Functional connectomics from neural dynamics: probabilistic graphical models for neuronal network of

Area of Science:

• Computational Neuroscience
• Systems Neuroscience
• Network Biology

Background:

• Understanding neuronal network dynamics is crucial for deciphering brain function.
• Existing methods often struggle to capture causal relationships and stimulus propagation within complex neural systems.

Purpose of the Study:

• To develop a novel approach for representing neuronal network dynamics as a probabilistic graphical model (PGM).
• To create a functional connectome that elucidates causal dependencies between neurons and stimuli.
• To validate the methodology using a model of the Caenorhabditis elegans nervous system.

Main Methods:

• Collecting time series of neuronal responses from a neuronal network model.
• Applying singular value decomposition (SVD) for low-dimensional projection of time-series data.
• Extracting dominant patterns to identify pairwise dependencies and construct the graphical model.

Main Results:

• A functional connectome was generated, effectively representing causal dependencies and stimulus propagation.
• The PGM successfully mapped known neuronal pathways in the C. elegans model.
• The approach demonstrated potential for identifying novel pathways and understanding network behavior.

Conclusions:

• The proposed PGM approach offers a powerful framework for analyzing neuronal network dynamics.
• This method provides insights into functional connectomics and causal inference in neural systems.
• The validated model aids in understanding neural circuit function and behavior, particularly in C. elegans.