surface EMG (HDsEMG) is one of the most powerful non-invasive tools for studying the neural control of movement. But the field is held back by a fundamental problem: there is no shared benchmark. Methods are evaluated on private datasets, compared against different baselines, and reported inconsistently — making it nearly impossible to track real progress.
MUnitQuest is our attempt to fix this, together as a community with support from the International Society of Electrophysiology and Kinesiology (ISEK). The competition runs two parallel challenges:
📊Data challenge: for experimental researchers using HDsEMG and simulation scientists developing electrophysiological models, who want to contribute reliably labeled motor unit spike trains to a shared, open dataset.
🤖Algorithm challenge: for algorithm developers working on computational methods that can reconstruct the activity of single neurons from complex mixtures, with two tasks: motor unit identification from (A) isometric, and (B) dynamic contractions; and two phases: (1) Familiarization for the development and optimization of algorithms, and (2) Showdown to determine the final Leaderboard.
All data contributed to MUnitQuest will be shared publicly in accordance with FAIR (Findable, Accessible, Interoperable, Reusable) principles. Top algorithm and dataset contributors will be invited to contribute to a special issue of the Journal of Electromyography and Kinesiology.
🗓️ Key dates:
- June 1 — The data challenge opens. Bring along your painstakingly curated dataset and share it with the rest of the community!
- June 15 — The Familiarization phase of the algorithmic challenge opens: download our train and validation datasets to prepare your algorithms and pipelines!
- June 25 — MUnitQuest Symposium at ISEK 2026 (12:30–14:00 CET)
- Sep 15 — Data submission closes.
- Nov 1 — The Showdown phase of the algorithmic challenge opens: blind evaluation on a subset of the data curated during the Data Challenge.
- Jan 15, 2027 — Winners announced 🏆
We'd love to see broad participation — from HDsEMG experimentalists and computational modellers, to signal processing and machine learning teams who have never worked with EMG before. The starter kit and baseline algorithms will make it easy to get started regardless of your background.
👉 Website: https://munitquest.github.io/
📜 Registration: register on Codabench then navigate to the Data and Algorithm challenges!
📧 Get in touch or become a partner