Skip to main content

MNE Inspect

MNE Inspect

MNE Inspect is a 3D brain visualization and source analysis application. It provides an interactive viewer for FreeSurfer-reconstructed cortical surfaces, BEM models, source estimates, sensor layouts, functional connectivity networks, and evoked sensor-field maps — all within a single, integrated GUI.

MNE Inspect – 3D brain visualization

Getting Started​

Launch MNE Inspect from the command line:

mne_inspect [options]

The application opens with the default MNE sample dataset. You can specify your own data via command-line options (see below) or load files interactively through the GUI controls.

Command-Line Options​

OptionDescriptionDefault
--subjectPath <path>Path to the FreeSurfer subjects directory<app>/resources/data/MNE-sample-data/subjects
--subject <name>Subject namesample
--bem <path>BEM surface file (FIFF)—
--trans <path>Head-to-MRI transformation file (FIFF)—
--stc <path>Source estimate file (can be repeated for multiple STCs)—
--digitizer <path>Digitizer / sensor layout file (FIFF)—
--srcSpace <path>Source space or forward solution file (FIFF)—
--atlas <path>Atlas annotation file (lh or rh; the sibling hemisphere is auto-detected)—
--evoked <path>Evoked / average data file (FIFF)—

Example​

mne_inspect \
  --subjectPath /data/subjects \
  --subject sample \
  --stc /data/sample-stc-lh.stc \
  --bem /data/subjects/sample/bem/sample-5120-bem.fif \
  --trans /data/sample-trans.fif \
  --atlas /data/subjects/sample/label/lh.aparc.annot

Features​

Cortical Surface Visualization​

  • Load and display FreeSurfer-reconstructed surfaces (inflated, pial, white, sphere, etc.)
  • Switch between surfaces at runtime via the Surface combo box
  • Apply different shader modes for surface rendering
  • Toggle left and right hemisphere visibility independently

Atlas Overlays​

  • Load FreeSurfer atlas annotations (e.g., aparc, aparc.a2009s)
  • Overlay parcellation regions on the cortical surface with color coding
  • Both hemispheres are loaded automatically when one is specified

BEM Models​

  • Load and display the three BEM layers: inner skull, outer skull, and head surface
  • Toggle individual layers on or off
  • Separate shader control for BEM surfaces
  • Optional fixed-color mode for BEM rendering

Source Estimates (STC)​

  • Load and visualize MNE/dSPM/sLORETA source estimate time courses
  • Animated playback with adjustable speed (0.25x – 4x)
  • Real-time accurate playback mode with fractional sample stepping
  • Configurable colormap (Hot, Hot Negative, Jet, Bone, Red-Blue, Cool-Warm)
  • Adjustable threshold controls (min, mid, max) for activation display
  • Timeline scrubbing via slider
  • Support for loading multiple STC datasets and switching between them

Sensor and Digitizer Visualization​

  • Load sensor layouts and digitizer points from FIFF files
  • Toggle visibility by sensor type: MEG, EEG
  • Toggle digitizer point categories: cardinal, HPI, EEG, extra points
  • Apply head-to-MRI coordinate transformation in real time

Dipole Fitting Results​

  • Load and display dipole fit results
  • Toggle dipole visibility

Source Space​

  • Load and display source space points from a forward solution file
  • Toggle source space visibility

Functional Connectivity Networks​

  • Visualize connectivity networks in 3D
  • Adjustable threshold slider to filter network edges by strength
  • Configurable network colormap

Evoked Data and Sensor-Field Mapping​

  • Load evoked / averaged data files
  • Display MEG and EEG sensor-field interpolation on the scalp surface
  • Show contour lines for MEG and EEG fields
  • Visualize the MEG sensor helmet (convex hull or point cloud)
  • Timeline scrubbing synchronized with STC playback (optional)

Sensor-Field Streaming​

  • Stream sensor data with configurable modality (MEG / EEG)
  • Adjustable averaging window
  • Loop mode for continuous playback
  • Configurable colormap for sensor-field display

Multi-Viewport​

  • Support for multiple simultaneous 3D viewports (1–4)
  • Independent camera control per viewport
  • Camera presets (e.g., anterior, posterior, left, right, dorsal, ventral)
  • Per-viewport editing target selection

Usage​

The simplest way to get started is to launch MNE Inspect directly from the command line or by double-clicking the executable. When started without any arguments, the application opens with an empty scene and you can load data interactively through the GUI — use the toolbar buttons and file dialogs to add surfaces, overlays, sensors, or any other supported data type one at a time. You can also pass individual files via the command-line options listed above to pre-load specific datasets on startup.

For day-to-day work, however, you typically want to inspect a complete subject context at once — cortical surfaces together with BEM meshes, source estimates, sensor positions, coordinate transforms, and evoked data. Assembling all of those --stc, --bem, --digitizer, … flags by hand quickly becomes tedious. This is where the convenience launch scripts come in.

Quick Launch Scripts​

The source tree ships with ready-made scripts that launch MNE Inspect with all MNE sample-data files pre-loaded — surfaces, BEM, STC, digitizer, atlas, source space, evoked, and coordinate transform — so you can start exploring immediately without assembling a long command line.

PlatformScript
macOS / Linuxsrc/applications/mne_inspect/run.sh
Windowssrc/applications/mne_inspect/run.bat

What the Scripts Do​

  1. Locate the build — resolve the mne_inspect executable relative to the script directory (supports both flat builds and macOS .app bundles).
  2. Set data paths — default to ~/mne_data/MNE-sample-data (or %USERPROFILE%\mne_data\MNE-sample-data on Windows). Override by setting the MNE_DATA_PATH environment variable.
  3. Auto-discover STC files — scan the processed/ subdirectory for all *-lh.stc files and pass each one via --stc. If no processed STCs exist, the script falls back to the original sample_audvis-meg-eeg-lh.stc.
  4. Launch with full context — the final command line includes --subjectPath, --subject, --bem, --digitizer, --trans, --srcSpace, --atlas, --evoked, and all discovered --stc arguments.

Running the Scripts​

# From the project root (macOS / Linux)
./src/applications/mne_inspect/run.sh

# Or point to a custom data location
MNE_DATA_PATH=/data/my_mne_sample ./src/applications/mne_inspect/run.sh
:: From the project root (Windows)
src\applications\mne_inspect\run.bat

:: Or with a custom data location
set MNE_DATA_PATH=D:\data\my_mne_sample
src\applications\mne_inspect\run.bat
tip

The scripts are a good starting point for creating your own launch configurations. Copy one and adjust the paths to load your own subject, BEM, and STC files.

Online (Experimental)​

An experimental WebAssembly build of MNE Inspect is available. It runs entirely locally inside the browser — no installation needed, no data uploaded. Learn more about data privacy.

Dependencies​

MNE Inspect depends on the mne_disp3D library and is only built when this library is available. It also links against mne_fs, mne_fiff, mne_fwd, mne_mne, mne_disp, mne_inv, mne_utils, and mne_connectivity.

See Also​