An immersive modeling platform for computational neuroscience and neurology — focused on cell and tissue physiology, stochastic nano-morphology, and experimental-design replication.
BRAINCELL is developed by the Savtchenko / Rusakov Lab (UCL) as a structured simulation environment that combines the numerical rigor of NEURON (HOC + MOD) with a Python-driven export framework, JSON biophysics presets, and a manager-based architecture. It lets experimentalists and theorists model neurons, astrocytes, and surrounding tissue with nano-scale geometric precision — and reproduce in silico the conditions of the bench.
- Stochastic nano-morphology generation — procedurally seed spines, processes, and fine nano-structures on imported cell skeletons, producing statistically realistic ultrastructure rather than smoothed cylinders.
- Stop-save-go simulation control — pause long-running simulations, save full state to disk, and resume later (or on a different machine) without loss of fidelity. Essential for parameter sweeps and interactive exploration.
- Adaptive morphology import — ingest a wide range of reconstruction formats (SWC, ASC, HOC, NRX, XML, ESWC, IMS, and more) and adaptively refine segmentation for biophysically meaningful compartmentalization.
- Dynamic extracellular interactions — coupled inside-out / outside-in ion diffusion engines for realistic tissue-scale ionic dynamics (K⁺ buffering, glutamate spillover, Ca²⁺ waves).
- Manager-driven architecture —
BioManager,SynManager,GapJuncManager,InhomManager,StochManager, andExportManagerprovide clean separation of biophysics, synapses, gap junctions, inhomogeneity, stochasticity, and export. - JSON biophysics presets — version-controllable, reproducible configurations for astrocyte and neuron models.
- Ready-to-run simulations — including
SimMyelinatedAxon, calcium dynamics, FRAP, calcium waves, glutamate dynamics, and voltage distributions.
Important
Forum registration is required. Downloads and the setup password are distributed through the Neuroalgebra Forum. Please register at forum.neuroalgebra.net before attempting installation.
| Resource | Minimum | Recommended |
|---|---|---|
| RAM | 4 GB | 8 GB or more |
| Disk space | 5 GB | 10 GB (for saved simulation states) |
| Python | 3.10+ (Anaconda suggested) | 3.11 via Anaconda |
| NEURON | 8.x | 8.2 or newer |
| Graphics | Any OpenGL-capable GPU | Discrete GPU for nano-geometry rendering |
| Platform | Method | Notes |
|---|---|---|
| Windows | All-in-One Installer (.exe) |
Recommended path. Bundles NEURON, compiled mechanisms (nrnmech.dll), and BRAINCELL. Run as administrator. |
| macOS | Source build + NEURON wheel | Install NEURON via pip install neuron, then clone the repository. Compile MOD files with nrnivmodl in each Mechanisms/*/MOD_files directory. Apple Silicon is supported. |
| Linux | Source build + system NEURON | Install NEURON from your distribution or via pip. Compile mechanisms with nrnivmodl. Works on Ubuntu 22.04+, Debian 12+, and Fedora 38+. |
Warning
Password required during setup. The All-in-One Installer and source archives are protected. Obtain the current password from the Neuroalgebra Forum after registration.
After installation, launch the main entry point. You should see the BRAINCELL GUI load with the Main UI panel, geometry selectors, and manager buttons (MechManager, GapJuncManager, SynManager, ExportManager).
BRAINCELL ships with two companion AI assistants that help new users navigate the codebase and run simulations.
- Setup & Installation AI Guide — step-by-step walk-through of Anaconda, API keys, the
anthropicpackage, file placement, and first-run diagnostics. SeeBrainCell_Agent_Setup_Manual.htmlin the repository root. - End-User AI Manual — how to ask questions about MOD files, KINETIC schemes, biophysics, architecture, and request code modifications. See
BrainCell_Agent_User_Manual.html.
The agents themselves are provided as two Python scripts in the repository root:
| Script | Purpose |
|---|---|
braincell_mapper.py |
Indexes all MOD, HOC, and Python files into braincell_map.json. Run once after installation and again whenever files are added. |
braincell_agent.py |
Interactive AI assistant that answers questions about the codebase and proposes changes. Also available as a GUI via braincell_panel.py. |
Note
The AI agents use the Anthropic API and require an API key. Running cost is typically $5–20 / month for active research use.
| Resource | Link |
|---|---|
| Community Forum | forum.neuroalgebra.net |
| Documentation PDFs | Available through the Forum downloads area |
| GitHub repository | github.com/RusakovLab/BRAINCELL |
| NEURON simulator | neuron.yale.edu |
BRAINCELL is organized into clearly separated layers. Contributors should respect these boundaries:
- Geometry (classic + nano) — cell morphologies, with or without procedural nano-structures
- Biophysics — JSON presets under
Biophysics/Astrocyte/andBiophysics/Neuron/ - Mechanisms — MOD files, split into
Astrocyte/,Neuron/, andCommon/trees - Managers —
BioManager,SynManager,GapJuncManager,InhomManager,StochManager,ExportManager - Simulation layer — ready-to-run scenarios in
_Code/Simulations/ - Extracellular engines — inside-out and outside-in diffusion calculators
- Export framework — marker-driven (
@meta,py:) Python generators and skeleton templates - Reduced Inhomogeneous / Stochastic system — segmentation, distribution, and variable mapping
- GUI widgets — Tk-based control panels, interleaved with the engine
- Testing entry points —
_Testing/init_*.hocfor development
A CLAUDE.md file in the repository root provides architecture guidance and common commands for Claude Code. If you use Claude Code to work in this repository, it will be loaded automatically.
BRAINCELL is distributed under the 3-clause BSD license. See LICENSE for the full text.
If you use BRAINCELL in published research, please cite the platform and the Savtchenko / Rusakov Lab. Canonical citation details are available on the Forum.