Pycabnn: Efficient and Extensible Software to Construct an Anatomical Basis for a Physiologically Realistic Neural...

Area of Science:

• Computational Neuroscience
• Biophysics
• Neuroinformatics

Background:

• Physiologically detailed neural network models are essential for understanding how biophysical mechanisms influence neural information processing.
• Accurate anatomical modeling, including cell placement and connectivity, is a fundamental prerequisite for constructing these complex network models.

Purpose of the Study:

• To present pycabnn, an open-source software tool specifically designed for generating the anatomical foundation of neural network models.
• To provide efficient algorithms for creating physiologically realistic cell positions and determining connectivity based on neuronal morphology.

Main Methods:

• Implementation of efficient algorithms within pycabnn for cell positioning and connectivity mapping.
• Utilizing extended geometrical structures like axonal and dendritic morphology for connection determination.
• Demonstration using a large-scale model of the cerebellar granular layer.

Main Results:

• pycabnn successfully generated anatomical models with over half a million cells and computed their mutual connectivity.
• The software demonstrated efficiency, completing tasks on a laptop within a reasonable runtime.
• pycabnn is capable of running in parallel computing environments for larger-scale simulations.

Conclusions:

• pycabnn is an efficient, user-friendly, and extensible Python-based tool for generating anatomical models for neural networks.
• Its capabilities are suitable for large-scale network simulations, such as the cerebellar granular layer.
• pycabnn is poised to be a valuable asset for future computational neuroscience studies.