Skip to content

Poirazi-Lab/dendrify

Repository files navigation

Dendrify

Introducing dendrites to spiking neural networks

Documentation Status Contributor Covenant

Neuronal dendrites play a crucial role in shaping how individual neurons process synaptic information, yet their contributions to network-level functions remain largely underexplored. While current spiking neural networks (SNNs) often oversimplify or neglect essential dendritic properties, circuit models with morphologically detailed neuron representations are computationally expensive, limiting their practicality for large-scale network simulations.

To address these challenges, we introduce Dendrify, a free and open-source Python package designed to work seamlessly with the Brian 2 simulator. Dendrify enables users to generate reduced compartmental neuron models with biologically relevant dendritic and synaptic properties using simple commands. These models strike a good balance between flexibility, performance, and accuracy, making it possible to study the impact of dendrites on network-level functions.

https://github.com/Poirazi-Lab/dendrify/assets/30598350/b6db9876-6de4-458a-b27e-61d4edd360db

If you use Dendrify for your published research, we kindly ask you to cite our article:

Pagkalos, M., Chavlis, S., & Poirazi, P. (2023). Introducing the Dendrify framework for incorporating dendrites to spiking neural networks. Nature Communications, 14(1), 131. https://doi.org/10.1038/s41467-022-35747-8

Documentation for Dendrify can be found at https://dendrify.readthedocs.io/en/latest/

The project presentation for the INCF/OCNS Software Working Group is available on google drive.