Chaotic dynamics in spatially distributed neuronal networks generate population-wide shared variability.

Area of Science:

• Computational Neuroscience
• Systems Neuroscience
• Neural Dynamics

Background:

• Cortical neural activity shows high variability to repeated stimuli.
• This variability is shared across large neuronal populations and confined to a low-dimensional space.
• The origin of this widespread, shared neural variability remains largely unknown.

Purpose of the Study:

• To investigate the source of population-wide shared neural variability in the cortex.
• To analyze the dynamical regimes of spatially distributed excitatory and inhibitory neuronal networks.
• To determine if specific network dynamics can replicate experimentally observed neural variability patterns.

Main Methods:

• Analysis of dynamical regimes in simulated networks of excitatory and inhibitory neurons.
• Modeling spatially distributed neuronal networks with specific projection widths.
• Investigating the impact of globally correlated noisy inputs on network dynamics.

Main Results:

• Chaotic spatiotemporal dynamics were identified in networks mirroring cortical anatomical features (excitatory/inhibitory projection widths).
• These chaotic dynamics produced broadband frequency power in neural firing rates.
• The model exhibited distance-dependent and low-dimensional correlations, aligning with experimental data.
• Globally correlated noisy inputs were shown to induce rate chaos.

Conclusions:

• Spatiotemporal chaos in cortical networks provides a plausible mechanism for the observed shared neural variability.
• The findings suggest that network dynamics, specifically chaos, can account for low-dimensional, population-wide correlations in neural responses.
• This study offers a theoretical framework linking network structure and dynamics to emergent population-level neural activity patterns.