SPINE
Simulator for Plasticity and Integrated Neuronal Events
SPINE simulates neurons in mechanistic detail, from single cells to large networks on CPUs and GPUs.
- A coupled voltage–calcium model: PMCA, NCX, VDCC, SERCA, RyR, IP3R, SOC and leaks are built in, with Hodgkin–Huxley or your own channels and AMPA, NMDA, GABA and calcium-coupled synapses with plasticity.
- Extracellular stimulation: point-source, uniform-field and TMS fields, and myelinated axons.
- One entry point,
solve(), from one neuron on one CPU process up to multi-GPU and MPI populations. - Output: checkpoint and restart, streamed and HDF5 recording, 3-D rendering and animation.
The User Guide documents the models, every setting and the API. It is available at https://zacharymiksis.com/research/software/. The source archive on Zenodo (10.5281/zenodo.24226059) holds the guide, the examples and the tests. The PyPI package holds the library only.
Installation
pip install spine-neuro
This needs Python ≥ 3.10; NumPy, SciPy and threadpoolctl are installed with it. The package is imported as spine.
Extras combine freely, for example pip install "spine-neuro[viz,mpi]": examples (matplotlib), viz (3-D rendering and animation), mpi, gpu-cuda12 and gpu-cuda11, multi-gpu-cuda12 and multi-gpu-cuda11, hybrid-cuda12 and hybrid-cuda11, and metis. HDF5 output needs h5py (pip install h5py); no extra includes it. The User Guide lists the dependencies in its Dependencies section.
A natural-language front end that writes SPINE scripts is a separate package: pip install spine-assistant.
Quick start
from spine.utils import NeuronModel
from spine.solver import solve, read_results
from spine.graph import gen_beam
from spine.synapse import SynapseFactory
neuron = NeuronModel()
neuron.settings.top_folder = './'
neuron.settings.data_folder = 'calcium_wave/'
neuron.settings.output_folder = 'calcium_wave-output/'
for key in ('CYT_buffer', 'PMCA', 'NCX', 'SERCA', 'PM_leak', 'ER_leak', 'RyR', 'IP3', 'synapse'):
neuron.model[key] = True # the calcium mechanisms
neuron.geom_builder['const_rad'] = True
neuron.settings.rad = 0.2 # µm
gen_beam(neuron, 64, 1000) # a 64 µm beam of 1000 nodes
neuron.settings.T = 70.e-3 # 70 ms
neuron.settings.set_time_step(5.e-6)
neuron.synapse_instances = [ # calcium and IP3 influx at node 500
SynapseFactory.create_calcium_linear(nodes=[500], amplitude=2.5e-12, duration=1.e-3),
SynapseFactory.create_ip3_linear(nodes=[500], amplitude=5.0e-12, duration=200.e-3),
]
neuron.recorder['full'] = True # cytosolic Ca²⁺ → results/calcium_wave-output/cyt.bin
neuron.recorder['full_er'] = True # ER Ca²⁺ → results/calcium_wave-output/er.bin
neuron.recorder['format'] = 'binary' # one binary table per neuron
neuron.recorder['precision'] = 'float32'
solve(neuron)
res = read_results('results/calcium_wave-output')
res.t # time, s
res['cyt'] # (rows, 1000) cytosolic Ca²⁺ in µM
examples/basics/calcium_wave.py is a runnable version with plots. Internal units are µm, s, V and µmol/µm³; output is in mV and µM (see Units and conventions in the User Guide).
Examples
examples/ has one directory per topic (basics/, synapses/, morphology/, stimulation/, visualization/, gpu/, scaling/, cells/); every script writes to results/ below the current directory. The Index of examples in the User Guide describes each one. cells/ holds published cell models: the Jarsky CA1 cell under TMS (jarsky/), the Aberra layer-5 cell under field stimulation (aberra/), and a cerebellar Purkinje cell based on Masoli et al. (2015) with its SCA1[82Q] variant (purkinje/). The Jarsky and Purkinje directories have a README of their own.
python examples/basics/calcium_wave.py # start here
python examples/gpu/calcium_wave_gpu.py # needs a GPU extra
mpirun -n 8 python examples/scaling/parallel_scaling_test.py # needs the mpi extra
Parallel and GPU runs
solve(neuron) runs one neuron on one CPU process. mpirun -n N python script.py is detected and needs no code change. solve(..., backend='gpu', batched=True) and solve(..., backend='gpu', devices=[...]) run populations on one or several GPUs, and MPI ranks can each hold GPUs. Testing covers the ranks and devices of one node. Several nodes are untested. See Parallel and GPU Backends in the User Guide.
Testing and contributing
With pytest installed:
pytest tests/ # about three minutes on a workstation, CPU only
pytest tests/ -k synapse # a subset by keyword
The suite runs the User Guide listings, compares the batched, multi-device and MPI solvers with the serial solver, and checks checkpointing and recording. Tests that need CuPy, mpirun or pymetis are skipped when those are missing. On a GPU machine, run python tests/gpu_population_check.py.
The User Guide chapters Extending SPINE, Architecture and Design, and Contributing and Testing cover new channels, receptors and flux mechanisms, the package layout, and the workflow for changes (fork, branch, add a test, open a pull request).
Citation
@software{spine2026,
author = {Miksis, Zachary M. and Queisser, Gillian},
title = {{SPINE}: Simulator for Plasticity and Integrated Neuronal Events},
version = {1.0.0},
year = {2026},
doi = {10.5281/zenodo.24226059}
}
CITATION.cff holds the same entry for GitHub's "Cite this repository" button.
License
Copyright (C) 2026 Zachary M. Miksis and Gillian Queisser. Licensed under the GNU Lesser General Public License, version 3 or later (LGPL-3.0-or-later).
Acknowledgments
Works SPINE is based on:
- Standard model (calcium parameters and coupled electro-calcium formulation): P. R. Borole, J. M. Rosado, M. Neal and G. Queisser, "Neuronal Resilience and Calcium Signaling Pathways in the Context of Synapse Loss and Calcium Leaks: A Computational Modeling Study and Implications for Alzheimer's Disease," SIAM J. Appl. Math., 2023. doi:10.1137/23M1557842
- FJ model (Extended-Flux Fick-Jacobs, under development): Z. M. Miksis and G. Queisser, "A New Fick-Jacobs Derivation with Applications to Computational Branched Diffusion Networks," 2025. arXiv:2501.08247
Support
Email: miksis@temple.edu
Metadata
Release files for spine-neuro 1.0.0.post1
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| spine_neuro-1.0.0.post1.tar.gz | 561.0 kB | Without provenance | Details for spine_neuro-1.0.0.post1.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| spine_neuro-1.0.0.post1-py3-none-any.whl | Python 3 | none | any | Details for spine_neuro-1.0.0.post1-py3-none-any.whl |
Total release size: 1.1 MB
Release files / spine_neuro-1.0.0.post1.tar.gz
| Download URL | spine_neuro-1.0.0.post1.tar.gz |
|---|---|
| Size | 561.0 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
fd0e1472f85134dcb30196795843973cc6d2e692812be73d124e14dd19a620cd
|
|
BLAKE2b-256 checksum How to use checksums |
c1dd1758f58975bad232aeae9f75368304092f9827389d72676dced4aa87e25c
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/7.0.0 CPython/3.10.9
|
Release files / spine_neuro-1.0.0.post1-py3-none-any.whl
| Download URL | spine_neuro-1.0.0.post1-py3-none-any.whl |
|---|---|
| Size | 535.2 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
83839b58cfbda5fdf2b49e6b553ce5a271103f0ab8b57737e17b108c53644038
|
|
BLAKE2b-256 checksum How to use checksums |
436e7a1a0f9cd3f87548e656530d4f85430655f14b888233af691a1928383583
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/7.0.0 CPython/3.10.9
|
Security for spine-neuro 1.0.0.post1
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.