Software & Datasets
Open-source software and publicly released datasets.
Software
OpenBMI
Open-source toolbox and benchmarking platform for brain-computer interface development, with standardized, reproducible pipelines for MI, ERP, and SSVEP paradigms. Lead developer and technical architect.
GitHubNeuroTalk
Generative framework for synthesizing speech from brain signals, implementing the methods from our AAAI 2023 paper on voice reconstruction during imagined speech. Includes reproducible demos.
GitHubcIOL
Noise- and movement-artifact reduction method for ambulatory EEG, implementing the techniques reported in our IEEE Transactions on Neural Systems and Rehabilitation Engineering paper.
GitHubOMCR
Open-source Claude Code plugin providing agentic workflows for scientific research: literature synthesis, manuscript revision, citation checking, and reviewer response.
GitHubOMCX
Open-source multi-agent research workflow for OpenAI's Codex CLI, with specialized agents for hypothesis development, analysis, manuscript writing, figures, citation management, and peer review.
GitHubDeep-BCI
Collaborative open-source framework that uses deep learning to recognize user intentions from brain signals, spanning intelligent, ambulatory, cognitive-state, and zero-training BCIs.
GitHubAIHub-Brain
AI Innovation Hub repository for brain-computer interface and EEG research, including voice reconstruction from imagined speech (NeuroTalk) and diffusion-based EEG decoding (Diff-E).
GitHubDatasets
OpenBMI EEG Dataset
Standardized EEG dataset for three BCI paradigms (MI, ERP, SSVEP), released with the OpenBMI toolbox and published as a peer-reviewed data descriptor in GigaScience.
Dataset PaperMobileBCI Dataset
Scalp- and around-ear EEG recorded in real-world settings (standing, walking, running) with ERP and SSVEP paradigms, released with end-to-end preprocessing and analysis code (Scientific Data).
Dataset Paper