Spectral sensitivity of dichromats: Role of postreceptoral processes

Pergamon

0042_6989(94)OOO9H

Vision Res. Vol. 34, No. 22, pp. 2983-2990, 1994 Copyright 0 1994 Elsevier Science Ltd Printed in Great Britain. All rights reserved 0042-6989/94 $7.00 + 0.00

Spectral Sensitivity of Dichromats: Postreceptoral Processes

Role of

STEVEN H. SCHWARTZ* Received 22 January 1993; in revised form 22 February 1994; in final form 27 April 1994

Increment spectral sensitivity functions were determined for dichromatic subjects; these functions were interpreted in the context of parvo and magno pathways. Increments of either 200 or 10 msec in duration were presented on a spatially coincident, 1000 td white background. When compared to the control trichromatic data, the 200 msec functions of dichromats manifest a reduction in sensitivity at middle and long wavelengths. These data are consistent with the notion that 200 msec threshold increments reveal the sensitivity of the parvo pathway; the sensitivity of this pathway is reduced in dichromacy due to pigment replacement with a resultant loss of spectral opponency. The 10msec functions of dichromats do not show a comparable reduction in sensitivity. Therefore, it is concluded that 10msec increments reveal the sensitivity of a pathway whose spectral-opponent status is not substantially altered in dichromacy, presumably the magno pathway. Dichromacy

Magno pathway

Parvo pathway

Spectral opponency

INTRODUCTION The majority of parvo neurons in the primate retina and dorsal lateral geniculate nucleus (dLGN) manifest some form of opponency between the M- and L-cones (DeMonasterio, Gouras & Tolhurst, 1975; Zrenner, 1983; Reid & Shapley, 1992). Figure l(a) shows a schematic receptive field of such a neuron. These neurons, which will be referred to as M-L parvo neurons, presumably form the basis for a third stage red-green perceptual opponency (DeValois & DeValois, 1993). An interesting question regards the effect that red-green dichromacy has on the spectral-opponent status of M-L parvo neurons (Dain & King-Smith, 198 1). Retinal densitometry and molecular genetics suggest that in dichromacy, one of the photopigments is absent (Rushton, 1963, 1965; Nathans, Thomas & Hogness, 1986; Nathans, Piantanida, Eddy, Shows & Hogness, 1986). For example, in deuteranopia the M-cone photopigment is missing. According to the pigment replacement model of dichromacy, the M-cone photopigment is replaced with L-cone photopigment (Dain & King-Smith, 1981). Assuming that the postreceptoral synaptic arrangement is unaltered, this results in a loss of spectral opponency: all cones that contribute to the parvo neurons under consideration would have the same spectral sensitivity [Fig. l(b)]. Likewise, the equivalent parvo neurons of a protanope are not expected to manifest spectral opponency because they receive input from only cones containing the M-cone photopigment [Fig. l(c)]. *Department of Biomedical 1245 Madison Avenue,

Sciences, Southern College of Optometry, Memphis, TN 38104, U.S.A.

Spectral sensitivity

The effects of red-green dichromacy on the spectralopponent status of magno neurons are expected to be of a different nature. As indicated in the schematic receptive field in Fig. 2(a), the magno neurons of trichromats are spectrally nonopponent (DeMonasterio & Gouras, 1975; Dreher, Fukada & Rodieck, 1976; Schiller and Malpeli, 1978; Shapley, 1990). Photopigment replacement, as presumably occurs in dichromacy, does not affect this lack of spectral opponency; both the magno neurons of trichromats and dichromats are expected to be spectrally nonopponent [Fig. 2(b, c)]. The model of dichromacy that has been presented in this Introduction leads to hypotheses that may be addressed in psychophysical studies on human dichromats. The first hypothesis is that the parvo pathway of dichromats manifests a reduction in sensitivity to middle and long wavelength stimuli. This is due to a reduction in spectral opponency secondary to pigment replacement. Consider, for example, the response of the parvo neuron in Fig. l(a) to a large stimulus of 530 nm that covers its entire receptive field. The center of this neuron’s receptive field shows maximal sensitivity to 530 nm, while the surround is maximally sensitive to 555 nm. Consequently, the center response is stronger than the surround’s response, and the neuron manifests a relatively vigorous discharge. In contrast, an equivalent parvo neuron of a protanope is expected to respond less vigorously to this stimulus and manifest a higher threshold [Fig. l(c)]. This is because both the neuron’s center and surround respond strongly to 530 nm, but with opposite signs.

2983

STEVEN H. SCHWARTZ

2984

(a)

(b)

(cl

Trichromacy

Deuteranopia

Protanopia

FIGURE 1. (a) Simplified receptive field of an M-L parvo neuron located in either the retina or dLGN of a trichromatic primate. The depicted receptive field manifests spatially segregated cone inputs; other parvo neurons display cone inputs that are spatially coextensive (Wiesel & Hubel, 1966; Reid & Shapley, 1992). (b) Simplified receptive field of a parvo neuron in a deuteranope assuming a pigment replacement model of dichromacy. (c) Postulated parvo cell receptive field in protanopia. Note the loss of spectral opponency in both deuteranopia and protanopia.

An alternate model of dichromacy, pigment loss with “make-up cones”, assumes that the missing class of cones are replaced by the remaining class of cones; these make-up cones, however, are hypothesized to contribute to the signal of the remaining cones rather than opposing it (Cicerone, Nagy & Neger, 1987). If this alternative model is correct, we would expect no reduction in sensitivity for the parvo pathway. This model is examined in more detail in the Discussion. The second hypothesis is that the magno pathway will not show a reduction in sensitivity in dichromats, except in those regions of the spectrum where the missing pigment was responsible for detection. This result is expected because there is relatively little difference in the spectral opponent status of magno neurons in trichromats and dichromats; both the center and surround of magno neurons undergo similar changes in dichromacy (Fig. 2). The above hypotheses were addressed by determining increment spectral sensitivity functions with stimuli expected to isolate the parvo and magno pathways (Sperling & Harwerth, 1971; King-Smith & Carden, 1976). Increments of 200msec duration were used to identify the parvo pathway and durations of 10msec were used to identify the magno pathway. Data were obtained in

(a) Trichromecy

trichromatic and dichromatic subjects. The data show that dichromats manifest a reduction in sensitivity for middle and long wavelength stimuli when these stimuli are presumably detected by the parvo pathway. In comparison, a comparable reduction in sensitivity is not found for stimuli presumably detected by the magno pathway.

METHOD Subjects Three trichromats (JS, BD, and SS), two deuteranopes (SP and SH), and two protanopes (CO and SR) participated in the study. All subjects except the author (SS) were naive with regard to the basic design, aims and theoretical considerations of this research. Except for JS, the subjects were male. The naive observers were in their mid-twenties, and SS was in his mid-thirties. Diagnosis of color vision status was made with a Nagel I anomaloscope. Srimuli and apparatus

Increment stimuli were presented with a two-channel Maxwellian view system. Light sources for each channel were 12-V tungsten-halogen bulbs. The filaments were

(b)

(cl

Deumfanopla

FIGURE 2. (a) Simplified receptive field of a magno neuron located in either the retina or dLGN of a trichromatic primate. (b) Simplified receptive field of a magno neuron in a deuteranope assuming a pigment replacement model of dichromacy. (c) Postulated magno cell receptive field in protanopia. Note that none of these three receptive fields manifest spectral opponency.

SPECTRAL SENSITIVITY OF DICHROMATS

2985

2.0

i

I

420

I

I

I

I

I

I

460

500

540

560

620

660

I

I

420 460

I

I

500 540

I

I

I

I

620 660

560

I

I

I

420 460 500 540

I

I

560

I

620 660

Wavelength (nm) FIGURE 3. Increment spectral sensitivity, in relative quanta1 units, for three trichromats. The dashed curves represent data for stimuli of 200msec duration and the solid curves for IOmsec duration. The data have not been scaled. Error bars represent SDS.

imaged entirely within the subject’s pupil. An artificial pupil was not used. The stimulus and background were spatially coincident and 5deg in diameter (Schwartz & Loop, 1982; Snelgar, Foster & Scase, 1987). Diameter

2.0

was controlled by a single aperture located prior to the final lens of the system. Head position was maintained by use of a dental bite firmly attached to the apparatus. Stimulus and background intensity were controlled by I, /\

-1

y’

t

SH \ \

1.5 E .I> .= cn s v)

1.0

9 ‘Z = B 0” 4

.5

I

I

I

I

I

I

I

I

I

I

I

I

I

I

420

460

500

540

560

620

660

420

460

500

540

560

620

660

Wavelength (nm) FIGURE 4. Increment spectral sensitivity, in relative quanta1 units, for two deuteranopes. The dashed curves represent data for stimuli of 2OOmsec duration and the solid curves for IOmsec duration. The data have not been scaled. Error bars represent SDS.

2986

STEVEN

H. SCHWARTZ.

neutral density filters in the background channel and a circular neutral density wedge in the stimulus channel. The background retinal itlumination was 1000 td (Nygaard & Frumkes, 1982). The chromatic characteristics of the stimulus were controlled with a monochromater. The monochromater was calibrated in the apparatus with a Tektronix radiance probe (56502) positioned at the point where light emerged from the monochromater. Subsequent loss of light due to lenses and mirrors, which remained fixed for the duration of the experiment for all subjects, were ignored. Measurements were converted to log relative quanta1 units. Stimulus presentation was accomplished with solid state programming modules interfaced with a Uniblitz electronic shutter. Temporal presentation of the increment stimuli was step onset and offset. Stimulus durations of 200 and 10 msec were used.

In a single session, a complete spectral sensitivity curve for either the 200 or 10 msec stimuli was determined (King-Smith & Carden, 19’76). Thresholds were determined by the method of adjustment. A minimum of three measurements were made at each wavelength before moving to the next wavelength. The sequence of wavelength presentation was pseudo-random. Sensitivity was measured every 20 nm from (typically) 420-W nm. Subjects wore headphones playing white noise to mask shutter noise.

Increment spectral sensitivity functions for three trichromats are displayed in Fig. 3. The dashed curves show the spectral sensitivities for stimuli of 200 msec duration and the solid curves for stimuli of 10 msec duration. The data have not been scaled. The 200 msec functions manifest three distinct peaks. Reduction of the stimulus duration to 10 msec results in a dramatic change in the form of the spectral sensitivity function. The two long wavelength peaks are replaced by a single broad curve that peaks in the region of 550 nm, Note the relative sensitivities of the 200 and 1Omsec functions. The 200 msec function shows substantially greater sensitivity at all wavelengths. Data obtained on dichromats are presented in Fig. 4 (deuteranopes) and Fig. 5 (protanopes). The 200 msec functions (dashed curves) show only two peaks. These generally correspond to the peaks of the S-cone and the remaining M- or L-cone. The 10 msec functions of dichromats (solid curves) are less broad than those of trichromats, and they are displaced toward the location of the remaining M- or L-cone. Of particular relevance to this study are the relative sensitivities of the 200 and 10 msec functions. For trichromats, the 200 msec functions show substantially greater sensitivity in the middle and long wavelength region of the spectrum (Fig. 3). In comparison, the dichromats’ 200 and 10msec functions show considerably closer sensitivities over this region of the spectrum. This is due to a reduction in sensitivity for the 200 msec increments.

Li\ \

CO

2.0

at 2

RESULTS

SR

1.0

ljii 5

a

4

.5

t

- 420

460

500

540

580

620 Wa

660

4&&&Os(;O&6~~& &W

FIGURE 5. Increment spectral sensitivity, in relative quanta1 units, for two protanopes. The dashed curves represent data for stimuli of 200 msec duration and the solid curves for 10 msec duration. The data have not been scaled. Error bars represent SDS.

SPECTRAL SENSITIVITY OF DICHROMATS

2987

Trich Des .._ ___.__ Plot ---

Oi

420

I

440

I

460

I

460

I

500

I

520

I

540

I

560

I

580

I

600

I

620

I

640

I

860

I

680

Wavelength (nm) FIGURE 6. Increment spectral sensitivity functions for 200 msec increments. These functions represent the averaged data from Figs 3, 4, and 5. Functions for trichromats, deuteranopes, and protanopes are represented by the solid, dotted, and dashed curves respectively. SEMs are given for the trichromatic data.

The reduction in middle and long wavelength sensitivity for 200 msec increments can be seen by plotting the averaged data for the trichromatic and dichromatic subjects on the same graph (Fig. 6). For wavelengths greater than about 460 nm, the dichromats manifest substantial reductions in sensitivity. Figure 7 displays the averaged data for the same subjects when the stimulus duration is reduced to 10 msec. Plotting the data in this manner clearly shows the displacement of the dichromatic functions toward the location of the remaining cone. This plot also shows that deuteranopes have essentially the same sensitivity as trichromats for wavelengths of 580nm and longer. Likewise, protanopes show no reduction in sensitivity for stimuli whose wavelength is 540 nm and shorter.

DISCUSSION The data presented in this paper demonstrate that dichromats manifest a reduction in sensitivity to middle and long wavelength increments when these increments have a duration of 200 msec. These long duration increments are presumably detected by the spectrally opponent parvo pathway (Schwartz, 1992, 1994). Verriest and Uvijls (1977) obtained similar results with 500 msec increments presented with a perimeter. These psychophysical data may be interpreted in the context of spectral opponent interactions within the parvo pathway. The M-L parvo receptive field model, presented in the Introduction, assumes one cone type (M or L) provides input to the receptive field center, while the other cone type provides input to the surround. This

STEVEN H. SCHWARTZ

Trich De* ___.._._. Prot ---

I

420

I

440

I

460

I

480

I

!MO

I

520

I

540

I

560

I

580

I

f300

t

620

I

640

I

060

1

680

Wavelength (nm) FIGURE 7. Increment spectral sensitivity functions for 10 msec increments. These functions represent the averaged data from Figs 3, 4, and 5. Functions for trichromats, deuteranopes, and protanopes are represented by the solid, dotted, and dashed curves respectively. SEMs are given for the trichromatic data.

cone-specific model is supported by detailed mapping of the cone contributions to dLGN neurons (Reid & Shapley, 1992). Such a receptive field arrangement is expected to produce a substantial degree of spectral opponency. The pigment replacement hypothesis of dichromacy holds that in red-green dichromats, one of the two longer wavelength sensitive photopigments is replaced by the other. In its most simple expression, this model does not suggest changes in postreceptoral synaptic arrangements. Consequently, the remaining cones oppose each other; this reduction in spectral opponency is expected to produce a decrease in sensitivity to middle and long wavelength stimuli (Fig. 1). The cone-specific model of M-L parvo cell receptive field organization is controversial. A “mixed-surround hypothesis” suggests that the receptive field center of an

M-L parvo neuron manifests cone specificity, while the surround is formed by an indiscriminate combination of M- and L- cones (Lennie, Haake & Williams, 1991). For example, an excitatory M-center is opposed by a surround composed of inputs from both M- and L-cones. This mixed-surround model produces a spectrally opponent neuron; however, it could be argued that the strength of the opponency would be less marked than that predicted by a cone-specific model. Nonetheless, the replacement of one of the cone photopigments, as predicted by the pigment replacement hypothesis of dichromacy, is expected to produce a substantial reduction in the degree of spectral antagonism and a resultant reduction in sensitivity to middle and long wavelengths. The photopigment replacement hypothesis of dichromacy is not universally accepted (Cicerone ef al.. 1987). In

SPECTRAL

SENSITIVITY

(a) Normal Opponency

M+/L-

-

(b) Pigment Replacement

M+/M-

(c) Pigment Loss with Make-up Cones

+

:-M;;4___------

_-

M+/M+

WI

FIGURE 8. Schematics showing the basis of postreceptoral processing in the parvo pathway of a normal trichromat (a) and a protanope as predicted by the pigment replacement model of dichromacy (b) and the pigment loss with make-up cones model of dichromacy (c). Note that the photopigment replacement model predicts a reduction in sensitivity to middle and long wavelength stimuli since the remaining cones oppose each other. The pigment loss with make-up cones model does not predict a reduction in sensitivity for these same stimuli because the remaining cones show the same sign.

OF DICHROMATS

2989

instance, a protanope would have the original M-cones along with a new class of M-cones; both classes of cones would produce signals of the same sign. Such an arrangement would result in a loss of spectral opponency; however, it would not produce a reduction in sensitivity. Consequently, the results reported in this paper are not consistent with this model of dichromacy. Reduction of the stimulus duration to 1Omsec produces displacement of the dichromatic function toward the location of the remaining cone (Hsia & Graham, 1957). This displacement, presumably secondary to pigment replacement, produces a reduction in sensitivity at short wavelengths for deuteranopes and long wavelengths for protanopes. However, there is no reduction in sensitivity in the remaining regions of the spectrum. This suggests that 10 msec increments are detected by a pathway that has a similar spectral-opponent status in both trichromacy and dichromacy, presumably the magno pathway (Schwartz, 1993). We have assumed that magno neurons manifest a complete absence of spectral opponency. There are neurophysiological and psychophysical data that suggest this is not entirely correct (Finkelstein & Hood, 1982, 1984; Schiller & Colby, 1983; Logothetis, Schiller, Charles & Hurlbert, 1990). Magno neurons may show some spectral opponency; however, this opponency is considerably less robust than that expressed by parvo neurons (Reid & Shapley, 1992; DeValois & DeValois, 1993). Consequently, the opponent status of magno neurons in trichromats and dichromats is expected to be similar: neither is strongly spectrally opponent. In summary, dichromats manifest a reduction in sensitivity to increments of middle and long wavelengths, if these increments are of 200 msec duration. In contrast, there is not a comparable reduction in sensitivity for 10 msec increments. These data suggest that 200 msec increments are detected by a pathway that incurs a substantial loss of spectral opponency, secondary to pigment replacement, in dichromacy. This is presumably the parvo pathway. The 10 msec increments, however, are apparently detected by a pathway that shows less profound changes in spectral-opponent status in dichromacy, the magno pathway. REFERENCES

a cone loss model, an entire class of cones is essentially absent; these cones are not replaced. This model is at odds with the apparently normal resolution acuity displayed by red-green dichromats. Another model proposes pigment loss with “make-up cones”. These make-up cones contain the same pigment as the remaining M- or L-cones. However, unlike the pigment replacement model, it is suggested that the make-up cones contribute to the signal of the remaining M- or L-cones, rather than opposing it (Fig. 8). The pigment loss with make-up cones model does not predict a reduction in sensitivity for the parvo neurons of a dichromat. In this model, synapses that are opponent in a trichromat have the same sign in a dichromat. For

Cicerone, C. M., Nagy, A. L. & Neger, J. L. (1987). Equilibrium hue judgements of dichromats. Vision Research, 27, 983-991. Dain, S. J. &King-Smith, P. E. (1981). Visual thresholds in dichromats and normals; the importance of post-receptoral processes. Vision Research, 21, 573-580. DeMonasterio, F. M. & Gouras, P. (1975). Functional properties of ganglion cells of the rhesus monkey retina. Journal of Physiology, London, 251, 167-195. DeMonasterio, F. M., Gouras, P. & Tolhurst, D. J. (1975). Trichromatic colour opponency in ganglion cells of the rhesus monkey retina. Journal of Physiology, 25I, 197-216. DeValois, R. L. & DeValois, K. K. (1993). A multi-stage color model. Vision Research, 33, 1053-1065. Dreher, B., Fukada, Y. & Rodieck. R. W. (1976). Identification, classification and anatomical segregation of cells with X-like and Y-like properties in the lateral geniculate nucleus of old-world primates. Journal qf Physiology, London, 258, 433-452.

2990

STEVEN H. SCHWARTZ

Finkelstein, influence Finkelstein, of small,

M. A. & Hood, D. C. (1982). Opponent-color cells can detection of small, brief lights. Vision Reseurch, 22,89995. M. A. &Hood, D. C. (1984). Detection and discrimination brief lights: Variable tuning of opponent channels. l&ion Research, 24, I75- I8 I. Hsia, Y. & Graham, C. H. (1957). Spectral luminosity curves for protanopic, deuteranopic and normal subjects. Proceedings of the ~atianui Academy of SciencesU.S.A., 43, 101lliOl9. King-Smith, P. E. & Carden, D. (1976). Luminance and opponentcolor contributions to visual detection and adaptation and to temporal and spatial integration. Journal of the Optical Sociefy of America, 66, 109-7 17.

Lennie, P., Haake, W. & Williams, D. R. (1991). The design of chromati~lly opponent receptive fields. In Landy, M. S. & Movshon, J. A. (Eds), Computafional models of visual processing (pp. 71-82). Cambridge, Mass.: MIT Press. Logothetis, N. K., Schiller, P. H., Charles, E. R. & Hurlbert, A. C. (1990). Perceptual defects and the activity of the color-opponent and broad-band pathways at isoluminance. Science, 247, 214-217. Nathans, J., Thomas, D. & Hogness, D. S. (1986). Molecular genetics of human color vision: The genes encoding blue, green, and red pigments. Science, 232, 193-202. Nathans, J., Piantanida, T. P., Eddy, R. L., Shows, T. B. & Hogness, D. S. (1986). Molecular genetics of inherited variation in human color vision. Science, 232, 203-210. Nygaard, R. W. & Frumkes, T. E. (1982). Calibration of the retinal illuminance provided by Maxwellian views. Vision Research, 22, 433-434.

Schwartz, S. H. (1992). Reaction time distributions and their relationship to the transient,sustained nature of the neural discharge. Vision Research. 32, 2087.-2092.

Schwartz, S. II. (1993). Colour and Ricker thresholds for high frequency increments. Ophthalmic and Physiologicot Optics, 13. 299.-302.

Schwartz, S. H. (1994). Visual perwprion: A clinical orwn~ation. East Norwalk, Conn.: Appleton & Lange. Schwartz, S. H. & Loop, M. S (1982). Evidence for transient luminance and quasi-sustained color mechanisms. Vision Research, 22, 445- 447.

Shapley, R. (1990). Visual sensitivity and parallel retinocortical pathways. Annual Review of Psychology. 41, 635 658. Snelgar, R. S., Foster, D. H. & Scase, M. 0. (1987). Isolation of opponent-colour mechanisms at increment threshold. Vi.vion Research, 27, 1017-1027.

Sperling, H. G. & Harwerth, R. S. (1971). Red-green cone interactions in the increment-threshold spectral sensitivity of primates. Science, (72, 180@184. Verriest, G. & Uvijls, A. (1977). Central and peripherat increment thresholds for white and spectral lights on a white background in different kinds of congenitally defective colour vision. Atri Fondazione Giogio Ronchi, 32, 2 13--254. Wiesel, T. N. & Hubel. D. H. (1966). Spatial and chromatic interactions in the lateral geniculate body of the rhesus monkey. Journa/ of Neurophysiology, 29, 1115-I156. E. (1983). Neurophysiological aspects of color vision in primates. Berlin: Springer.

Zrenner,

Reid, R. C. & Shapley, R. M. (1992). Spatial structure of cone inputs to receptive fields in primate lateral geniculate nucleus. Nature, 356, 716-718.

Rushton, W. A. H. (1963). A cone pigment in the protanope. Journal

ofPhysiology, 168, 345-359.

Rushton, W. A. H. (1965). A fovea1 pigment in the deuteranope. Journal of Physiology, 176, 2437.

Schiller, P. I-I. & Colby, C. L. (1983). The responses of single cells in the lateral geniculate nucleus of the rhesus monkey to color and luminance contrast. Vision Research, 23, 1631-1641. Schiller, P. H. & Malpeli, J. G. (1978). Functional specificity of lateral geniculate laminae of the rhesus monkey. Journal qf Neurophysiology, 41, 788-797.

Ackno~tedgeme~r.~-~

would like to thank Bryan Deck, Jeannie Stone. Shane Presson, Chad Overman. Steven Reed, and Scott Henry for their participation in this experiment, Pat McCollum for secretarial assistance, and Robert Herold for assistance with the graphics. I would also like to thank two anonymous reviewers for their thoughtful comments on an earlier version of this manuscript. A preliminary version of this work was presented at the 1993 meeting of the American Academy of Optometry in Boston, Mass. Some of the 2OOmsec functions appear in the textbook entitled Visual perceprion: A clinical orientation by S. H. Schwartz, published in 1994 by Appleton & Lange (East Notwaik, Corm.).