Brian simulator

69 Followers
7 Following
19 Posts
I am a spiking neural network simulator, developed by @neuralreckoning, @romainbrette, @mstimberg (and many others)
homepagehttps://briansimulator.org
GitHubhttps://github.com/brian-team/brian2
Documentationhttps://brian2.readthedocs.io
Discussionshttps://brian.discourse.group

🎁 We are happy to announce the Brian 2.10 release, now available from pypi and conda-forge, and soon from other sources. 🎁

This release extends the functionality of PopulationRateMonitor and SpikeMonitor, which now provide a unified interface to for smoothed and binned rates, and adds a new run_at function to execute code at specific time points. It also fixes several important bugs and comes with performance improvements for C++ standalone mode. Behind the scenes, it made important progress in refactoring aspects of our code generation system, working towards the long-term goal of having a single C++ code generation system for standalone and runtime mode, and retiring the current Cython code generation mechanism.

For more details, head over to the release notes.

Release notes β€” Brian 2 2.10.1 documentation

πŸš€ We are happy to announce the **Brian 2.9** release, now available from pypi and conda-forge, and soon from other sources. πŸš€

This release extends the functionality of Brian’s *linked variable* feature (in particular, it can now be used to refer to a global modulation term in `Synapses`), and adds a β€œgraceful interrupt” feature, allowing users to interrupt a simulation with Ctrl+C without triggering the usual Exception. As usual, the release also fixes a number of bugs and incompatibilities and comes with new examples in its documentation.

For more details, head over to the release notes: https://brian2.readthedocs.io/en/2.9.0/introduction/release_notes.html

Release notes β€” Brian 2 2.9.0 documentation

🧐

3 Replace Brian's just-in-time compilation mechanism (Full-time)

πŸ€” In Brian's "runtime mode", the generated code can interact with Python code by accessing the same data structures in memory. To make this work seamlessly, Brian makes use of Cython. This approach comes with a few downsides, since Cython is much more general than what we need, and is not meant for dynamic code generation.

πŸ‘‰ This project's aim is to replace the Cython-based runtime code generation by a simpler framework.

https://neurostars.org/t/gsoc-2025-project-8-brian-simulator-replace-brians-just-in-time-compilation-mechanism-350h/32066

GSoC 2025 Project #8 Brian Simulator :: Replace Brian's just-in-time compilation mechanism (350h)

Mentors: Marcel Stimberg <[email protected]>, Dan Goodman <[email protected]>, Benjamin Evans <[email protected]> Skill level: Intermediate/Advanced Required skills: Python, C++ Time commitment: Full time (350 hours) Forum for discussion About: Brian is a clock-driven spiking neural network simulator that is easy to learn, highly flexible, and simple to extend. Written in Python, it allows users to describe and run arbitrary neural and synaptic models without needing to ...

Neurostars

2 Brian2Wasm: simulations in the browser (Full-time or part-time)

πŸ€” Brian's code generation pipeline can generate C++ code, and the emscripten toolchain can compile this code into WebAssembly. We have recently created a proof-of-concept, brian2wasm, showing the feasibility of this approach.

πŸ‘‰ The aim of this project is to turn this proof-of-concept into a convenient tool for the wider community.

https://neurostars.org/t/gsoc-2025-project-7-brian-simulator-brian2wasm-simulations-in-the-browser-175-350h/32065

1 Serialization and deserialization for Brian simulator models (part-time)

πŸ€” Currently, the Brian simulator lacks advanced support for complex simulation pipelines that require storing or restoring the state of simulations (e.g., checkpointing) or saving a complete network architecture to a file.

πŸ‘‰ The aim of this project is to improve and consolidate Brian's the existing tools for model and state (de)serialization to overcome their current limitations.

https://neurostars.org/t/gsoc-2025-project-idea-6-brian-simulator-serialization-and-deserialization-for-brian-simulator-models-175-h/31999

GSoC 2025 Project Idea #6 Brian Simulator :: Serialization and deserialization for Brian simulator models (175 h)

Mentors: Benjamin Evans <[email protected]>, Dan Goodman @d.goodman <[email protected]>, Marcel Stimberg @mstimberg <[email protected]> Skill level: Novice/intermediate Required skills: Python, C++ Time commitment: Part-time (175 hours) Forum for discussion About: Brian is a clock-driven spiking neural network simulator that is easy to learn, highly flexible, and simple to extend. Written in Python, it allows users to describe and run arbitrary neural and synaptic models...

Neurostars

πŸŽ‰ The winter is coming to an end in the Northern Hemisphere, and we are getting geared up to participate in this year's Google's Summer of Code program (thanks as always to @INCF for taking us under their umbrella). If you are interested in contributing (either full- or part-time) to the Brian simulator over the summer – and getting paid for it –, this might be for you!

https://briansimulator.org/posts/2025/gsoc-2025/

#SummerOfCode #GSoC2025

A short thread with our 3 project proposals below πŸ‘‡

GSoC 2025

We are very happy that the INCF has been again selected to be a mentor organization for the Google Summer of Code (β€œa global, online program focused on bringing new contributors into open source softw

The Brian spiking neural network simulator

In the spirit of end-of-year celebrations, we've just released Brian 2.8 🌠
Alongside the usual helping of small improvements and fixes, this release comes with an important performance improvement for random number generation in C++ standalone mode.

Head to the release notes for more details! https://brian2.readthedocs.io/en/2.8.0/introduction/release_notes.html

#FOSS #ComputationalNeuroscience #Python

Release notes β€” Brian 2 2.8.0 documentation

Tootprint! πŸ“„ πŸ€“ I am happy to share my latest preprint (along with some code!), a collaboration with Claudia Cusseddu and Julijana Gjorgjieva in the Gjorgieva lab. https://www.biorxiv.org/content/10.1101/2024.10.12.618014

Inhibitory neurons keep neural circuits stable, preserving the inhibition-stabilized dynamical regime we all like. But they can do more than just provide a blanket of inhibition – they can also form specific excitatory-inhibitory (E/I) connection patterns.

In this work we consider simple inhibitory spike-timing dependent plasticity (STDP) rules (the kind you can implement with a few lines of code in Brian2), and find out that one can achieve both network stabilization and the formation of E/I structures!

Depending on the STDP rule, we get either mutual connections (connecting to excitatory neurons that also connect to them) or lateral connections (targeting excitatory neurons without reciprocal connections). This selectivity depends on how we tune the rule and on the STDP shape. 1/3

It's release day again πŸš€

We are happy to announce the release of Brian 2.7 and Brian2CUDA 1.0a5. Apart from the usual mix of small improvements and bug fixes, the main new feature is the compatibility with the upcoming @numpy 2.0 release – Brian will not hold you back from upgrading to it on day 1 😊

For more details, head over to https://brian2.readthedocs.io/en/2.7.0/introduction/release_notes.html

#FOSS #ComputationalNeuroscience #Python

Release notes β€” Brian 2 2.7.0 documentation