Circuit mechanisms for colour vision in zebrafish

Spectral inference reveals principal cone-integration rules of the zebrafish inner retina

“…Accordingly, in these non-mammalian lineages, the expectation is that up to tetrachromatic color vision should be possible based on stereotyped cone-opponent ancestral circuits that are specified during development, without a necessity for building additional spectral opponencies in the brain. In agreement, physiological recordings from retinal neurons in cone-tetrachromatic species, including turtles 18 and diverse species of fish, 9 , 19 , 20 , 21 , 22 , 23 consistently revealed a rich complement of complex spectral signals, including diverse spectral opponencies.…”

“…Of the three spectral axes that dominate the zebrafish inner retina ( Figures 5 I–5N and 7 A), those functionally linked with green- (RH2) and blue-cone (SWS2) circuits are unlikely to have a direct counterpart in mammals where these cone types are lost. 1 , 9 However, the third axis, formed by functional opposition of UV-cone circuits against red-, green-, and blue-cone circuits, may relate to one or multiple of the well-studied mammalian SWS1:LWS opponent circuits ( Figure 7 B). 50 , 51 …”

“…In contrast, most non-mammalian vertebrate lineages, including fish, amphibians, reptiles, and birds, retain the full complement of ancestral cone types based on four opsin-gene families: SWS1 (UV cones), SWS2 (blue cones), RH2 (green cones), and LWS (red cones). 1 , 9 , 10 , 11 Each of these four ancestral cones provides type-specific extracellular matrix proteins that developmental programs use to build cone-type selective circuits in the outer retina (e.g., zebrafish 12 , 13 , 14 and chicken 15 , 16 , 17 ). Accordingly, in these non-mammalian lineages, the expectation is that up to tetrachromatic color vision should be possible based on stereotyped cone-opponent ancestral circuits that are specified during development, without a necessity for building additional spectral opponencies in the brain.…”

“…However, what the dominant opponencies are and how they are built at the circuit level remains incompletely understood in any cone-tetrachromat vertebrate. 9 This is in part because already horizontal cells in the outer retina functionally interconnect and potentially retune cone types, 10 , 14 , 24 , 25 , 26 thus limiting the possibility of making inferences about spectral processing based on recordings from downstream neurons. To address this, we recently measured the in vivo spectral tuning of the synaptic outputs from the four cone types in larval zebrafish using spatially widefield but spectrally narrow flashes of light.…”

“…Accordingly, in larval zebrafish, already the cone output provides up to three axes of spectral opponency. 9 , 27 However, the opponent axis provided by UV cones was weak, which left its role in zebrafish color vision unclear. Moreover, in view of expected extensive mixing of cone signals in downstream circuits, 12 , 28 whether and how the cones’ spectral axes are propagated downstream remains unknown.…”