University of Crete (UOC)
Incorporating dendritic mechanisms in spiking neural networks through the Dendrify framework
Abstract
dc:descriptionComputational modeling has been indispensable for understanding how subcellular neuronal features influence neural network operations. However, the role of numerous dendritic mechanisms in brain functions remains largely unknown. This knowledge gap persists partly because existing tools are unsuitable for developing realistic and efficient network models that account for dendrites. Current spiking neural networks (SNNs), while efficient, are often overly simplistic, overlooking essential dendritic properties. On the other hand, circuit models with morphologically and biophysically detailed neurons are complex and computationally costly, making them impractical for large-scale simulations. To bridge the gap between these two approaches we introduce Dendrify. Dendrify is a theoretical framework for developing reduced compartmental neuron models with simplified, yet biologically realistic dendritic and synaptic integrative properties. It builds upon established theoretical work and is easily adaptable across various programming languages and platforms. For better accessibility, it is also available as a free, open-source Python package that works seamlessly with the Brian 2 simulator. Through user-friendly commands, Dendrify automatically generates reduced compartmental neurons, managing all necessary equations and parameters for Brian simulations. Meanwhile, it leverages Brian’s powerful features such as ease of use, GPU acceleration and wide compatibility. The reduced models developed with Dendrify strike a balance between flexibility, efficiency, and biological accuracy, enabling the exploration of dendritic contributions to network functions. By integrating dendritic computations into spiking neural networks, Dendrify may not only enhance our understanding of neuronal and circuit functions but also pave the way for developing more realistic and powerful neuro-inspired systems and algorithms.
Degree
thesis:*- Grantor dc:publisher
- University of Crete (UOC)
- Year dc:date
- 2024
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Pagkalos, Michail
- Πάγκαλος, Μιχαήλ
Subjects
dc:subject × 16Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/57096