Refractory sampling links efficiency and costs of sensory encoding to stimulus statistics

Area of Science:

• Neuroscience
• Computational Biology
• Sensory Systems

Background:

• Sensory neurons adapt information sampling based on environmental changes.
• The mechanistic basis of how stimulus statistics influence neural encoding is poorly understood.
• Photoreceptor function is critical for visual perception.

Purpose of the Study:

• To systematically analyze how stimulus statistics affect information encoding in sensory neurons.
• To investigate the role of photoreceptor microvilli refractoriness in information processing.
• To compare information sampling efficiency between naturalistic and artificial stimuli.

Main Methods:

• Intracellular recordings from fly photoreceptors (Drosophila melanogaster and Coenosia attenuata).
• Stochastic simulations using biophysically realistic photoreceptor models.
• Analysis of information capture, signal-to-noise ratio, and encoding efficiency.

Main Results:

• Photoreceptors sample more information from naturalistic light intensity time series (NS) compared to Gaussian white-noise (GWN) or other artificial stimuli.
• Higher signal-to-noise ratio and encoding efficiency were observed for NS, particularly during bursty contrast changes.
• Photoreceptor information capture is critically dependent on the stochastic refractoriness of microvilli.
• Microvilli refractoriness enhances encoding of intensity changes in naturalistic conditions but is energetically costly for artificial stimuli like GWN.

Conclusions:

• Information sampling in sensory neurons is more efficient for naturalistic stimulation due to mechanisms like microvilli refractoriness.
• The findings provide mechanistic insights into sensory neuron operation, design, and evolution.
• Understanding stimulus statistics is crucial for deciphering neural coding and perception.