Neural circuits in the mouse retina...


Home » The Institute » News & Events » News » News Article

Neural circuits in the mouse retina support color vision in the upper visual field

In the last decade, mice have become an important model system for biomedical research. Many ideas of how the healthy visual system works stems from studies in mice. Recently, a paper by Szatko, Korympidou and colleagues in Nature Communications answered a fundamental question about mouse vision – how do mouse eyes convey color information to the brain?

Top left: Visualization of the UV/green sensitivity gradient across the flattened retina. Most cones in the upper retina are sensitive to green light, while most cones in the lower retina are only sensitive to UV light. Top right: Schematic of the studied cell types in the retina. Cone photoreceptors detect light, bipolar cells transmit the signal further to the ganglion cells, which send the information to the brain. Top middle: Main results of the study. Each point on these maps is a cell, and its color illustrates its color preference. The upper map shows the cells' preference for center stimuli, the lower map shows the surround preference. Bottom: As only the lower part of the retina (marked in red) has cells with color opposing center/surround responses, likely only this area of the retina allows color discrimination. Because the lower retina looks at the visual world above the animal, this would lead to "sky-only" color vision (illustrated in the photograph on the left).

In the mammalian retina, two classes of photoreceptors detect light: rods and cones. Rods are traditionally thought to mediate black-and-white night vision and are most sensitive to green light. Cones, in contrast, are active during the day and are sensitive to different wavelengths or “colors”. For example, the human retina has three cone types, roughly preferring red, green or blue light. Neural circuits in the retina compare the cone signals, and this information is used by the brain to decode the color of a specific object. For example, a neuron can detect a blueberry in front of green leaves by comparing blue signals in a central location (“center”) with green signals in a surrounding location (“surround”). This neuron would be classified as center-surround color-opponent.

In contrast to primates, mice like most other mammals only have two cone types, which are sensitive to ultraviolet (UV) and green light. The majority of cones are ”green”-cones, and they show an interesting pattern of sensitivity: They are indeed green-sensitive in the upper part of the retina, which “looks” at the ground, but despite their name become increasingly more UV-sensitive towards the lower part, which “looks” at the sky. Thus, a given region of the retina is predominantly sensitive to one “color” and, therefore, mice should be colorblind. In contrast to this prediction, a recent behavioral study showed that mice can discriminate colors, with the best performance in the upper visual field – indicating that they can make color comparisons in this region. The study by Szatko, Korympidou and colleagues investigated which tricks the mouse retina uses to enable this behavior.

The authors systematically recorded the isolated mouse retina’s activity to colored light by “flattening it out” between a special light projector and a microscope. The projector then played patterns of UV and green light to the retina, thereby exciting retinal neurons. The light patterns tested the color preference of the neurons: The center stimulus was a small dot of light, covering only the recorded cells, while the surround stimulus was a ring of light that did not reach the recorded cells, but only their immediate neighborhood. To measure the neurons’ activity, the researchers introduced a fluorescent molecule into the tissue that lights up whenever the neurons are active, so the microscope provided a live view of the retinal activity.

First, the authors focused their microscope on cones. When the center stimulus was presented in UV and green, the authors saw the expected response pattern: Cones in the lower retina responded more to UV, while the cones of the upper retina preferred green. Also in line with the UV/green-gradient, the same cones in the upper retina preferred the green surround stimulus. Therefore, at the level of the cones, the upper retina showed no signs of color opponency. In contrast, the UV-sensitive cones in the lower retina showed a strong preference for green in their surround – despite the lack of green-sensitive cones in this area.

What could be the source of such a preference in the lower retina? One possibility are rods, which are generally green-sensitive and could signal about surrounding green light to cones via a specific interneuron-type, the horizontal cells. This hypothesis was confirmed when the researchers blocked the horizontal cells: Without active horizontal cells, cones in the lower retina showed no response to the green surround. This indicated that in the lower retina rods and cones work together to establish center-surround color opponency, thereby laying the foundation of mouse color vision.

After establishing that only cones in the lower retina show color-opponent signals, while cones from the upper retina do not, the authors followed this regional difference further through the retina. They next imaged the color responses of bipolar cells, which transmit cone signals to downstream retinal neurons, and of ganglion cells, which send the retinal output to the brain. Both experiments revealed that cells of the upper retina were mostly colorblind, while many cells in the lower retina displayed center-surround color opponent responses. Interestingly, further analysis showed that the ganglion cells process the color signal before sending it to the brain, presumably to enhance the color information for downstream brain areas.

In summary, this work shows how color processing arises along the layers of the retina to allow color discrimination. In the mouse, everything starts with a regional difference in cone color preference: The ground-focused upper retina is green-dominated, while the sky-gazing lower retina prefers UV. Then, UV-cones and green-sensitive rods of the lower retina act together to create color opponency, the neural basis of mouse color vision.

The UV/green sensitivity gradient is also present in other mammals like guinea pig and rabbit. Until now, it was believed that these species were colorblind in their lower retina. However, this work shows that rods and cones together can still extract color information in those cases. Further studies will be needed to see if this has consequences on a behavioral and ecological level. At least in humans, a similar rod-cone interaction might exist: In blue-cone monochromatic patients, only the blue sensitive cone type is functional. Zrenner and colleagues showed that these patients can still see colors, likely by using their rods as the required second color channel. Therefore, the mechanism described by Szatko, Korympidou and colleagues likely contributes to understanding the foundations of color vision in mammals in general.

Original Publication under

The Institute thanks Jan Lause for the compiliation of the text above and Katrin Franke, Thomas Euler and Philipp Berens for the support to stengthen the visibility of the Institute for Ophthalmic Research.