Real-time biosignal experiment GUI builder

Build EMG experiments in Python.

One short script gives you a window with live signals, recording, training and prediction. Your code is plain functions and widget calls.

$ pip install myogestic

The Playground runs in your browser and takes about 30 s to load the first time.

P-001 / myocontrol · live2048 Hz · 64 ch
StreamEMG
sourceLSLSourcechannels64rate2048 Hzwindow1000 ms
PredictionLDA
Fist
p 0.94
Recordingclick a gesture
Gesture
RecordingFist
ModelPipeline
Trained4 classes
Predictinglive
Targets2
Virtual HandLIVE
LSL outletIDLE
The background shows 64 channels of synthetic EMG.Click a gesture to change it.
A minimal app

Hello EMG

hello_emg.pyPython 3.12+
from myogestic import App, Streamfrom myogestic.sources import LSLSourcefrom myogestic.widgets import RecordingControls, SignalViewer
Hello EMG app with SignalViewer and RecordingControls in the dark theme
Playground

Try MyoGestic in your browser.

Record a few seconds of each gesture, train an LDA classifier and test its predictions. Nothing needs to be installed.

nsquaredlab.github.io/MyoGestic/playground
Recording
RestFistPinchIndex
Model
Trained · LDA
Prediction
Fist

Built on Pyodide, imgui-bundle and scikit-learn. The first load takes about 30 s; later loads use the browser cache.

Complete protocols

Run them from a clone of the repository.

force_ramps.py

Isometric force ramps

Track a trapezoidal force target at a set percentage of maximum voluntary contraction. EMG, force and target are recorded together.

uv run python examples/start_here/force_ramps.py
Read the guide →
myocontrol.py

Myocontrol

Record cued Rest and Fist trials, train a classifier or regressor, and drive the Virtual Hand with it.

uv run --extra examples --extra grpc python examples/start_here/myocontrol.py
Read the guide →
pong.py

Pong

Follow a cursor with the paddle to train a signed wrist command from −1 to +1, then play.

uv run --extra examples --extra grpc python examples/start_here/pong.py
Read the guide →
Set up a deviceQuattrocentoMuoviMuovi+SessantaquattroSessantaquattro+Any LSL streamSynthetic amplifier
Cite
Science AdvancesResearch article · 2025

MyoGestic: EMG interfacing framework for decoding multiple spared motor dimensions in individuals with neural lesions

First page of the MyoGestic paper in Science Advances

Sîmpetru, Braun, Simon, März, Cnejevici, Souza de Oliveira, Weber, Walter, Franke, Höglinger, Prahm, Ponfick, Del Vecchio

Sci. Adv. 11(15), eads9150 · doi:10.1126/sciadv.ads9150

Read the paper