MNE Analyze Python

High-Performance 3D Brain Visualization & Analysis

MNE Analyze Python is a cutting-edge desktop application designed for the visualization and analysis of MNE-Python source estimation data. Built on PySide6 for a modern, responsive user interface and wgpu-py (WebGPU) for next-generation hardware-accelerated 3D rendering, it offers a distinct performance advantage over traditional OpenGL-based viewers.

This tool is specifically engineered for researchers and clinicians working with MEG/EEG data who need to visualize cortical activity in real-time, explore raw signal recordings, and inspect source time courses with high fidelity.

• WebGPU Powered: Utilizes the latest graphics API standards via wgpu-py to render high-resolution cortical surfaces efficiently.
• Dynamic Visualization: Real-time rendering of "Electric" source activity superimposed on the brain surface.
• Atlas Support: Toggle between dynamic source activation and static Atlas region visualization (e.g., Desikan-Killiany).
• Butterfly Traces: Overlay global signal traces directly on the 3D view for temporal context.

• Subject Configuration Panel: A dedicated sidebar allows you to easily manage and switch between files for a specific subject:
  • Raw Recordings: Load .fif files to inspect sensor-level data.
  • Surfaces: Import cortical surface geometries (.gii, .obj, etc.).
  • Atlases: Load parcellation labels/annotations.
  • Source Estimates (STC): Independently load Left Hemisphere (-lh.stc) and Right Hemisphere (-rh.stc) source time courses.

• Raw Data Browser: A fully integrated, interaction-rich plot for exploring raw MEG/EEG sensor data. Features include zooming, panning, and time-locking with the 3D view.
• Source Traces Browser: A dedicated tab for visualizing source-level time courses (butterfly plots), automatically synchronized when you load an STC file.

• Time-Locked Playback: Scrub through time or play back the neural activity.
• Interactive Navigation: Intuitive camera controls for the 3D view.
• Keyboard Shortcuts: Rapidly toggle modes (T), traces (P), or playback (Space).

Before installing, ensure your system meets the following requirements:

• Operating System: macOS 12+, Windows 10/11, or modern Linux distribution.
• Python: Version 3.10 or higher.
• Graphics Hardware: A GPU compatible with WebGPU, Vulkan, Metal, or DirectX 12.
  • Note: This application relies on wgpu-py which accesses the GPU directly. Older integrated graphics requiring legacy OpenGL drivers may not work.

git clone https://github.com/your-username/mne-analyze-python.git
cd mne-analyze-python

It is highly recommended to use a virtual environment to manage dependencies.

Using venv:

python -m venv .venv
source .venv/bin/activate  # On Windows: .venv\Scripts\activate

Using uv (Faster):

pip install uv
uv venv
source .venv/bin/activate

Install the package in editable mode to ensure all dependencies (including MNE, PySide6, and wgpu) are installed correctly.

pip install -e .

Once installed, you can launch the application directly from the root directory:

python main.py

The application window is divided into three main areas:

• Left Sidebar (Controls): Manage Visualization settings (Electric vs Atlas) and Subject Configuration.
• Central View: The main 3D Brain Viewport.
• Tabs: Switch between the Brain View, Raw Browser (sensor data), and Source Traces (source data).

Use the Subject Configuration section in the top-left to load your files:

• Click the Folder Icon next to Recording to open a .fif file. This populate the "Raw Browser".
• Click the Folder Icon next to Source (LH) or Source (RH) to load source estimates. This will populate the "Source Traces" tab and the 3D view.

• 3D View: Click and drag to rotate. Scroll to zoom. Right-click and drag to pan.
• Keyboard Shortcuts:
  • Space: Play / Pause animation.
  • T: Toggle between Source Activation view and Atlas Region view.
  • P: Toggle "Butterfly Traces" overlay on the 3D view.

• "WebGPU device not available": Ensure your GPU drivers are up to date and support Vulkan (Windows/Linux) or Metal (macOS).
• Performance Issues: If rendering is slow, check if the application is using your dedicated GPU instead of integrated graphics (common on laptops).

MIT License. See LICENSE for details.