Recent study findings may prove previous estimations of retinal cell communication, low-light vision

In an exclusive interview with Optometry Times, Yale School of Medicine researchers Yao Xue, PhD; Z. Jimmy Zhou, PhD; and Seunghoon Lee, PhD; and Co-director of the Yale Legacy Tissue Donation Program, Pathology Marcello DiStasio, MD, PhD, discussed findings from their recent groundbreaking research into retinal circuitry, specifically how bipolar cells in the retina process visual information. The study, “A hierarchical electrical synaptic circuit mechanism for integrative parallel visual processing in the retina” published in Neuron, found that information channels in the retina are more integrated than previously thought.

Traditionally, retinal processing has been conceptualized as occurring through parallel, largely independent channels. Photoreceptors (rods and cones) capture light and pass signals to bipolar cells, which then relay information to downstream ganglion and amacrine cells. In the mammalian retina—here, primarily mouse retina—there are over 15 distinct types of bipolar cells, each thought to extract particular visual features and perform specialized tasks within separate processing streams.

Xue, a postdoctoral scholar in Zhou’s lab, describes how his thesis work challenges this long-standing view. Using physiological approaches, he examined how different bipolar cell types handle visual information. Contrary to the classical model of independent parallel channels, his findings reveal that bipolar cells engage in coherent, network-level processing. This is mediated not only by classical chemical synapses but also by electrical synapses—gap junctions—linking different bipolar cell types. While electrical synapses are known in neuroscience, they are difficult to detect and had not been directly demonstrated as functionally organizing bipolar cell networks in this way.

Zhou elaborated that the presence of these electrical synapses between bipolar cells forms a network across what were previously considered independent channels. This introduces an additional level of computation at the very first synapse in the visual system (from photoreceptors to bipolar cells). Within this network, some bipolar cell channels act as “driver” channels and others as “passive” channels, suggesting a hierarchy or weighting of influence among parallel pathways. This reframes how early visual processing may integrate features before information ever reaches the brain’s higher visual centers.

The work extends beyond animal models. With the collaboration of the Yale Legacy Tissue Donation Program—a rapid research autopsy initiative that provides high-quality post-mortem human tissue—Yao and colleagues were able to perform electrophysiological recordings in whole-mount human retina within 2–3 hours after death. Remarkably, the retinal neurons remained functional, allowing them to test whether the network-level bipolar cell processing seen in mice is conserved in humans. The confirmatory findings in human retina strongly suggest that integrated, gap junction–mediated processing by bipolar cell networks is a general and conserved feature of mammalian vision, enabled by unique access to fresh human tissue and a multidisciplinary collaboration.