Mouse retinal specializations reflect knowledge of natural environment statistics

Abstract

SummaryPressures for survival drive sensory circuit adaption to a species’ habitat, making it essential to statistically characterise natural scenes. Mice, a prominent visual system model, are dichromatic with enhanced sensitivity to green and UV. Their visual environment, however, is rarely considered. Here, we built a UV-green camera to record footage from mouse habitats. We found chromatic contrast to greatly diverge in the upper but not the lower visual field, an environmental difference that may underlie the species’ superior colour discrimination in the upper visual field. Moreover, training an autoencoder on upper but not lower visual field scenes was sufficient for the emergence of colour-opponent filters. Furthermore, the upper visual field was biased towards dark UV contrasts, paralleled by more light-offset-sensitive cells in the ventral retina. Finally, footage recorded at twilight suggests that UV promotes aerial predator detection. Our findings support that natural scene statistics shaped early visual processing in evolution.Lead contactFurther information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Thomas Euler (thomas.euler@cin.uni-tuebingen.de)

Ancestral circuits for vertebrate colour vision emerge at the first retinal synapse

“…To our knowledge, a similarly specific visuoecological purpose of UV-vision in Drosophila remains unknown. More generally, UV-light can be highly informative about edges in space, as it tends to accentuate objects' silhouettes against bright backgrounds (50)(51)(52).…”

Circuit mechanisms for colour vision in zebrafish

“…In contrast, almost any solid overhead object blocks out the skylight, and thus appears as an approximately achromatic silhouette (Figure 1B). Both this effect, and spectral attenuation with distance mentioned above, do also occur above the water -in this case, however, they are usually much less obvious because air does not strongly attenuate light with (short) distance: For example, we can usually tell the hue of canopy seen from below, because it is well-illuminated from light reflecting off the ground 34 . Overall, vertical gradients in spectral variance of natural light have meant that colour-computing circuits are often found in specific eye positions examining specific regions of visual space.…”

“…A conceptually related UV-silhouette effect is also available for terrestrial vision where this background subtraction works well, for example, for removing clouds from the sky, but less obviously so, for example, for delineating an approaching bird against the treetops 34,42 . In fact, in addition to serving colour vision per se, differentially combining 'visible light' and UV signals at a circuit level (discussed below) could potentially serve to accentuate achromatic contrasts 34 (Figure 1G).…”

“…A conceptually related UV-silhouette effect is also available for terrestrial vision where this background subtraction works well, for example, for removing clouds from the sky, but less obviously so, for example, for delineating an approaching bird against the treetops 34,42 . In fact, in addition to serving colour vision per se, differentially combining 'visible light' and UV signals at a circuit level (discussed below) could potentially serve to accentuate achromatic contrasts 34 (Figure 1G). Beyond silhouette detection or potential contrast enhancement, the same UV-scatter can also be exploited at very short distances (millimetres) to highlight some of the particles that scatter UVlight in the first place: many small aquatic animals, including larval zebrafish, feed on approximately transparent microorganisms that, however, appear as tiny UV-bright objects when illuminated by the sun, ready for the taking 40,43 (Figure 1D; see also Figure 5J,K).…”

Natural Image Statistics for Mouse Vision

“…This suggests that the dorsovental functional division of the mouse retina is not optimal for maximizing the bandwidth of information transmission. Factors besides the coding efficiency, such as visual behavioral requirements, will thus need to be considered to fully explain the characteristic organization of the mouse retina.While natural image statistics have been extensively studied thus far [1,21], those outside the spectral domain of human vision remain to be fully explored [2,18,22]. Here we thus developed a multi-spectral camera system to sample high-quality images that spectrally match the mouse photopic vision, and analyzed the first-and second-order statistics of the UV and green image data sets.…”