Wavelength dependence of visual acuity in goldfish.

Visual acuity was measured in a two-choice training experiment with food reward. Four goldfish were trained to select a homogeneously illuminated testfield when a high-contrast grating (transparency) was shown for comparison at the second testfield. Measurements were performed for white and monochromatic testfield illuminations in the light adapted state. Fourteen wavelengths between 404 nm and 683 nm were tested. For each wavelength (and white light) the testfield intensity was determined for which spatial resolution was highest. Between 446 nm and 683 nm maximal values of 2.0 cycles/deg (corresponding to a visual acuity of 15' of arc) were found. At 404 nm and in the ultraviolet resolution was lower (0.6 and approximately 0.25-0.35 cycles/deg, respectively). Cone and small ganglion cell densities may equally account for visual acuity. The action spectrum of maximal visual acuity is very similar to the spectral sensitivity function representing recognition of "colour". Measurements under reduced room illumination and after treatment with Ethambutol further indicate that the detection of high contrast gratings is processed by the same "channel" as colour vision. A similar separate and parallel processing of "colour" and "form" on the one hand, and "brightness" and "motion" on the other hand was found in humans.

Color constancy in goldfish: the limits.

Color constancy was investigated in behavioral training experiments on colors ranging from blue to yellow, located in the color space close to Planck's locus representing the main changes in natural skylight. Two individual goldfish were trained to peck at a test field of medium hue out of a series of 13-15 yellowish and bluish test fields presented simultaneously on a black background. During training the tank in which the fish were swimming freely was illuminated with white light. Correct choices were rewarded with food. During the tests differently saturated yellow or blue illumination was used. The degree of color constancy was inferred from the choice behavior under these illuminations. Perfect color constancy was found up to a certain degree of saturation of the colored light. Beyond this level test fields other than the training test field were chosen, indicating imperfect color constancy. Color constancy was quantified by applying color metrics on the basis of the goldfish cone sensitivity functions.

Simultaneous color contrast in goldfish--a quantitative study.

A set of 9-15 colored test fields was presented to goldfish. In Experiment 1, test field hues ranged from green through yellow to red; in Experiment 2, the hues varied from blue through gray to yellow. In the training conditions, the test fields were presented with a gray or black surround. The fish learned to choose one intermediate test field hue by rewarding them with food. In the test conditions, the color of the surround was changed from gray to green, or red (Experiment 1), and from black to blue, or yellow (Experiment 2). The choice behavior of the goldfish changed substantially: one of the test fields other than the training test field was preferred. Direction and strength of simultaneous color contrast was quantified in goldfish color space. The effect of spatial stimulus configuration was investigated by changing test field size and using narrow annular surrounds. With test field radii ranging between 2 and 7.5 mm simultaneous color contrast was optimal whenever the ratio between surround width and test field radius had a value of about 1:1.

Motion detection in goldfish investigated with the optomotor response is "color blind".

The action spectrum of the optomotor response in goldfish was measured to investigate which of the four cone types involved in color vision contributes to motion detection. In the dark-adapted state, the action spectrum showed a single maximum in the range of 500-520 nm, and resembled the rod spectral sensitivity function. Surprisingly, the action spectrum measured in the light-adapted state also revealed a single maximum only, located in the long wavelength range between 620 and 660 nm. A comparison with spectral sensitivity functions of the four cone types suggests that motion detection is dominated by the L-cone type. Using a two colored, "red-green" cylinder illuminated with two monochromatic lights separately adjustable in intensity, it could be shown that motion vision is "color-blind": the optomotor response disappeared whenever "isoluminant" red and green stripes were offered. Under this condition, calculations revealed that the L-cones were only slightly modulated by the "red-green" stimulus.

Reduction of red-green discrimination by dopamine D1 receptor antagonists and retinal dopamine depletion.

Reduction of wavelength discrimination ability in the 560-640 nm range, but not in the 404-540 nm range, has been demonstrated in goldfish after intravitreal injection of D1-dopamine receptor antagonists. Intravitreal injection of the dopaminergic neurotoxin 6-OH-dopamine severely reduced wavelength discrimination ability in the 540-661 nm range within 3 days. Discrimination ability could be reconstituted by the D1-agonist SKF 38393. Animals recovered from injection of 6-OH-dopamine within 14-16 days. No change of wavelength discrimination was induced by 6-OH-dopamine in the 461-540 nm range. We conclude that under photopic conditions dopamine modulates retinal mechanisms involved in red-green colour coding via D1-dopamine receptor-like binding sites.

The goldfish--a colour-constant animal.

A series of either thirteen or fifteen coloured test fields with hues from blue through grey to yellow were presented on a black background. Goldfish were trained on a bluish-grey test field by food reward. In the training situation, the setup with the coloured papers was illuminated with white light. In the test situation, the colour of the illumination was changed to blue or yellow. In both test illuminations the goldfish preferred the training field in the same way as under white illumination despite the fact that this test field stimulated the cone types very differently from the training situation. As test fields were present that excited the cones in exactly the same way as under white light, but were not chosen, colour constancy can be concluded. By means of colour metrics, it was possible to quantify direction and strength of colour constancy.

Wavelength discrimination of the goldfish in the ultraviolet spectral range.

Wavelength discrimination ability of the goldfish was measured with a behavioural training technique in the UV spectral range. First, spectral sensitivity was determined for the two fish to adjust the monochromatic lights (between 334 and 450 nm) to equal subjective brightness. The results of the wavelength discrimination experiment show that, independent of which wavelength the fish were trained on, the relative choice frequency reached values above 70% only at wavelengths longer than 410 nm. Wavelength discrimination between 344 and 404 nm was not possible. Accordingly, the delta lambda function increases steeply between 400 and 380 nm, with values between about 12 and 90 nm, respectively. Model computations indicate that the delta lambda function cannot be explained on the basis of the cone sensitivity spectra. Instead, inhibitory interactions have to be assumed which suppress the short wavelength flanks of the short-, mid-, and long-wavelength sensitive cone types in the UV range.

Induced color blindness in goldfish: a behavioral and electrophysiological study.

To answer the question whether, like man, ethambutol treated fish would become color-blind, wavelength discrimination was measured behaviorally in goldfish, preceding, during and after ethambutol treatment. The results are that of the three high discrimination abilities at around 400, 500 and 600 nm, ethambutol affected the latter one. Red-green discrimination is lost reversibly leaving the discriminations in the blue-green and violet range unaffected. This red-green discrimination deficiency cannot be accounted for by a loss of long wavelength cones since the ERG and luminosity functions remain unaffected. Intracellular horizontal cell recordings in goldfish show that ethambutol hyperpolarizes all three types of cone driven horizontal cells and changes their color coding such that their spectral characteristics become cone-like as is the case in dark adapted retina. So, the initial effect induced by ethambutol seems to be an adaptation deficiency in color vision related tasks. Human wavelength discrimination and increment threshold spectral sensitivity functions obtained at low luminance levels are compared to behavioral functions in ethambutol treated goldfish. The high similarity between the ethambutol effects in man and goldfish, and the effects observed in the horizontal cell responses in goldfish are highly indicative that horizontal cells play a key role in color vision. So far their function has been puzzling.

Separate processing of "color" and "brightness" in goldfish.

Spectral sensitivity was measured under different adaptation levels using a behavioral training technique in which the fish had to discriminate between a dark test field and a test field illuminated with monochromatic light. Depending on which of the two test fields was used as training test field, two functions were obtained which differ (1) in absolute sensitivity and (2) in shape. When trained on the dark test field, the fish seems to discriminate on the basis of a "color" cue, but it uses a "brightness" cue when trained on the illuminated test field. This was concluded from measurements of wavelength discrimination. Under low levels of the adaptation light (1.5 and 0.2 lx instead of 20 lx), the L-cone type contributes to perception of "brightness" but not to color vision. This difference in the adaptation behavior in the long-wavelength range was used to identify the ganglion cells which may represent channels for "brightness" and "color" in the retina. Action spectra were recorded extracellularly at different levels of dark- and light-adaptation.

Tetrachromatic color vision in the goldfish becomes trichromatic under white adaptation light of moderate intensity.

Spectral sensitivity of the goldfish was measured under white room light of 5 lx and 1.5 lx illuminance, using a behavioral training technique. Compared with the result obtained under 25 lx (Neumeyer, 1984), the functions differed remarkably in the mid- and longwave spectral ranges. Under 1.5 lx, the longwave maximum was absent, and wavelength discrimination was impossible in the mid- and longwave range (between 555 and 663 nm). This indicates that the longwave cone type does not contribute to color vision in these conditions. Since discrimination ability was not affected in other spectral ranges, we conclude that color vision is trichromatic then, being subserved by the ultraviolet, the short- and the midwave cone types only. Under 5 lx, the longwave cone type contributes to color vision, but, as shown in color mixture experiments, to a lesser extent.

Wavelength discrimination in the turtle Pseudemys scripta elegans.

Wavelength discrimination was measured with two individuals using a behavioral training technique. The delta lambda function has three minima, indicating best discrimination at 400, 510 and at 570 nm (S2) or 600 nm (S1). In the range between 450 and 510 nm the ability to discriminate wavelengths was absent. This can be explained by the filter effect of the colored oil droplets, as indicated by a model computation. Color mixture experiments revealed that the high discrimination ability at 400 nm is based on an ultraviolet sensitive photoreceptor which is the fourth cone type involved in turtles' color vision. Spectral sensitivity in the ultraviolet region (measured for one turtle) is maximal at 370-380 nm.

Spectral sensitivity of the freshwater turtle Pseudemys scripta elegans: evidence for the filter-effect of colored oil droplets.

Spectral sensitivity was measured in the light adapted state for two individuals (S1 and S2) of Pseudemys scripta elegans using a behavioral training technique. The curves have three pronounced maxima at 450 nm, at 570 nm (S1) or 540 nm (S2), and at 640 nm. The spectral sensitivity functions can be described by the over-envelope of the effective cone sensitivity functions which are determined by the absorption spectra of the cone photopigments (lambda max at 450, 518 and 620 nm) and by the filter properties of the colored oil droplets.

On spectral sensitivity in the goldfish. Evidence for neural interactions between different "cone mechanisms".

Spectral sensitivity was measured in freely moving, light adapted goldfish using a behavioral training technique. The curve reveals 3 pronounced maxima at 470, 540 and 660 nm. Compared with the absorption spectra of the cone photopigments with lambda max at 450, 530 and 625 nm, the maxima were much narrower and shifted towards longer wavelengths. Results under chromatic adaptation indicate that spectral sensitivity measured under the white adaptation light is determined by inhibitory interactions between the spectrally different "cone mechanisms". For the short-wave maximum, a summation process probably has to be assumed in addition.

Brightness matching, brightness cancellation, and increment threshold in the Ehrenstein illusion.

Matching and cancellation techniques were used to measure the relative strength of the Ehrenstein illusion in dark figures on a light background (negative contrast) and light figures on a dark background (positive contrast). Brightness enhancement on the former was shown to be maximally 0.28 log unit (relative to the detection threshold), and darkness enhancement on the latter 0.43 log unit. Values differed little with figure-ground contrast (down to a minimum of +/- 0.5), but decreased with decreasing level of illumination. The luminance increment (decrement) needed to match the illusory brightness (darkness) was similar in size to the luminance decrement (increment) needed to cancel the illusion. The increment threshold for a small test flash measured in three locations relative to the subjective contour delineating the illusion did not differ systematically. The results are compatible with a neurophysiological explanation of the Ehrenstein illusion in terms of line-induced lateral interaction in hypercomplex receptive fields.