---
language:
- en
pipeline_tag: tabular-classification
tags:
- Computational Neuroscience 
license: mit
---

## Model description
The model is trained on recordings from 11 mice in the V1, SC, and ALM brain regions using Neuropixels probes.
Each recording was labeled by at least two independent annotators, with different combinations of labelers, achieving an 80% agreement rate.
This model utilizes a subset of metrics that are computationally efficient while maintaining robust classification performance.

# Intended use
Used to identify Noise clusters automatically in SpikeInterface.

# How to Get Started with the Model
This can be used to automatically identify SUA units in spike-sorted outputs. If you have a sorting_analyzer, it can be used as follows:

``` python
    from spikeinterface.curation import auto_label_units
    labels = auto_label_units(
        sorting_analyzer = sorting_analyzer,
        repo_id = "AnoushkaJain3/noise_neural_classifier_lightweight",
        trusted = ['numpy.dtype']
    )
```

## 📜 Citation  

If you find [UnitRefine](https://github.com/anoushkajain/UnitRefine) models useful in your research, please cite the following DOI:  
**[10.6084/m9.figshare.28282841.v2](https://doi.org/10.6084/m9.figshare.28282841.v2)**.  

We will be releasing a **preprint soon**. In the meantime, please use the above DOI for referencing.  

## 🔗 Resources  

- **GitHub Repository:** [UnitRefine](https://github.com/anoushkajain/UnitRefine)  
- 📖 **SpikeInterface Tutorial – Automated Curation:**  
  [View Here](https://spikeinterface.readthedocs.io/en/latest/tutorials_custom_index.html#automated-curation-tutorials)  

UnitRefine is **fully integrated with SpikeInterface**, making it easy to incorporate into existing workflows. 🚀  


# Authors

Anoushka Jain and Chris Halcrow