Neural decoding · Decoding
torch_brain
A library for designing and training deep-learning models on neural population data.
- Active
- Early
- Research-focused
What problem does it solve?
Pretraining across sessions, subjects and datasets needs data structures and models built for heterogeneous spiking data.
Who is it for?
Researchers building neural foundation models
At a glance
- Primary use
- Neural decoding
- Licence
- Apache-2.0 · permissive
- Language
- Python
- Current release
- v0.2.0, published 7 Jul 2026
- Development activity
- Last commit 2 Oct 2026 · repository created 24 Jun 2024
- Maintainers
- neuro-galaxy
- GitHub stars
- 98 as of 3 Oct 2026
- Standards
- NWB
Repository facts are a snapshot read from the GitHub API on 3 Oct 2026; they are reported as found, not scored.
Supported modalities and use
Common use cases
- Multi-session pretraining
- Population decoding models
How it fits into a BCI stack
Hardware
2 tools
Acquisition
3 tools
Streaming
3 tools
Processing
6 tools
Decoding
torch_brain
Applications
2 tools
Data & standards
3 tools
- Analyse an ECoG speech datasetas Decoder
- Build an online motor decoderas Model
Sources
- 01
github.com/neuro-galaxy/torch_brain (opens in a new tab)
Open-source repositoryGitHubChecked against source 3 Oct 2026
Repository metadata read from the GitHub API.