Dynamics of multistable states during ongoing and evoked cortical activity

Area of Science:

• Neuroscience
• Computational Neuroscience
• Systems Neuroscience

Background:

• Cortical circuits dynamics evolve through temporal sequences of metastable states.
• Metastability is crucial for sensory coding, memory, and decision-making, but its network mechanisms remain unclear.
• Existing models often assume external stimuli trigger state sequences.

Purpose of the Study:

• Investigate network mechanisms underlying metastable state sequences in cortical circuits.
• Explore spontaneous generation of state sequences without external stimulation.
• Develop a recurrent spiking network model explaining observed neural dynamics.

Main Methods:

• Analysis of multielectrode recordings from the gustatory cortex of alert rats.
• Development and simulation of a recurrent spiking network model.
• Comparison of model predictions with experimental data, including responses to external stimuli.

Main Results:

• Observed spontaneous sequences of metastable states in the absence of overt sensory stimulation.
• Identified single-neuron multistability, where neurons exhibit multiple firing rates across states.
• Model successfully reproduced spontaneous state sequences and single-neuron multistability.
• Model predicted and experimental data confirmed quenching of multistability into bistability upon stimulation.

Conclusions:

• A recurrent spiking network model can explain both spontaneous and evoked cortical dynamics.
• The model provides a mechanistic framework for understanding metastable state sequences and single-neuron multistability.
• Findings challenge existing network models and offer insights into neural computation.