2026-08-05 00:00 China Japan Poland Papers Foundations & Methods Translated from EN

AutoMI: Hands-Free Motor Imagery EEG Classification via LLM Multi-Agents

Summary The study presents AutoMI, a framework that uses LLM multi-agents to automatically and rapidly iterate on motor imagery EEG classification models, combining a Q-learning policy with deterministic rules and integrating planning, execution and output agents with predefined tools, plus experience tracking and rollback. Models built by AutoMI reached 77.62%, 78.08% and 83.02% accuracy on the IV2a, OpenBMI and ECUST-MI datasets — up 18.42%, 9.27% and 19.25% over automated optimization algorithms.

Why it matters Handing model optimization to LLM multi-agents attacks the human-expertise bottleneck that has kept BCI model tuning a specialist's job.

This study, published in Journal of Neuroscience Methods, proposes AutoMI, a framework that uses multi-agent automated rapid iterations to construct state-of-the-art motor imagery EEG classification models, coupling Q-learning strategies with deterministic rules and integrating experience tracking and rollback mechanisms.

On the IV2a, OpenBMI and ECUST-MI datasets, the models built by AutoMI achieved accuracies of 77.62%, 78.08% and 83.02%, exceeding automated optimization algorithms by 18.42%, 9.27% and 19.25% respectively. PMID 42551560, DOI 10.1016/j.jneumeth.2026.110871

Compiled by BCIwiki from public sources

Sources · 1

pubmed.ncbi.nlm.nih.gov 2026-08-05

East China University of Science and Technology Decoding Algorithms Motor Decoding China Japan Poland

East China University of Science and Technology timeline

2026-08 EFS-Net Fuses EEG and fNIRS for Hybrid BCI Decoding 2026-04 Enhancing the Capability and Accuracy of Motor Imagery Classification: A Deep Neural Network-Powered Multifaceted Strategy Model All entries →

Read original ↗

Suggest a correction Revisions · none 中 / EN