Kernel-based LFP estimation in detailed large-scale spiking network model of mouse visual cortex.

Area of Science:

• Computational neuroscience
• Neural modeling
• Brain signal analysis

Background:

• Simulations of large-scale neural activity are vital for understanding brain function.
• Estimating measurable brain signals, such as the local field potential (LFP), from these simulations is critical for bridging models and experiments.
• Current methods for accurate LFP simulation often require computationally intensive, highly detailed models, limiting practical applications.

Purpose of the Study:

• To demonstrate a kernel-based method for accurate and efficient LFP estimation from large-scale neural network models.
• To analyze the contributions of different neuronal populations to the LFP in a mouse primary visual cortex (V1) model.
• To investigate the role of external synaptic inputs versus local activity in shaping V1 LFPs.

Main Methods:

• Utilized a kernel-based method to estimate LFPs from a detailed multicompartmental network model of the mouse V1.
• Simulated responses to visual stimuli, including drifting gratings and full-field flashes.
• Applied the method to disentangle contributions of neuronal populations and synaptic inputs to the LFP.

Main Results:

• The kernel method accurately and efficiently estimated LFPs in the V1 network model.
• External synaptic inputs, specifically feedback from lateromedial visual areas and thalamic afferents, were found to dominate the V1 LFP.
• Local synaptic activity from V1 neuronal families contributed marginally to the LFP, and correlations could mask this finding in experimental data.

Conclusions:

• The kernel method is a powerful and accurate tool for LFP estimation in complex neural network models.
• This method can provide novel insights into the neural mechanisms underlying measurable brain signals.
• External synaptic inputs play a dominant role in shaping LFPs in the mouse V1, challenging previous assumptions about local circuit contributions.