Synchronous firing and higher-order interactions in neuron pool.

Area of Science:

• Computational Neuroscience
• Statistical Physics
• Information Geometry

Background:

• Neuronal populations exhibit complex firing patterns, including synchronous firing.
• Understanding the underlying mechanisms of neuronal synchrony is crucial for neuroscience.
• Independent neuron firing typically results in concentrated activity distributions due to the law of large numbers.

Purpose of the Study:

• To investigate the stochastic mechanism of synchronous firing in neuronal populations using information geometry.
• To identify the role of higher-order neuronal interactions in generating widespread activity distributions.
• To explore how neuronal networks can exhibit synchronized activity and quiescence.

Main Methods:

• Analysis of the probability distribution q(r) of neuronal activity (r), the fraction of firing neurons.
• Information geometry framework to study neuronal interactions.
• Modeling a simple neural network with common overlapping inputs.

Main Results:

• Higher-order neuronal interactions, irreducible to pairwise correlations, are proven to exist in synchronous firing.
• Widespread activity distributions, particularly those with two peaks, indicate synchronous firing and quiescence.
• Pairwise and third-order interactions alone cannot explain the observed widespread distributions, highlighting the necessity of higher-order interactions.

Conclusions:

• Higher-order stochastic interactions are necessary to generate widespread neuronal activity distributions observed in synchronous firing.
• A model with common overlapping inputs demonstrates the generation of higher-order stochastic interactions and widespread activity.
• The findings provide insights into the complex dynamics of neuronal populations and information processing.