PymoNNto

PymoNNto provides a Python framework for constructing and simulating modular, brain-inspired neural network models for computational neuroscience research.

Key Features:

• Modular network class: A network class that contains multiple neuron- and synapse-groups for structured model specification.
• Exchangeable modules: Group behavior is defined through interchangeable modules that can be customized per group.
• Library interoperability: Implementations can use Numpy and TensorFlow to perform computations on CPUs and GPUs.
• Differential-equation behavior modules: High-level behavior modules support differential equation–based implementations similar to those in Brian2.
• Minimal implementation restrictions: The framework imposes fewer restrictions on implementation and execution compared to Brian2 and NEST.
• Python integration: Core functionality is implemented in Python to enable flexible scripting and customization.

Scientific Applications:

• Computational neuroscience: Constructing and simulating brain-inspired neural networks for research studies.
• Model development: Creating complex and customizable neuron and synapse models for hypothesis testing.
• Algorithm prototyping: Implementing and evaluating custom neural dynamics and synaptic mechanisms.
• Brain-inspired computing exploration: Exploring novel neural computation and simulation approaches in brain-inspired computing.

Methodology:

Models use a network class with neuron- and synapse-groups whose behavior is specified by exchangeable modules; implementations may employ Numpy or TensorFlow for CPU and GPU computation and include high-level differential equation–based behavior modules analogous to Brian2.