Self-Organization of Nonlinearly Coupled Neural Fluctuations Into Synergistic Population Codes

Area of Science:

• Computational neuroscience
• Neural coding
• Brain dynamics

Background:

• Correlated neural fluctuations significantly impact neural population coding.
• The origins and computational consequences of these fluctuations in neural circuits are not fully understood.
• Resolving nonlinear coupling between fluctuations and system dynamics is a key challenge.

Purpose of the Study:

• To investigate the emergence of synergistic neural population codes from intrinsic dynamics of correlated neural fluctuations.
• To model realistic nonlinear noise coupling in spiking neuron networks.
• To understand how differential and noise correlations interact to produce synergistic codes.

Main Methods:

• Developed a neural circuit model with realistic nonlinear noise coupling of spiking neurons.
• Analyzed spatial correlation patterns in a bump attractor network.
• Investigated the dynamical regimes governing the interplay of correlations.

Main Results:

• A rich repertoire of spatial correlation patterns emerged naturally.
• Identified dynamical regimes where differential and noise correlations yield synergistic codes.
• Observed that negative correlations can induce stable bound states between two bumps, a novel finding for firing rate models.
• Demonstrated noise-induced effects enhancing working memory capacity and spatiotemporal multiplexing.

Conclusions:

• Correlated neural fluctuations play a crucial role in synergistic neural population coding.
• Negative correlations and noise-induced effects offer computational advantages for neural systems.
• This work provides a dynamical framework for understanding neural fluctuations and their role in cortical computations, potentially explaining working memory phenomena.