Analysis of ongoing dynamics in neural networks

Area of Science:

• Computational neuroscience
• Systems neuroscience
• Neural dynamics

Background:

• Spontaneous neural activity in the cerebral cortex displays complex spatio-temporal patterns without sensory input.
• Understanding these ongoing activities is crucial for deciphering the brain's computational principles.
• The exact mechanisms driving these phenomena remain largely unknown.

Purpose of the Study:

• To model the ongoing dynamics of generic neural networks with attractor states.
• To investigate the mechanisms underlying spontaneous cortical activity and state transitions.

Main Methods:

• Utilized a conductance-based neuron model for realistic neural network simulations.
• Analyzed the spatio-temporal dynamics of simulated neural activity, including up-states and down-states.
• Investigated the roles of excitatory and inhibitory inputs, synaptic depression, and potassium channels in state transitions.

Main Results:

• Modeled neural networks exhibiting distinct up-states and down-states in membrane potential.
• Identified up-states as spatially clustered patches that propagate through the network.
• Demonstrated that balanced excitatory and inhibitory inputs sustain up-states.
• Showed that synaptic depression and potassium channels differentially regulate transitions from up-states to down-states.
• Found that patch velocity is influenced by excitatory neuron firing frequency.

Conclusions:

• The switching dynamics between cortical states can arise from local network interactions.
• These findings provide insights into the constraints governing autonomous neural dynamics in the cortex.