State dependence of stimulus-induced variability tuning in macaque MT

Area of Science:

• Neuroscience
• Computational Neuroscience

Background:

• Behavioral states influence cortical response properties like firing rates and correlations.
• The precise mechanisms driving these state-dependent changes and their impact on stimulus decoding remain unclear.

Purpose of the Study:

• To investigate how behavioral states modulate the Fano factor (response variance-to-mean ratio) in cortical neurons.
• To determine the impact of state-dependent Fano factor tuning on stimulus discriminability.
• To explore the underlying sources of neural variability in different behavioral states.

Main Methods:

• Recorded activity of motion-sensitive neurons in the middle temporal cortex (MT) of macaques.
• Compared neural responses during alert fixation versus light opioid anesthesia.
• Analyzed firing rates, trial-to-trial variability (Fano factor), and temporal correlations in neural activity.
• Developed computational models to explain observed Fano factor distributions and their relation to gain fluctuations.

Main Results:

• Anesthesia lowered average spike counts but did not decrease trial-to-trial variability.
• Anesthetized neurons exhibited supra-Poisson Fano factors due to longer time-scale excitability fluctuations.
• Alert neurons showed higher mean firing rates and stimulus-dependent, largely sub-Poisson variability.
• The stimulus-induced variability tuning observed in the alert state was absent under anesthesia.
• A model suggested suppressed variance in gain fluctuations contributes to alert-state Fano factor tuning.

Conclusions:

• State-dependent modulation of Fano factor tuning significantly affects stimulus discriminability.
• Neural variability sources differ between alert and anesthetized states.
• Alertness facilitates enhanced stimulus discriminability through specific neural variability tuning and correlations.