Benefits of pathway splitting in sensory coding

Area of Science:

• Neuroscience
• Computational Neuroscience
• Sensory Systems Biology

Background:

• Mammalian retinas exhibit complex neural signal processing with ~20 types of retinal ganglion cells transmitting visual information.
• Early splitting of neural signals into parallel pathways, such as ON and OFF neurons, is common in sensory systems.
• The functional significance of this profuse early pathway divergence remains largely unexplained.

Purpose of the Study:

• To investigate the hypothesis that parallel pathway splitting, specifically ON-OFF neuron diversification, optimizes sensory information encoding.
• To compare the information transmission efficiency of an ON-OFF neural system versus an ON-ON system.
• To determine the role of firing rate constraints in the efficiency of ON-OFF versus ON-ON neural coding.

Main Methods:

• Computational modeling comparing information transmission in ON-OFF and ON-ON neural systems.
• Optimization of neural systems under constraints of maximal and mean firing rates.
• Analysis of coding efficiency and downstream linear decoding performance.

Main Results:

• The ON-OFF system transmits information as effectively as an optimized ON-ON system under maximal firing rate constraints.
• The ON-OFF system achieves this with significantly fewer average spikes, indicating higher metabolic efficiency.
• This efficiency advantage is comparable to decorrelation and is amplified for sparse, high-impact stimuli.

Conclusions:

• ON-OFF pathway splitting provides a more efficient neural code for sensory information, particularly concerning metabolic cost.
• The evolution of ON-OFF diversification in sensory systems may be driven by the need for energy conservation in environments with sparse, critical stimuli.
• ON-OFF systems offer superior information extraction for linear downstream decoders.