Inferring neuronal network functional connectivity with directed information.

Area of Science:

• Neuroscience
• Computational Neuroscience
• Systems Neuroscience

Background:

• Inferring neural circuit connectivity from large-scale neuronal activity is a significant challenge.
• Existing methods often struggle with nonlinear relationships and require extensive prior knowledge.

Purpose of the Study:

• To develop and validate a data-driven method for inferring directed information and synaptic connections from neural spike trains.
• To identify both linear and nonlinear relationships between neurons for a more comprehensive understanding of neural circuits.

Main Methods:

• Utilized context tree maximizing (CTM) to estimate directed information (DI) from neural spike trains.
• Applied the CTM-DI method to simulated conductance-based networks and voltage-sensitive dye recordings from Aplysia.

Main Results:

• The CTM-DI method reliably identified circuit structures in simulations and inferred properties from Aplysia buccal ganglion recordings.
• The method proved robust against synaptic plasticity and capable of distinguishing excitatory and inhibitory connections.

Conclusions:

• The CTM-DI method offers a systematic tool for mapping neural network connectivity and tracking changes in network structure.
• This approach provides insights into how learning-induced modifications are distributed within neural networks and can detect network structures mediating motor patterns.