Neuroline

A unified pipeline for neurophysiology data processing

Neuroline is a wrapper that brings together powerful, established neuroscience software packages into a single pipeline. Rather than reinventing the wheel, it provides a streamlined interface to combine electrophysiology preprocessing, spike sorting, two-photon imaging data processing, video analysis, and behavioural tracking and exports everything into the NWB (Neurodata Without Borders) standard.

Installation & Documentation

Please see the documentation for the full installation guide and how to get started. In short:

1. Create an environment using conda

conda create -n neuroline python=3.11 pip
conda activate neuroline

2. Install PyTorch matching your GPU's compute capability, e.g., for CUDA 13.0:

pip install torch torchvision torchaudio

or for CUDA 12.6:

pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu126

3. Install Neuroline

pip install neuroline

or pip install neuroline[cu126] (for example) if you used a custom index url for PyTorch.

Dependencies

Neuroline is essentially a wrapper around many different software packages, the chief among which:

• Spikeinterface
• Kilosort
• Phy
• Open Ephys Python Tools
• Suite2P
• DeepLabCut
• FaceMap
• Neuroconv
• pyNWB

Compatibility

For electrophysiology, Neuroline currently supports data recorded with Neuropixels 1.0 and 2.0 probes and Rhythm interface-based headstages using the Open Ephys software, stored in either binary or NWB format.