NeuXelec is a standalone desktop application for researchers working with stereoelectroencephalography (SEEG). It covers the full post-implantation workflow, from MRI/CT coregistration to electrode reconstruction, oblique reslicing, anatomical labelling, functional imaging and 3D visualization, including MNI-space group analysis, without requiring any programming knowledge. It also reads the NeuroInspire implantation plan, so the planned trajectories can be compared with the electrodes actually implanted.
Research use only. NeuXelec is not a certified medical device and must not be used as the sole basis for any clinical, diagnostic, or treatment decision.
Website and documentation: https://neuxelec.com
- Import & Coregistration - align MRI, CT, PET, ictal and inter-ictal SPECT and fMRI into a shared space (ANTs, rigid only for the CT), review every alignment before it counts, automatic brain masking, SISCOM.
- Electrode Reconstruction - two clicks per electrode on the CT; automatic contact localization in millimetre coordinates; with a plan loaded, the name, hemisphere and reference are proposed from the nearest planned trajectory.
- Oblique & Parcellation - reslice along each electrode axis, blend modalities, name every contact with the SEEG2parc measure.
- 3D View & Export - render electrodes on the cortical surface with the functional overlays, normalize to MNI152, export BIDS-compatible coordinates and publication figures.
- A project opens on another computer. Image paths are stored both as they were and relative to the project file, with a fingerprint of each file, so a project carried to another machine, another drive letter or another network share finds its data again on its own. When a file has genuinely moved, NeuXelec walks the neighbouring folders for it, and learns from a folder you point at once to resolve every other file that moved with it. Only what is still missing is asked for, and a fingerprint keeps another patient's same-named scan from being bound by mistake.
- One registration to the template instead of several. The brain mask, the defacing and the MNI normalization shared the same MRI-to-template registration and each recomputed it. It is now computed once and reused, which removes the bulk of the files the brain mask used to leave behind.
- Planning MRI and electrodes can be saved together, and a reopened project finds the registered planning MRI again.
- Contacts restricted to grey matter, in the contact table and in both views.
- A background colour for the 3D view and for the oblique slices, so a screenshot or an animation drops straight onto a coloured slide. In the oblique view the air around the head fades into the colour instead of being cut, leaving no dark outline.
- NeuroInspire plan import.
.nipfiles are read by a pure-Python reader that needs no database and no proprietary tool. Planned trajectories are drawn in the 3D scene next to the implanted electrodes, the name, hemisphere and Dixi reference of the nearest planned trajectory are proposed when the deepest contact is picked, and a divergence of more than 25 degrees between the reconstructed axis and the planned direction is reported. - Functional MRI as a full modality. Its own cockpit row, coregistration, brain mask, Save button and project round-trip. On import NeuXelec classifies the file: statistical map, scanner colour fusion (the activation is recovered from the colour ramp and realigned through its grey anatomy), raw 4D BOLD (refused with an explanation) or a map already in MNI space (warning). Shown as a 3D blob, projected on the pial surface or painted on the slice planes, with threshold, opacity, colour map, colour scale, minimum cluster size in mm3 and restriction to the cortical ribbon.
- Contacts table aligned on SEEG2parc. The dominant region of the weighted
3x3x3 voxel cube around each contact, identical to the BIDS export, with the
wmparcsubstitution that names white matter gyrus by gyrus, the full breakdown on hover and a tooltip on every column header. - Built-in user guide. A guided tour on the first launch of a new version, and a User guide window opened from the left menu or with F1, documenting every page, every right-click menu and every shortcut.
- Defacing correction. The exported region was inverted and kept the face while erasing the occiput. Face, eyes, nose, mouth and ears are now removed, the occipital scalp is preserved and the brain is never touched.
- Interface. Imaging cockpit split into Structural and Functional, right-click menus grouped in labelled sections, and a warning when a project is saved while a validated coregistration has not reached the disk.
Download the latest installer from https://neuxelec.com/download and run it. No Python or dependencies required.
Requires Python 3.10+ on Windows.
git clone <repository-url> NeuXelec
cd NeuXelec
python -m venv .venv
.venv\Scripts\activate
pip install -r requirements.txt
python scripts/run_neuxelec.pyExternal data not tracked in git (see .gitignore) and required at runtime:
tools/ants/- ANTs executables (coregistration, MNI normalization).templates/- MNI152 template, brain mask and parcellation atlases.
src/neuxelec/
├── app.py # entry point: QApplication, logging, project loop
├── main_window.py # main window, page lifecycle, safe teardown
├── state.py # AppState - single source of truth
├── project_io.py # project (.json) save / load, schema versioning
├── logging_config.py # rotating file logging
├── coregistration.py # ANTs coregistration wrappers
├── siscom.py # SISCOM computation
├── seeg2parc.py # weighted 3x3x3 contact to region measure
├── controllers/ # interaction controllers (electrodes, menu)
├── help/ # manual text, guided tour, User guide window
├── pages/ # the four workflow pages (files, recon, oblique, 3D)
├── ui/ # dialogs and reusable widgets
├── workers/ # QThread workers for long-running tasks
└── utils/ # imaging, coordinates, fMRI ingestion, plan reader
resources/ # Qt .ui files and images (icons, logo)
scripts/run_neuxelec.py # dev launcher
tests/ # pytest test suite
See ARCHITECTURE.md for the design and data flow.
pip install -r requirements.txt -r requirements-dev.txt
pytest # run the test suite
ruff check src tests # lint
black src tests # format
mypy # static type checkingSee CONTRIBUTING.md for conventions.
pyinstaller NeuXelec_windows.spec --clean --noconfirm
"C:\Program Files (x86)\Inno Setup 6\ISCC.exe" NeuXelec_setup.issThis produces installer/NeuXelec_Setup_<version>.exe.
Released under the GNU General Public License v3.0 (GPL-3.0). You are free to use, study, modify and redistribute NeuXelec, provided that any distributed derivative work is also released under GPL-3.0 with its complete source code made available. See LICENSE for the full text.
Research use only. NeuXelec is not a certified medical device and must not be used as the sole basis for any clinical, diagnostic or treatment decision.
If you use NeuXelec in your research, please cite:
@software{veber2026neuxelec,
author = {Veber, Jules},
title = {NeuXelec: an all-in-one application for stereo-EEG electrode reconstruction and multimodal brain visualization},
year = {2026},
url = {https://neuxelec.com},
institution = {HUG Geneva, University of Geneva}
}NeuXelec builds on Voxeloc (Monney et al., 2024), a MATLAB tool for SEEG electrode localization developed in the Human Neuron Lab. NeuXelec reimplements its two-point localization approach as a standalone Python application and extends it across the full SEEG workflow (coregistration, SISCOM, oblique reslicing, parcellation, MNI normalization and 3D visualization).
Monney J, Dallaire SE, Stoutah L, Fanda L, Mégevand P. Voxeloc: A time-saving graphical user interface for localizing and visualizing stereo-EEG electrodes. Journal of Neuroscience Methods. 2024;407:110154. doi:10.1016/j.jneumeth.2024.110154
Jules Veber · Hôpitaux Universitaires de Genève (HUG) and the Human Neuron Lab (HNL), University of Geneva · contact@neuxelec.com