Inferring Cortical Variability from Local Field Potentials.

Area of Science:

• Neuroscience
• Computational Neuroscience
• Systems Neuroscience

Background:

• Sensory neuron responses exhibit significant trial-to-trial variability, not fully explained by stimulus properties alone.
• Understanding this variability is key to deciphering the complex computations within sensory cortical networks.

Purpose of the Study:

• To investigate the direct link between network activity, specifically local field potentials (LFPs), and the variability of single-neuron spiking responses in the visual cortex.
• To develop a predictive model incorporating LFPs to better understand sensory cortical function.

Main Methods:

• Simultaneous electrophysiological recordings of single-neuron spikes and LFPs from the middle temporal (MT) area of macaque visual cortex.
• Presentation of continuous optic flow stimuli.
• Application of a maximum likelihood-based modeling framework to predict spiking responses using stimulus, LFP, and neuronal activity data.

Main Results:

• MT neuron responses showed strong phase-locked relationships with gamma (approx. 40 Hz) and delta (1-4 Hz) oscillations in LFPs.
• LFP-modulated responses were largely complementary to stimulus-driven and other neuron-driven modulations.
• LFP-based predictions accounted for nearly half of the trial-to-trial spiking variability and accurately captured noise correlations between neurons.

Conclusions:

• Network activity, reflected in LFPs, plays a critical role in shaping sensory neuron responses and their variability.
• Integrating network activity measures significantly improves the prediction of neural responses compared to stimulus-based models alone.
• This work provides a foundation for understanding how sensory cortex integrates sensory information with ongoing network dynamics and cognitive variables.