Purchase this article with an account.
GH Jacobs, JA Fenwick, GA Williams; Evaluating the Prospects for Color Vision in the Mouse . Invest. Ophthalmol. Vis. Sci. 2002;43(13):1780.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Purpose: Many mammals segregate different cone pigments into two classes of receptor thereby setting the stage for dichromatic color vision. Antibody labeling experiments have shown that the cones of adult mice can express two different opsins, although there are varying claims about the nature and extent of such coexpression. This experiment examined whether, as has been suggested, cone photopigment coexpression may make it impossible for the mouse to achieve any color vision. Methods: Adult mice were trained to make visual discriminations in a three-alternative, forced-choice task. Increment-threshold spectral sensitivity functions were measured on photopic level backgrounds (6200 K, 57 cd/m2) at 10 nm steps from 370 nm to 600 nm. Next, animals were trained in wavelength discrimination tests [either, 500 nm (positive) vs. 370 nm (n = 2), or 370 nm (positive) vs. 500 nm (n = 3)]. Results from direct brightness matches and the spectral sensitivity data were used to render irrelevant any luminance-related cues. Results: Mouse increment-threshold spectral sensitivity functions show clear contributions from both UV (λmax= 359 nm) and M (λmax = 509 nm) pigments. Following extensive training, all the mice succeeded at making pure-wavelength discriminations. Once the initial discrimination was learned, the positive stimulus was progressively changed in 5 or 10 nm steps toward the wavelength of the negative lights. From both the long-wavelength side and from the short-wavelength side mice failed the color discrimination when the test wavelengths were in the vicinity of 400-410 nm. Conclusion: Mice appear to be capable of making dichromatic color discriminations. Very extensive training was required to demonstrate such discriminations, but whether this reflects an inherent weakness of the capacity or some limitation on this particular test situation is uncertain. If cone co-expression of opsins is near complete, as has been suggested, then the neural networks of mouse retina must be capable of extracting and reliably transmitting onward very small spectral difference signals.
This PDF is available to Subscribers Only