Back

High-fidelity backpropagation through primate foveal cones

Wienbar, S. R.; Bryman, G. S.; Do, M. T. H.

2026-01-30 neuroscience
10.64898/2026.01.28.701353 bioRxiv
Show abstract

Primate vision has exceptionally high spatial acuity and contrast sensitivity. This performance originates in specialized photoreceptors of the fovea. These cones transduce light into electrical signals in the outer segment, and convey these signals to the presynaptic terminal for transmission. Backpropagating signals are also possible, as the terminal receives gap-junctional input from neighboring cones. Such signals could influence phototransduction itself. To test this idea, we recorded electrophysiologically from both ends of single cones dissociated from the macaque fovea. We found that backpropagation was effective despite the extreme slenderness and length of these cells. Backpropagation was also effective in a passive compartmental model, indicating that amplification by voltage-gated channels is not required. We then modeled mosaics of foveal cones coupled by terminal gap junctions. Despite faithful backpropagation of these inputs, they appear unlikely to influence phototransduction. Thus, even though foveal cones exhibit effective backpropagation, their encoding of visual information remains compartmentalized. SIGNIFICANCEHumans, like other primates, see a fineness of detail that eludes other mammals. This capability is used for tasks like reading and recognizing faces. It is lost in leading forms of vision impairment, such as age-related macular degeneration. Investigating it therefore provides insight into the origin of exceptional sensory performance while strengthening the foundation for preserving and restoring sight. This study examines cells that initiate high-acuity vision, the foveal cones, which produce electrical signals from light and send them forward. It reveals that electrical signals also travel effectively in reverse, from the site of transmission to that of production, and how the production of light responses can remain independent nonetheless.

Published in The Journal of Neuroscience (predicted rank #1) · training set

Matching journals

The top 3 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.