Stochastic nature of precisely timed spike patterns in visual system neuronal responses

Area of Science:

• Neuroscience
• Computational Neuroscience
• Systems Neuroscience

Background:

• The representation of cognitive and psychological information by single neuron responses remains unclear.
• A hypothesis suggests precisely timed spike patterns carry information beyond spike count or latency.

Purpose of the Study:

• To investigate if precisely timed spike patterns (1-ms precision) in LGN and V1 neurons carry unique information.
• To develop and test a statistical model that accounts for temporal spike patterns in neural responses.

Main Methods:

• Analysis of stimulus-elicited spike trains from LGN and V1 neurons in awake rhesus monkeys.
• Information theoretical analysis to quantify information carried by precisely timed spike patterns.
• Development and application of a novel stochastic model incorporating peristimulus time histogram, interspike interval, and spike count distributions.

Main Results:

• Numerous precisely timed spike triplets and quadruplets were identified in neural responses.
• Information theoretical analysis indicated that precisely timed spike patterns did not carry information beyond spike count.
• The novel stochastic model successfully predicted observed spike patterns, demonstrating their potential to arise by chance.
• Deviations from a Poisson distribution in spike count significantly altered expected precisely timed spike patterns.

Conclusions:

• The precise temporal structure of stimulus-elicited neuronal responses in LGN and V1 can occur by chance.
• Fine temporal structures are interpretable only after considering coarse temporal statistics, such as firing rate.
• The findings challenge the notion that precisely timed spike patterns inherently encode distinct information in early visual pathways.