Bioelectronics Neurophysiology and Engineering Lab (BNEL)#
This is the central documentation hub for software tools developed in the Bioelectronics Neurophysiology and Engineering Lab (BNEL) at Mayo Clinic. The tools are designed to facilitate the analysis of brain electrophysiology, behavior, and dynamics.
The Bioelectronics Neurophysiology and Engineering Laboratory, led by Gregory A. Worrell, M.D., Ph.D., is a multidisciplinary discovery and translational research laboratory focused on exploring the human brain. Learn more on the official website.
BrainMaze toolbox 
#
BrainMaze is a modular Python toolbox for analyzing brain electrophysiology,
behavior, and dynamics. Each component is a standalone, pip-installable
package with its own documentation.
EEG/iEEG analysis: interictal spike detection (Janca, Barkmeier, and a gap-aware wrapper for recordings with dropouts), slow-wave detection, sleep classification, hypnogram scoring, and CRP analysis of evoked responses.
Ready-to-use trained EEG/iEEG models on ONNX Runtime (CPU, or NVIDIA
(CUDA) GPU via onnxruntime-gpu; no PyTorch needed): seizure
probability and OpenSpindleNet sleep-spindle detection.
Shared signal, file, and statistics utilities used across the toolbox, including resampling and helpers for NaN gaps in recordings.
ZeroMQ communication wrappers for real-time acquisition pipelines.
Superseded by BrainMaze EEG Models (brainmaze-eeg-models), which
carries on its seizure-probability model without needing PyTorch. New
projects should use brainmaze-eeg-models.
MEF3 tools#
Open-source toolchain for the Multiscale Electrophysiology Format (MEF 3.0):
libraries, a streaming server, and a desktop app. See the
MEF3 tools page for details, including the deprecated
mef_tools library.
Read and write MEF 3.0 recordings from Python, MATLAB and C++, and
de-identify them from Python and C++. Replaces mef_tools.
gRPC server and client for efficient, concurrent streaming of MEF3 data, with caching and prefetch for viewers and data-processing applications.
Desktop app for exploring and anonymizing MEF 3.0 recordings, built on mef3io (Windows, macOS, Linux). Test versions are downloadable from GitHub Releases.
Read two MEF3 recordings that were coregistered (aligned onto one clock) as if they were one recording. Built on mef3io.
More#
Code examples and links to papers from the Bioelectronics Neurophysiology and Engineering Lab (BNEL) at Mayo Clinic.