Developmental self-construction and -configuration of functional neocortical neuronal networks

Area of Science:

• Computational neuroscience
• Developmental biology
• Systems neuroscience

Background:

• Neural circuits require specific configurations for postnatal behaviors.
• Understanding the integration of genetic and cellular processes in network formation is limited.

Purpose of the Study:

• To simulate and demonstrate how a competitive co-operative ('winner-take-all', WTA) network architecture can develop from a single precursor cell.
• To explore the role of a simplified gene regulatory network in guiding developmental processes.
• To investigate how homeostatic unsupervised learning shapes synaptic weights based on input patterns.

Main Methods:

• Detailed computational simulations of neural network development.
• Modeling of a precursor cell with a gene regulatory network controlling mitosis, differentiation, migration, neurite outgrowth, and synaptogenesis.
• Application of homeostatic unsupervised learning with wave-like input patterns to established axonal connections.

Main Results:

• Demonstrated the autonomous developmental sequence from a single precursor cell to a self-calibrated WTA network.
• Showcased how genetic programming and activity-dependent learning interact to form functional neural architectures.
• Validated simulation outcomes against biological data.

Conclusions:

• Genetically directed developmental sequences can autonomously generate complex, functional neural network architectures like WTA.
• The interplay between genetic regulation and homeostatic learning is crucial for self-calibration of neural circuits.
• This model provides insights into the fundamental principles of neural development and network self-organization.