Connectome-based reservoir computing with the conn2res toolbox.

Area of Science:

• Computational Neuroscience
• Artificial Intelligence
• Network Science

Background:

• Understanding how neural circuit connection patterns give rise to complex cognition and behavior is a central challenge in neuroscience.
• Advances in connectomics and artificial intelligence offer new avenues for investigating the computational capacity of biological brain networks.
• Reservoir computing provides a framework for approximating cognitive functions using high-dimensional, nonlinear dynamical systems.

Purpose of the Study:

• To introduce conn2res, an open-source Python toolbox for implementing biological neural networks as artificial neural networks.
• To enable researchers to explore structure-function relationships in brain networks by reconceptualizing function as computation.
• To facilitate novel research at the intersection of neuroscience and artificial intelligence.

Main Methods:

• Development of conn2res, a modular Python toolbox for creating artificial neural network models from biological connectomes.
• Inputting reconstructed connectomes from various techniques, including tract tracing and diffusion imaging.
• Imposing diverse dynamical systems, such as spiking neurons and memristive dynamics, onto the network models.

Main Results:

• conn2res supports arbitrary network architectures and dynamics, offering significant modularity and flexibility.
• The toolbox accommodates connectomes derived from multiple reconstruction methods.
• Researchers can utilize various dynamical systems, enhancing the versatility of network simulations.

Conclusions:

• conn2res provides a powerful platform for investigating brain network computation.
• The toolbox fosters a mechanistic understanding of how neural circuit structure relates to function.
• It opens new research possibilities by integrating connectomics, artificial intelligence, and computational neuroscience.