Rank Order Coding: a Retinal Information Decoding Strategy Revealed by Large-Scale Multielectrode Array Retinal...

Area of Science:

• Neuroscience
• Computational Neuroscience
• Vision Science

Background:

• Retinal ganglion cells (RGCs) are crucial for visual processing.
• Previous studies suggested RGC pairs encode spatial information via relative latencies.
• The population coding strategy of RGCs remains incompletely understood.

Purpose of the Study:

• To investigate how retinal ganglion cell populations encode visual scenes in mice.
• To test the hypothesis that concerted spiking across RGCs transmits visual information rapidly and efficiently.
• To compare population-based coding with individual RGC spike count and latency codes.

Main Methods:

• Simultaneous recording of light-evoked responses from hundreds of RGCs using a 4096-electrode multielectrode array.
• Pan-retinal recordings to capture population-level activity.
• Analysis of RGC spike timing, spike counts, and relative activities.

Main Results:

• No individual RGCs showed clear latency tuning to stimuli in mice.
• Significant visual information is encoded synergistically in the concerted spiking of large RGC populations.
• RGC population response, based on relative activities (ranks), is more informative than independent spike count or latency codes.
• The wave of first stimulus-evoked spikes across the population accurately indicates stimulus content.

Conclusions:

• In mice, RGC population coding relies on concerted spiking rather than individual RGC latency tuning.
• Relative activities and the wave of first spikes across the RGC population provide a fast and efficient visual code.
• This population coding strategy coexists with classical neural codes, enhancing visual information transmission speed and reliability at the retinal level.