pycabnn

pycabnn constructs anatomically grounded neural network models by generating neuron placements and computing connectivity constrained by experimental anatomical properties.

Key Features:

• Anatomical modeling: Generates anatomically accurate cell positions and interconnections using extended geometrical structures such as axonal and dendritic morphology.
• Efficient algorithms: Computes mutual connectivity with algorithms capable of handling large-scale models, demonstrated on a cerebellar granular layer model with over half a million cells.
• Versatility: Adapts methods developed for the cerebellar granular layer to other neural circuit models.
• Scalability and parallelization: Scales to large networks and supports parallel computing environments for enhanced computational efficiency.

Scientific Applications:

• Biophysical mechanism analysis: Provides an anatomically accurate structural basis for simulations that investigate how biophysical mechanisms influence neural information processing.
• Cerebellar cortex modeling: Applied to constructing detailed models of the cerebellar granular layer network.
• General neural circuit modeling: Enables construction of anatomically based models for other neural systems to study network dynamics and interactions.

Methodology:

Generates cell positions and interconnections from extended geometrical structures (axonal and dendritic morphology), computes mutual connectivity with efficient algorithms demonstrated on large-scale (>500,000 cells) models, supports parallel execution, and is implemented in Python.