- Email: [email protected]
Spectral efficiency measured by heterochromatic flicker photometry is similar in human infants and adults
Vision Res. Vol. 35, No. 10, pp. 138%1392, 1995
~
Pergamon
0042-6989(94)00245-2
Copyright ~'~ 1995 Elsevier Science Ltd Printed in Great Britain. All rights reserved 0042-6989/95 $9.50 + 0.00
Spectral Efficiency Measured by Heterochromatic Flicker Photometry is Similar in Human Infants and Adults MICHELLE L. BIEBER,* V|CKI J. VOLBRECHT,t JOHN S. WERNER* Received 3! January 1994; in revised form 20 July 1994; in final form 22 September 1994
Spectral efficiency functions based on heterochromatic flicker photometry (HFP) were measured for three adults and 42 infants using a rapid visually-evoked potential (VEP) method. A 5°-diameter, broadband standard (0.6 cd/m2) was presented in square-wave counterphase (15 Hz) with one of 13 monochromatic lights (420-660 nm; 20 nm steps). The intensity of the monochromatic light was continuously varied while extracting the phase-locked VEP amplitude of the fundamental component. HFP functions measured psychophysically by the method of adjustment were also obtained for the adults. Adult HFP functions from the two methods were found to be essentially the same. Both of these functions were compared to Vos'-modified 2° V(7.) function and to the 10° CIE V(7.) function. The mean adult data were slightly better fit to the 2° V(7.) function than to the 10° CIE V0~) function, although there was an elevation in sensitivity at 420 and 440 nm. Infant HFP functions were similar to Vos' modified V(~) except for an elevation in efficiency at short wavelengths. The mean infant HFP function agreed better with the 10°CIE V(7.) function than Vos'-modilled V(7.) function, but infant sensitivity was elevated by 0.4 log units at 420 nm compared to the 10° CIE observer. The elevation found at short wavelengths for both adults and infants is attributed to individual and age-related variation in the density of the ocular media, and to reduced macular pigment screening resulting from use of a 5° field size. An infant spectral efficiency function based on the HFP data collected in this study was modeled. This function may serve to equate stimuli with respect to luminosity for the average infant. Flicker photometry Evokedpotential Infant color vision
The Commission Internationale de l'l~clairage (CIE) defined a function, known as the CIE V(k) function, to represent the spectral efficiency of a standard adult observer under photopic conditions. This function was originally based on data from a large number of observers using hetertochromatic flicker photometry (HFP), step-by-step brightness matching, and direct brightness matching (Coblentz & Emerson, 1918; Hyde, Forsythe & Cady, 1918; Gibson & Tyndall, 1923; Hartman, 1918), but revisions have been proposed several times (Judd, 1951; Vos, 1978a). The CIE has also standardized a large-field V(~.) function, which was derived from 10° color-matching data (Stiles & Burch, 1959; Speranskaya, 1959). While the CIE V(X) functions have proven useful in a wide variety of applications in science and industry, it is often necessary to take into account individual differences in spectral efficiency, or what Kaiser (1988) has called sensation luminance. In this case, HFP is the *Department of Psychology, Campus Box 345, University of Colorado, Boulder, CO 80309~0345, U.S.A. tColorado State University, Ft Collins, CO 80525, U.S.A. VR 35/10 B
procedure of choice because, unlike step-by-step and direct brightness matching, the HFP function conforms to Abney's law of additivity, at least over relatively moderate photopic luminance levels (Ives, 1912). The HFP function is thought to represent the sensitivity of an achromatic mechanism, carried by the magnocellular retinostriate pathway and possibly also by parvocellular fibers that multiplex chromatic and achromatic signals (Lennie, Pokorny, & Smith, 1993). This function can be modeled quite well by summing the outputs of the long-wave sensitive (L-) and middle-wave sensitive (M-) cones, in a ratio of approx. 2: 1, provided that the sensitivities are corrected for the spectrallyselective absorption of the ocular media and macular pigment. It has also been suggested that there is a small subtractive component from the short-wave sensitive (S-) cones to the HFP function (Stockman, MacLeod & DePriest, 1987; Lee & Stromeyer, 1988; Vos, Est~vez & Walraven, 1990); however, this input is only demonstrated under special conditions, and therefore can be ignored for most purposes. The most important factors to account for individual
1385
1386
M1CHELLE L. BIEBER et al.
and age-related variations in HFP functions are likely to be differences in the density of the ocular media and macular pigment (Ruddock, 1972). For example, a progressive loss in short wavelength efficiency has been observed in HFP functions of normal observers between the ages of 10 and 80 yr (Verriest, 1971; Ruddock, 1972; Kraft & Werner, 1994) and this change can be explained simply by the age-related increases in the density of the ocular media. Because the ocular media density is lower in human infants compared to young adults (Werner, 1982; Weale, 1988; Hansen & Fulton, 1989), an additional elevation at short wavelengths might be expected for infant HFP functions. Depending on the field size of the stimuli, this elevation might be further enhanced due to the reduced density of macular pigment in the infant eye (Schultze, 1866; Hering, 1885; Bone, Landrum, Fernandez & Tarsis, 1988). Still other differences between the HFP functions of human infants and adults may result from morphological and physiological immaturities in the infant retina and visual pathways. The short, fat cones in the foveal region (Abramov, Gordon, Hendrickson, Hainline, & Dobson, 1982; Yuodelis & Hendrickson, 1986) are inefficient in capturing light due to poor wave-guide properties (Banks & Bennett, 1988). In addition, while the postreceptoral processing of cone signals by the magnocellular pathway may account for the adult HFP function (Lennie et al., 1993), the differential rate of maturation of the parvocellular and magnocellular pathways may contribute to differences in the HFP functions of human infants. Anatomical evidence suggests that the parvocellular pathway develops at a faster rate postnatally than does the magnocellular pathway (Hickey & Peduzzie, 1987). Despite previous measures of infant spectral sensitivity (Dobson, 1976; Peeples & Teller, 1978; Moskowitz-Cook, 1979; Ladenhiem & Gordon, 1986; Anstis, Cavanagh, Maurer & Lewis, 1986; Teller & Lindsey, 1986; Nusinowitz & Maurer, 1992), a complete spectral efficiency function based on HFP is not yet available. In this study, HFP functions were measured for 42 infants and three adults covering a broad spectral range (4200660nm), using a rapid visually-evoked potential (VEP) method similar to that described by Regan (1970; see also Padmos & Norren, 1972; Est6vez, Spekreijse, van den Berg & Cavonius, 1975). The VEP method was first validated by comparing adults' HFP functions obtained with the VEP to those obtained with the psychophysical method of adjustment. The results of this comparison indicate that the VEP method is capable of accurately measuring HFP spectral efficiency, consistent with previous comparisons of the VEP and psychophysics (Siegfried, Tepas, Sperling & Hiss, 1965; Regan, 1970; Siegfried, 1978). This method was then used to measure spectral efficiency in infants. The results indicate that the average infant HFP function is similar to the 10° CIEV0~) function except for an elevation in sensitivity at 420nm. This is consistent with reduced macular pigment screening as a result of using a 5° stimulus (4400520 nm), and with age-related variation in the ocular media
(420 nm). A spectral efficiency function based on the mean infant HFP data was modeled. This modeled function may be useful for equating stimulus luminosity in future infant research.
METHODS
Subjects Three female adults (22, 35, and 41 yr of age), served as observers. All displayed normal color vision when tested with the Neitz anomaloscope, Farnsworth panel D-15 test, and the American Optical HRR pseudoisochromatic plates. Ninety full-term infants were recruited from birth announcements in the local newspaper to participate in this study. Data were successfully collected from 21 2-month-old infants (14 male and 7 female) and 21 4-month-old infants (10 male and 11 female). Each infant was tested within -I-7 days of their 2nd- or 4th-month birthday. Each infant participated in 1-4 testing sessions that lasted approx. 45-60 min. The parents of all babies reported a negative family history of color deficiency, or one that was unlikely to affect the baby (verbal reports suggested that the mothers of two female infants could be heterozygous for a congenital red-green color defect).
Stimulus and apparatus Stimuli. A 5°, circular stimulus was presented as a broadband white light (CIE coordinates x,y = 0.37, 0.40; 4516 K; 0.6 cd/m 2 ~20 td) alternating in 15-Hz squarewave counterphase with one of 13 monochromatic lights (420°660 nm; 20 nm steps). The intensity of the monochromatic light was varied continuously in either an increasing or a decreasing series (30 sec each direction) with a computer-driven neutral density wedge (2 log unit range). Infants were held 32 cm from the stimulus by one of their parents. Adult subjects were positioned 32 cm from the stimulus by use of a forehead and chin rest. No fixation point was used. Optical system. The light source was a 1000-W xenon arc lamp operated at 980W with a d.c. power supply. Spectral lights were produced with a holographic grating monochromator (Instruments SA, Model H-20). Conventional optics were used to collimate and focus the light from two channels to a common focus at a sectored mirror that produced counterphase alternation between the broadband and monochromatic light. The intensity of the stimulus was controlled by calibrated neutral density filters and neutral density wedges. The light from both channels then passed through a field stop to be projected onto a rear-view projection screen inside a dark, electrically-shielded chamber. Radiometric measurements were made with a silicon photodiode and a linear read-out system (United Detector Technology, 81 optometer) that was calibrated against a standard traceable to the National Institute of Standards and Technology. A He-Ne laser (Spectra
INFANT SPECTRAL EFFICIENCY
Physics) and mercury-line interference filters were used to calibrate the monochromator. The spectral bandwidth of the monochromator was 8 nm at half power. The luminance and correlated color temperature of the broadband field were measured with a photometer/spectroradiometer (Photo Research, Model PR703-A). Recording s.vstem. Bipolar VEPs were recorded with SensorMedics EEG electrodes and Beckman earclip electrodes. For the infants, electrodes were placed 1 and 3 cm above the inion, and the ground was placed on the forehead. For the adults, electrodes were placed 2 and 8 cm above the inion and a ground was placed on the ear lobe. Inter-electrode impedance, measured before and after testing with a Grass Impedance Meter (model EZM5B), was ~<5000 Q for the adults and ~<15,000 Q for the infants, typical levels for adults and infants (Dobson, 1976; Moscowitz-Cook, 1978). The EEG signal was led first through an isolation amplifier (Analog Devices, Model 273K) and then into a battery-powered pre-amplifier (Princeton Applied Research, Model 113). The nominal gain of the pre-amplifier was 10,000 and the low and high frequency roll-offs were set at 0.3 and 30.0 Hz, respectively. Both the isolation amplifier and the pre-amplifier were in the shielded chamber with the subject. The amplified and filtered EEG signal was fed into an oscilloscope and then into a vector voltmeter (Ithaco Dynatrac 3, model 393). The vector voltmeter was synchronized to the chopper at 15 Hz, with a 1.25 sec time constant. The vector voltmeter uses the lock-in principle to selectively amplify the phase-locked amplitude of a specific frequency component of the signal. Because the vector voltmeter extracts the VEP in real time, continuous changes in amplitude could be plotted as a function of the intensity of the monochromatic light. We found the fundamental component of the VEP amplitude agreed well with psychophysical measures under our conditions, consistent with previous work by Siegfried et al. (1965; cf. Regan, 1970).
Procedure Psychophysics. Psychophysical HFP data were collected from adult subjects following 10min of dark adaptation. Using the method of adjustment, the observer's task was to vary the intensity of the monochromatic light until flicker was minimized. Psychophysical HFP spectral efficiency was defined by the log reciprocal quanta required to minimize flicker. The wavelength of the stimuli was varied in random order. Each adult participated in three 60-min sessions. Adult VEPs. Testing began after 10min of dark adaptation. During testing, adult subjects were continuously viewed with an infrared illuminating system and video camera to monitor head and eye movements, as well as blinking. If the subject moved or blinked, recording was momentarily interrupted. Evoked responses were recorded as the intensity of the monochromatic light was continuously decreased or increased for 30 sec. VEP-HFP spectral efficiency was defined by the log reciprocal quanta required to produce a minimum amplitude of the VEP. Wavelength was varied in random
1387
order. Each adult subject participated in 3-5 recording sessions that lasted approx. 1.5-2 hr. Infant VEPs. The procedure used to measure infant VEPs was almost identical to the procedure used for the adults. After the electrodes were in position, infants were brought into the shielded chamber and seated on a parent's lap. Prior to testing, infants were dark adapted for 5 min. A control study with one of the adult subjects indicated that there was no observable difference between psychophysical HFP functions obtained after 5 or 10 min of dark adaptation. During testing, infants were viewed continuously on a video monitor so that recording could be stopped whenever the infant was not attending to the stimulus. Unlike adults, it was usually not possible to record responses from infants to all 13 wavelengths in one session. Each infant participated in 1 4 sessions that lasted approx. 30-60 min each. RESULTS Figure 1 shows typical, individual and mean VEP records obtained from adult and infant subjects. Relative VEP amplitude is plotted as a function of the relative log quanta of the monochromatic light. For each adult observer, 4-15 (mean = 8.5) VEP records were obtained at each wavelength, and for each infant observer, 1-5 (mean=2.3) VEP records were obtained at each
Adult (440 nm)
Infant (440 nm) Record 1
Q.
"0
.__= E
< 13_ LLI >
3
.>_ t~ nMean
Vean I
I 1
0
1
Relative Log Quanta F I G U R E 1. Typical VEP records obtained from an infant and adult observer are shown in the left and right panels respectively, Relative amplitude of the VEP response is plotted as a function o f the intensity of the monochromatic light. The mean of each set of three records is shown with the best-fitting parabola. The VEP criterion response was defined by the m i n i m u m of the polynomial function.
1388
MICHELLE L. BIEBER et al.
wavelength. Notice that the individual and mean VEP records have a V or U shape. The minimum amplitude of the VEP of each of these records was found to correspond to the psychophysically-measured point of minimum flicker (i.e. equal sensation luminance between the monochromatic light and standard).
i
Adult spectral efficiency To validate the VEP measure of spectral efficiency, a comparison was made between the psychophysically-obtained H F P functions of each adult observer and their electrophysiologically-obtained H F P functions. Figure 2 presents these functions as well as a mean function for the three adult subjects. Log reciprocal quanta required for the criterion response is plotted as a function of wavelength. The position of each curve along the axis of ordinates is arbitrary. The solid and open symbols represent VEP and psychophysical measures, respectively. The V E P - H F P functions have been normalized to the psychophysical H F P functions using a least-squares criterion. Both of these functions are plotted with Vos' 2 ° V(;L) function (solid curve) and the 10° CIE V(X) function (dashed curve) for comparison. The 2 ° and 10° V(;L) functions have been normalized to the psychophysically obtained H F P functions at 580 nm, where all three functions are expected to have the same shape and are
f
r
[
l
l
,
Data analysis As expected, some records contained noise artifacts (due to abrupt movements or fussy behavior), so an objective method was needed to determine whether a record should be retained for further analysis. After each session, individual records were digitized and fitted with a polynomial regression equation. All records that were fit significantly better by an upward facing parabola compared to a straight line (0t = 0.001) were averaged for each subject and wavelength. Overall, 5-40% of the records had to be eliminated for each subject. While forty percent may sound large, some rather active infants provided only five records, and two had to be eliminated. The mean VEP function for records meeting the criterion was computed at each wavelength. As shown in Fig. 1, a parabola was fitted by the method of least-squares to each of these mean amplitude vs intensity functions, and the log reciprocal quanta required to produce the minimum of the parabola was taken as the H F P criterion response. A parabola was chosen for convenience in locating the minimum of the V-shaped functions. No theoretical significance is attached to this fitting method, although it may be noteworthy that the coefficient of the x 2 parameter varied little with wavelength, implying a relatively fixed shape as a function of wavelength. In addition, based on results of Pokorny, Smith and Lutze (1989), we compared H F P functions for 12 infants and one adult derived from fits to the VEP data in linear rather than log stimulus space. An analysis of variance indicated that there was no significant main effect (F1,13=0.257, P=0.6207) or interaction (F~j.t43--- 1.108, P = 0.3574) between the fitting method and wavelength.
l
MB
t-
v
r~ ,t-
.o iii
,~
vv
I1)
._> I1)
rr
(:7) 0
Mean
o
II 400
I
I 500
I
I
1 I
600
700
Wavelength (nm) FIGURE 2. Log relative efficiency (quanta) is plotted as a function of wavelength (nm). Individual adult functions and the mean adult function are arbitrarily separated by 2 log units. HFP functions obtained electrophysiologicallyand psychophysically are represented by the solid and open symbols respectively. Vos'-modified 2° V(~,) function (solid curves) and the 10° CIE V(~,) function (dashed curves) have been normalized at 580 nm to the psychophysical HFP functions for comparison.
relatively unaffected by variations in ocular media density. As shown in Fig. 2, there is good agreement between the VEP and psychophysical spectral efficiency functions for each observer at most wavelengths, as previously reported for studies with adults using similar methods (Siegfried, et al., 1965; Regan, 1970). The mean adult function is fit slightly better to the 2 ° V(X) function than to the 10° CIE V(X).
Infant spectral efficiency H F P spectral efficiency functions with nine or more wavelengths were obtained from 20 of the 42 infants, using the VEP criterion response. Log reciprocal quanta is plotted for these data as a function of wavelength in Fig. 3. The H F P functions are arbitrarily separated by 1 log unit along the ordinates to better display the data. A least-squares criterion fit for wavelengths ~>520 nm to Vos' modified 2 ° V(X) function (where age-related changes in ocular media density are minimal) was determined for all the infant functions. All remaining infant functions are normalized in the same way unless otherwise stated. This was done to prevent masking an elevation that might be present in the infant data at short
INFANT SPECTRAL EFFICIENCY wavelengths, due to reduced ocular media density (Werner, 1982; Weale, 1988; Hansen & Fulton, 1989) and reduced macular pigment density (Schultze, 1866; Hering, 1885; Bone et al., 1988). The Vos 2 ° V(X) function was used rather than the 10 ° C I E V(k) function because it provided the best fit to the mean adult data. The remaining comparisons using the mean infant data are made to both of these standard functions. There is a close correspondence between most of the infant data sets and the 2 c' V(~,) function. Spectral efficiencies were also measured for 22 of the 42 infants at a subset of the spectral points shown in Fig. 3. These data are consistent with the more complete data sets shown. The infant data are somewhat more variable than the adult data presented in Fig. 2. This is to be expected from the more variable infant behavior during data collection and also because more records were obtained at each wavelength for the adult observers than for the infant observers. The latter factor may be less important than the former. A spectral efficiency function for one adult subject was redetermined using only two randomlyselected VEP records per point (close to the mean number of records obtained from infants). This function looked nearly the same as that determined from the full data set, except for one deviant point. The mean absolute
I
I
I
I
I
I
I
I
2-Month Infants
I
I
~
I
I
I
4-Month Infants
,
•? o
E -1 I:)-
/
0 c-
._o w 0
._> OC: 0 --J
^
)
o- C
yl°/ 400
l
i 500
i
L i 600
i 700
I
Ifl
I
400
500
I
L
600
I
i
700
W a v e l e n g t h (nm)
FIGURE 3. Log relative efficiency(quanta) is plotted for all complete data sets (10 2-month olds and 10 4-month olds) as a function of wavelength (nm). Each data set was normalized to Vos'-modified 2° V(X) function (solid curve) using a least-squares criterion for wavelengths >t 520 nm.
1389
¢.0.0
>" e-
-0.5
.~_ 0
-1.0
lit 11) >
-1.5
a) nr'
-2.0
0 I
I
I
I
L
I
L
400
450
500
550
600
650
700
Wavelength (nrn)
FIGURE 4. Log relative efficiency(quanta) is plotted as a function of wavelength (nm) for the mean infant ((D) and mean adult (0) VEP-HFP functions. Overlapping data for infants and adults are denoted by tD. Both mean functions were normalized to Vos'-modified 2° V(~,) function (solid curve) using a least-squares criterion for wavelengths >~ nm. The 10° CIE V(~,)function (dashed curve) has been normalized at 580 nm to Vos' modified 2 V(X)function.
difference between the two H F P functions was 0.13 log units. The mean spectral efficiency of all infants is shown in Fig. 4 (open circles). Averaging the 2- and 4-month data was justified because there are no systematic differences between the mean functions of the two age groups (average deviation 0.07). There is a close correspondence between the 10 ° V(X) function and the mean infant spectral efficiency function except for a 0.39 log unit elevation in the mean infant function at 420 nm. Also plotted in Fig. 4. is the mean adult VEP spectral efficiency function (solid circles). This function has been similarly normalized to the 10 ° and 2 ° V(~,) functions. As shown in this plot, the adult function is fit better to the 2': V(~,) function except where there is an elevation at 420 and 440 nm. An elevation in sensitivity at short wavelengths from approx. 440-500 nm would be expected for both infants and adults due to the use of a 5 ° field compared to a 2 ° field. With the former, the macular pigment will influence sensitivity less because of its lower density at this retinal eccentricity (Stabell & Stabell, 1980a; Werner, Donnelly & Kliegl, 1987). The macular pigment should not have a significant influence at 420 nm; hence, this elevation is more likely due to lower than average ocular media density in infants and these adults. The elevation in sensitivity at 420 nm observed for infants compared to identically-tested adults and to the 10 ° V(X) function is expected because of even lower ocular media density in infants (Werner, 1982). DISCUSSION H F P spectral efficiency measured with the VEP is similar for both 2- and 4-month-old infants. The mean infant function is essentially the same as the 10°V(~.) function (solid curve), except for an elevation in infant
1390
MICHELLE L. BIEBER et al.
sensitivity at the short wavelength extreme of the spectrum. These results are in close agreement with previous investigations using different methods (Dobson, 1976; Moskowitz-Cook, 1979). Because the previous studies used methods other than HFP, their results may represent the efficiency of different neural mechanisms. The elevation in efficiency at short wavelengths found in this study might, at first glance, be thought to be due to rod and peripheral cone contributions, both of which would enhance sensitivity to short wavelength light (Wooten, Fuld & Spillmann, 1975; Abramov & Gordon, 1977; Stabell & Stabell, 1980b). There is good reason to believe however, that rods did not contribute to the elevation at short wavelengths found in the present study. First, the temporal frequency of the stimulus was 15 Hz, a frequency that rods do not resolve very well under the conditions used in this study (Hecht & Verrijp, 1933). Second, the VEP is usually dominated by signals originating in the fovea (Potts & Nagaya, 1965; Wooten, 1972; Siegfried, 1978). Finally, the field size of the stimulus was 5° . In the newborn eye, the rod-free zone (foveola) is roughly 5.4 °, decreasing to about 2.3 ° at maturity (Banks & Bennett, 1988). Of course, due to unsteady fixation, some contribution from the peripheral cones cannot be ruled out. We argue, instead, that the elevation in efficiency at short wavelengths for infants relative to the 2 ° and 10° standard observers may be explained by developmental changes in the density of the ocular media (420 nm), and reduced macular pigment filtering due to the use of a 5 ° stimulus field size (440--500 nm). Experimental evidence has shown that the density of the lens pigment in the infant eye is low, and increases over the life span (Werner, 1982; Weale, 1988; Hansen & Fulton, 1989), but at every age, there is also considerable individual variation in density that contributes to individual differences in spectral sensitivity (Ruddock, 1972). The macular pigment is also reduced in the infant eye (Schultze, 1866; Hering, 1885) increasing to adult levels over the first few years of life (Bone et al., 1988). At 5° however, macular pigment does not have high density in adults and presumably also not in infants so it is unlikely to contribute to adult-infant differences in spectral efficiency.
*S-conecontributions are difficultto detect in the luminosityfunction based on curve fitting, but can be identified under certain experimental conditions. Additional analyses were conducted to determine whether S-conecontributions might be detectableunder our testing conditions. STEPIT was used to minimize the squared log ratio differencebetweenmean VEP-HFP functionsand the sum of the L-, M- and S-cone fundamentals in the same manner as described above, except S-conecontributions were allowedto vary. For the infant function, STEPITestimatedan S-conescalar of 0.006 and a lens scalar of 0.28. This additional S-coneparameter did not improve the fit to the mean infant VEP function (Fj.~3=0.13994, P>0.05). For a comparably modeled adult function, STEPIT estimated an S-conescalar of 0.006 and a lens scalar of 1.0, exactly equal to that of the standard observer.As with the infant function, allowing S-cones to vary didn't provide a significantlybetter fit to the mean adult VEP function (F~.~3= 1.367, P > 0.05). Giventhe lack of statistical justification, possible S-cone contributions were not included in the modeled infant function.
Modeling the HFP function in the infant The combined signals of the L- and M-cones, in a ratio of approx. 2:1, provides an excellent fit to V00, provided the sensitivities are corrected for the spectrally-selective absorption of the ocular media and macular pigment. In addition to known differences between infants and adults in their preretinal screening pigments, another difference that might contribute to differences in H F P spectral efficiency functions is the length of the cone outer segments. The shorter outer-segments of infant cones (Yuodelis & Hendrickson, 1986) will result in a reduction in photopigment density which is manifested as a narrowing of the shape of the photopigment absorption spectrum (Dartnall, 1962). Sensitivity reductions at the spectral extremes would, therefore, be expected for infants relative to adults. A quantitative evaluation of this effect, assuming maximal differences between infants and adults, revealed that reducing photopigment density from 0.35 (Rushton, 1963; King-Smith, 1971) to 0.001 decreases sensitivity at the spectral extremes only slightly (0.14 and 0.15 log unit reduction in efficiency at 420 and 660 nm respectively). At short wavelengths, this loss in sensitivity would be expected to be more than compensated by the less dense ocular media. At longer wavelengths, reduced photopigment density in infants might be expected to lower sensitivity relative to adults, but it appears not to be detectable in our measurements. Alternatively, the reduction in photopigment density for infants may not be as high as that assumed above. Given our results, possible effects of reduced photopigment density can be ignored for purposes of modeling the mean infant H F P function. The mean infant H F P function was modeled to determine a standard infant spectral efficiency function that may be used to equate chromatic stimuli in terms of luminosity. STEPIT (Chandler, 1965) was used to minimize the squared log ratio difference between the mean infant spectral efficiency function and a sum of Land M-cone fundamentals* (Smith & Pokorny, 1975), no macular pigment screening and a variable multiplicative scalar to adjust ocular media density. The shape of the ocular media density spectrum was assumed to follow the tabled values for the standard observer of Vos (1978b). A 2:1 ratio of L : M cones was assumed based on previous evidence from adult psychophysical studies (Vos & Walraven, 1971; Cicerone & Nerger, 1989). It should be noted, however, that this assumption is not critical as long as the ratio favors L-cones (Pitt, 1935; Nerger, 1988; Kraft & Werner, 1994; cf. Lutze, Cox, Smith & Pokorny, 1990). This implies that, contrary to previous claims (Morrone, Burr, & Fiorentini, 1993), similarity of infant and adult red-green flicker nulls does not necessarily imply similar L : M cone ratios. The STEPIT analysis estimated an ocular media density scalar of 0.51 for the infant, which is in close agreement with previous estimates based on scotopic spectral sensitivity (Werner, 1982). The modeled function is shown in Fig. 5 by the bold solid curve along with the mean infant data (open circles). With the exception of two
INFANT SPECTRAL EFFICIENCY I el. t~
I
I
I
I
I
[
0.0 -0.5
o¢ -
-1.0
"5 -1.5 lit
._> ¢r 0 ..-I
-2.0 -2.5 -3.0
400
I
I
I
I
I
I
450
500
550
600
650
700
Wavelength (nm)
FIGU RE 5. Log relative efficiency is plotted as a function of wavelength (nm). Open circles represent the mean infant HFP function. Error bars denote + 1 SD. The bold solid curve represents the function modeled to the infant data. Vos'-modified 2° V(X)function (solid curve) and the 10c' CIE V(X) (dashed curve) were normalized at 580 nm to the modeled function for comparison.
points (460 and 620 nm), the mean infant data agree well with the modeled infant function. The Vos-modified 2 ° V(k) function (solid curve) and the 10 ° CIE V(~,) function have been normalized at 580 nm to the theoretical infant function for comparison. The modeled infant function fits extremely well to the 10 ° V(X) function except at 420 nm
TABLE 1. Modeled 5° infant spectral efficiency (log quanta) Wavelength (nm)
Efficiency
Wavelength (nm)
400 405 410 415 420 425 430 435 440 445 450 455 460 465 470 475 480 485 490 495 500 505 510 515 520 525 530 535 540 545 550
- 1.754" - 1.624" - 1.494" - 1.363" - 1.233 - 1.171 - 1.110 - 1.048 -0.987 -0.939 -0.890 -0.842 -0.794 -0.726 -0.657 -0.588 -0.519 -0.454 -0.389 -0.324 -0.259 -0.215 -0.171 -0.127 -0.082 -0.063 --0.044 -0.025 -0.006 --0.005 -0.003
555 560 565 570 575 580 585 590 595 600 605 610 615 620 625 630 635 640 645 650 655 660 665 670 675 680 685 690 695 700
Efficiency -0.002 0.000 -0.015 -0.031 -0.046 -0.062 -0.096 -0.130 -0.164 -0.198 -0.252 -0.307 --0.361 -0.415 -0.502 -0.588 -0.674 -0.760 -0.877 -0.994 - 1.111 - 1.228 - 1.370" - 1.511" - 1.653" - 1.795" -- 1.952" --2.109" --2.267* - 2.424*
*Values that are extrapolated beyond the limits of empirical measurement.
139 l
w h e r e r e d u c e d o c u l a r m e d i a filtering is e x p e c t e d to be most pronounced for the wavelength range examined. T h e s p e c t r a l efficiency v a l u e s o b t a i n e d f r o m this a n a l y s i s a r e listed in T a b l e 1 f o r w a v e l e n g t h s b e t w e e n 400 a n d 700nm. A s t e r i s k s d e n o t e t a b l e d v a l u e s t h a t are e x t r a p o l a t e d b e y o n d the s p e c t r a l limits o f the e m p i r i c a l d a t a . W h i l e s p e c t r a l efficiency is s i m i l a r f o r a d u l t s a n d i n f a n t s at m i d d l e a n d l o n g w a v e l e n g t h s , d i f f e r e n c e s at short wavelengths are important. T h e m o d e l e d i n f a n t s p e c t r a l efficiency f u n c t i o n p r o v i d e s a fairly a c c u r a t e d e s c r i p t i o n o f t h e m e a n i n f a n t H F P f u n c t i o n u n d e r s t a n d a r d c o n d i t i o n s o f testing. A s w i t h t h e C I E V(X) f u n c t i o n , this f u n c t i o n m a y be useful in e q u a t i n g s t i m u l i f o r a n a v e r a g e o b s e r v e r . B e c a u s e o f individual differences between observers, however, n e i t h e r the m o d e l e d 5 ° i n f a n t s p e c t r a l efficiency o r the a d u l t C I E V(~,) f u n c t i o n s c a n be e x p e c t e d to a c c u r a t e l y e q u a t e s t i m u l i f o r all i n d i v i d u a l o b s e r v e r s . I n m a n y i n s t a n c e s i n d i v i d u a l s e n s a t i o n l u m i n a n c e will be r e q u i r e d to a c c u r a t e l y e q u a t e s t i m u l i in t e r m s o f l u m i n o s i t y . H o w e v e r , this m o d e l e d s p e c t r a l efficiency f u n c t i o n m i g h t be useful to the r e s e a r c h e r w h o is i n t e r e s t e d in g r o u p d a t a , o r is r e s t r i c t e d b y t i m e o r e q u i p m e n t to d e t e r m i n e individual sensation luminance.
REFERENCES
Abramov, I. & Gordon, J. (1977). Color vision in the peripheral retina. I. Spectral sensitivity. Journal of the Optical Society of America, 67, 195-202. Abramov, I., Gordon, J., Hendrickson, A., Hainline, L., Dobson, V. & LaBossiere, E. (1982). The retina of the newborn human infant. Science, 217, 265-267. Anstis, S., Cavanagh, P., Maurer, D. & Lewis, T. (1986). Early maturation of luminous efficiency for colored stimuli. Investigative Ophthalmology and Visual Science (Suppl.), 27, 264. Banks, M. S. & Bennett, P. J. (1988). Optical and photoreceptor immaturities limit the spatial and chromatic vision of human neonates. Journal of the Optical Society of America A, 5, 2059-2079. Bone, R. A., Landrum, J. T., Fernandez, L. & Tarsis, S. L. (1988). Analysis of macular pigment by HPLC: Retinal distribution and age study. Investigative Ophthalmology and Visual Science, 29, 843-849. Chandler, J. P. (1965). "STEPIT", Quantum chemistry exchange program. Department of Chemistry, Indiana University, Bloomington, IN. Cicerone, C. M. & Nerger, J. L. (1989). The relative numbers of long-wavelength-sensitive to middle-wave-sensitive cones in the human fovea centralis. Vision Research, 29, 115--128. Coblentz, W. W. & Emerson, W. B. (1918). Relative sensibility of the average eye to light of different colors and some practical applications to radiation problems. U. S. Bureau of Standards Bulletin, 14, 167-236. Dartnall, H. J. A. (1962). The photobiology of visual processes. In Davson, H. (Ed.), The eye (Vol. II, pp. 321 533). New York: Academic Press. Dobson, V. (1976). Spectral sensitivity of the 2-month infant as measured by the visually evoked potential. Vision Research, 16, 367-374. Est6vez, O., Spekreijse, H., van den Berg, T. J. T. P. & Cavonius, C. R. (1975). The spectral sensitivities of isolated human color mechanisms determined from contrast evoked potential measurements. Vision Research, 15, 1205-1212, Gibson, K. & Tyndall, E. P. T. (1923). Visibility of radiant energy. U.S. Bureau of Standards Bulletin, 19, 131-191. Hansen, R. M. & Fulton, A. B. (1989). Psychophysical estimates of ocular media density of human infants. Vision Research, 29, 687-690.
1392
MICHELLE L. BIEBER et al.
Hartman, L. W. (1918). Visibility of radiation in the blue end of the visible spectrum. Astrophysical Journal, 47, 83 95. Hecht, S. & Verrijp, C. D. (1933). Intermittent stimulation by light. IV. A theoretical interpretation of the quantitative data of flicker. Journal of General Physiology, 17, 266-286. Hering, E. (1885). Ueber individuelle Verschiedenheiten des Farbensinnes. Lotos, 6, 142-198. Hickey, T. L. & Peduzzi, J. D. (1987). Structure and development of the visual system. In Salapatek, P. & Cohen, L. (Eds), Handbook of infant perception (Vol. 1, pp. 1-42). New York: Academic Press. Hyde, E. P., Forsythe, W, E. & Cady, F. E. (1918). The visibility of radiation. Astrophysical Journal, 48, 65-88. lves, H. E. (1912). Studies in the photometry of lights of different colours: The addition of luminosities of different colour. Philosophical Magazine, 24, 845 853. Judd, D. B. (1951). Report of the U. S. Secretariat Committee on colorimetry and artificial daylight. In CIE Proceedings, Twelfth Session, Stockholm, I, Part 7, pp. 1-60, Bureau Central CIE, Paris. Kaiser, P. K. (1988). Sensation luminance: A new name to distinguish CIE luminance from luminance dependent on an individual's spectral sensitivity. Vision Research, 28, 455-456. King-Smith, P. E. (1971). The optical density of erythrolabe determined by retinal densitometry. Journal of Physiology, 218, 100-101 p. Kraft, J. K. & Werner, J. S. (1994). Spectral efficiency across the life span: Flicker photometry and brightness matching. Journal of the Optical Society of America A, I1, 1213-1221. Ladenheim, B. & Gordon, J. (1986). Heterochromatic flicker photometry in neonates. Investigative Ophthalmology and Visual Science (Suppl.), 27, 76. Lee, J. & Stromeyer, C. F. Ill (1989). Contribution of human short-wave cones to luminance and motion detection. JournalofPhysiology, 413, 563-593. Lennie, P., Pokorny, J. & Smith, V. C. (1993). Luminance. Journal of the Optical Society of America A, 10, 1283-1293. Lutze, M., Cox, N. J., Smith, V. C. & Pokorny, J. (1990). Genetic studies of variation in Rayleigh and photometric matches in normal trichromats. Vision Research, 30, 149-162. Morrone, M. C., Burr, C. A. & Fiorentini, A. (1993). Development of infant contrast sensitivity to chromatic stimuli. Vision Research, 33, 2535-2552. Moscowitz-Cook, A. (1979). The development of photopic spectral sensitivity in infants. Vision Research, 19, 1133 1142. Nerger, J. L. (1988). The relative numbers of long-wavelength-sensitive to middle-wavelength-sensitive cones in the human fovea and parafovea. Ph.D. dissertation, University of California, San Diego, CA. Nusinowitz, S. & Mauer, D. (1992). The luminous efficiency of colored stimuli for the newborn. Investigative Ophthalmology and Visual Science (Suppl.), 33, 714. Padmos, P. & Norren, D. V. (1972). The vector voltmeter as a tool to measure electroretinogram spectral sensitivity and dark adaptation.
Im,estigative Ophthalmology, 11,783-788. Peeples, D. R. & Teller, D, (1978). White-adapted photopic spectral sensitivity in human infants. Vision Research, 18, 49-53. Pitt, F. H. G. (1935).)(IV. Characteristics of dichromatic vision. Special Report Series, No. 200. London: Medical Research Council. Pokorny, J., Smith, V. C. & Lutze, M. (1989). Heterochromatic modulation photometry. Journal of the Optical Society of America, 6, 1618 1623. Potts, A. M. & Nagaya, T. (1965). Studies on the visual evoked response. 1. The use of the 0.06 degree red target for evaluation of foveal function. Investigative Opththalmology, 4, 303 309. Regan, D. (1970). Evoked potential and psychophysical correlates of changes in stimulus colour and intensity. Vision Research, 10, 163 177. Ruddock, K. H. (1972). Light transmission through the ocular media and macular pigment and its significance for psychophysical
investigation. In Jameson, D. & Hurvich, L. M. (Eds), Handbook ~/ sensoo, physiology Vol. VII/4, Visual po,chophysics. Berlin: Springer. Rushton, W. A. H. (1963). The density of chlorolabe in the foveal cones of the protanope. Journal of Physiology, 168, 360-373. Schultze, M. (1866). Ueber den gelben Fleck der Retina, seinen Einfluss auf normales Sehen und auf Farbenblindheit. Bonn: von Max Cohen & Sohn. Siegfried, J. B. (1978). VECP: Its spectral sensitivity. In Armington, J. C., Krauskopf, J. & Wooten, B. R. (Eds), Visualpsvchophysics and physiology (pp. 257 266). New York: Academic Press. Siegfried, J. B., Tepas, D. I., Sperling, H. G. & Hiss, R. H. (1965). Evoked brain potential correlates of psychophysical responses: Heterochromatic flicker photometry. Science, 149, 321-323. Smith, V. C. & Pokorny, J. (1975). Spectral sensitivity of the foveal cone photopigments between 400 and 500nm. Vision Research, 15~ 161-171. Speranskaya, N. I. (1959). Determination of spectrum color coordinates for twenty-seven normal observers. Optics and Spectroscopy, 7, 424428. Stabell, U. & Stabell, B. (1980a). Variation in the density of the macular pigmentation and in short-wave cone sensitivity with eccentricity. Journal of the Optical Society of America, 70, 706-711, Stabell, B. & Stabell, U. (1980b). Spectral sensitivity in the far peripheral retina. Journal of the Optical Society of America, 70, 959-963. Stiles, W. S. & Burch, J. M. (1959). N, P. L. colour-matching investigation: final report. Optica Acta, 6, 1-26. Stockman, A., MacLeod, D. 1. A. & DePriest, D. D. (1987). The temporal properties of the human short-wave photoreceptors and their associated pathways. Vision Research, 31, 189 208. Teller, D. & Lindsey, D. T. (1989). Motion nulls for white versus isochromatic gratings in infants and adults. Journal of the Optical Society of America, A, 12, 1945-1954. Verriest, G. (1971). L'influence de l'fige sur les fonctions visuelles de l'homme. Bulletin De L'Academie Royale De Medicine De Belgique, I1,527 577. Vos, J. J. (1978a). Colorimetric and photometric properties of a 2' fundamental observer. Color Research and Application, 3, 125-128. Vos, J. J. (1978b). Tabulated characteristics of a proposed 2' fundamental observer. Institute for Perception, Soesterberg, The Netherlands. Vos, J. J. & Walraven, P. L. (1971). On the derivation of the foveal receptor primaries. Vision Research, I1,799-818. Vos, J. J., Est6vez, O. & Walraven, P. L. (1990). Improved color fundamentals offer a new view on photometric additivity. Vision Research, 30, 937 943. Weale, R. A. (1988). Age and the transmittance of the human crystalline lens. Journal o[' Physiology, 395, 577-587. Werner, J. S. (1982). Development of scotopic sensitivity and the absorption spectrum of the human ocular media. Journal of the Optical Society of America, 5, 247 258. Werner, J. S., Donnelly, S. K. & Kliegl, R. (1987). Aging and human macular pigment density: Appended with translations from the work of Max Schultze and Ewald Hering. Vision Research, 27, 257-268. Wooten, B. R. (1972). Photopic and scotopic contributions to the human visually evoked cortical potential. Vision Research, 6, 1647-1660. Wooten, B. R., Fuld, K. & Spillmann, L. (1975). Photopic spectral sensitivity of the peripheral retina. Journal of the Optical Society of America, 65, 334-342. Yuodelis, C. & Hendrickson, A. 0986). A qualitative and quantitative analysis of the human fovea during development. Vision Research, 26, 847-855.
Acknowledgements--Supported by NICHD grant HD19143. We thank Karen Fox, Catie Johnston, Susan McEvoy, Chris Monson, Patrick McFadden, Jamie Phillips and Kim Atkin for assistance in testing infants.
~
Pergamon
0042-6989(94)00245-2
Copyright ~'~ 1995 Elsevier Science Ltd Printed in Great Britain. All rights reserved 0042-6989/95 $9.50 + 0.00
Spectral Efficiency Measured by Heterochromatic Flicker Photometry is Similar in Human Infants and Adults MICHELLE L. BIEBER,* V|CKI J. VOLBRECHT,t JOHN S. WERNER* Received 3! January 1994; in revised form 20 July 1994; in final form 22 September 1994
Spectral efficiency functions based on heterochromatic flicker photometry (HFP) were measured for three adults and 42 infants using a rapid visually-evoked potential (VEP) method. A 5°-diameter, broadband standard (0.6 cd/m2) was presented in square-wave counterphase (15 Hz) with one of 13 monochromatic lights (420-660 nm; 20 nm steps). The intensity of the monochromatic light was continuously varied while extracting the phase-locked VEP amplitude of the fundamental component. HFP functions measured psychophysically by the method of adjustment were also obtained for the adults. Adult HFP functions from the two methods were found to be essentially the same. Both of these functions were compared to Vos'-modified 2° V(7.) function and to the 10° CIE V(7.) function. The mean adult data were slightly better fit to the 2° V(7.) function than to the 10° CIE V0~) function, although there was an elevation in sensitivity at 420 and 440 nm. Infant HFP functions were similar to Vos' modified V(~) except for an elevation in efficiency at short wavelengths. The mean infant HFP function agreed better with the 10°CIE V(7.) function than Vos'-modilled V(7.) function, but infant sensitivity was elevated by 0.4 log units at 420 nm compared to the 10° CIE observer. The elevation found at short wavelengths for both adults and infants is attributed to individual and age-related variation in the density of the ocular media, and to reduced macular pigment screening resulting from use of a 5° field size. An infant spectral efficiency function based on the HFP data collected in this study was modeled. This function may serve to equate stimuli with respect to luminosity for the average infant. Flicker photometry Evokedpotential Infant color vision
The Commission Internationale de l'l~clairage (CIE) defined a function, known as the CIE V(k) function, to represent the spectral efficiency of a standard adult observer under photopic conditions. This function was originally based on data from a large number of observers using hetertochromatic flicker photometry (HFP), step-by-step brightness matching, and direct brightness matching (Coblentz & Emerson, 1918; Hyde, Forsythe & Cady, 1918; Gibson & Tyndall, 1923; Hartman, 1918), but revisions have been proposed several times (Judd, 1951; Vos, 1978a). The CIE has also standardized a large-field V(~.) function, which was derived from 10° color-matching data (Stiles & Burch, 1959; Speranskaya, 1959). While the CIE V(X) functions have proven useful in a wide variety of applications in science and industry, it is often necessary to take into account individual differences in spectral efficiency, or what Kaiser (1988) has called sensation luminance. In this case, HFP is the *Department of Psychology, Campus Box 345, University of Colorado, Boulder, CO 80309~0345, U.S.A. tColorado State University, Ft Collins, CO 80525, U.S.A. VR 35/10 B
procedure of choice because, unlike step-by-step and direct brightness matching, the HFP function conforms to Abney's law of additivity, at least over relatively moderate photopic luminance levels (Ives, 1912). The HFP function is thought to represent the sensitivity of an achromatic mechanism, carried by the magnocellular retinostriate pathway and possibly also by parvocellular fibers that multiplex chromatic and achromatic signals (Lennie, Pokorny, & Smith, 1993). This function can be modeled quite well by summing the outputs of the long-wave sensitive (L-) and middle-wave sensitive (M-) cones, in a ratio of approx. 2: 1, provided that the sensitivities are corrected for the spectrallyselective absorption of the ocular media and macular pigment. It has also been suggested that there is a small subtractive component from the short-wave sensitive (S-) cones to the HFP function (Stockman, MacLeod & DePriest, 1987; Lee & Stromeyer, 1988; Vos, Est~vez & Walraven, 1990); however, this input is only demonstrated under special conditions, and therefore can be ignored for most purposes. The most important factors to account for individual
1385
1386
M1CHELLE L. BIEBER et al.
and age-related variations in HFP functions are likely to be differences in the density of the ocular media and macular pigment (Ruddock, 1972). For example, a progressive loss in short wavelength efficiency has been observed in HFP functions of normal observers between the ages of 10 and 80 yr (Verriest, 1971; Ruddock, 1972; Kraft & Werner, 1994) and this change can be explained simply by the age-related increases in the density of the ocular media. Because the ocular media density is lower in human infants compared to young adults (Werner, 1982; Weale, 1988; Hansen & Fulton, 1989), an additional elevation at short wavelengths might be expected for infant HFP functions. Depending on the field size of the stimuli, this elevation might be further enhanced due to the reduced density of macular pigment in the infant eye (Schultze, 1866; Hering, 1885; Bone, Landrum, Fernandez & Tarsis, 1988). Still other differences between the HFP functions of human infants and adults may result from morphological and physiological immaturities in the infant retina and visual pathways. The short, fat cones in the foveal region (Abramov, Gordon, Hendrickson, Hainline, & Dobson, 1982; Yuodelis & Hendrickson, 1986) are inefficient in capturing light due to poor wave-guide properties (Banks & Bennett, 1988). In addition, while the postreceptoral processing of cone signals by the magnocellular pathway may account for the adult HFP function (Lennie et al., 1993), the differential rate of maturation of the parvocellular and magnocellular pathways may contribute to differences in the HFP functions of human infants. Anatomical evidence suggests that the parvocellular pathway develops at a faster rate postnatally than does the magnocellular pathway (Hickey & Peduzzie, 1987). Despite previous measures of infant spectral sensitivity (Dobson, 1976; Peeples & Teller, 1978; Moskowitz-Cook, 1979; Ladenhiem & Gordon, 1986; Anstis, Cavanagh, Maurer & Lewis, 1986; Teller & Lindsey, 1986; Nusinowitz & Maurer, 1992), a complete spectral efficiency function based on HFP is not yet available. In this study, HFP functions were measured for 42 infants and three adults covering a broad spectral range (4200660nm), using a rapid visually-evoked potential (VEP) method similar to that described by Regan (1970; see also Padmos & Norren, 1972; Est6vez, Spekreijse, van den Berg & Cavonius, 1975). The VEP method was first validated by comparing adults' HFP functions obtained with the VEP to those obtained with the psychophysical method of adjustment. The results of this comparison indicate that the VEP method is capable of accurately measuring HFP spectral efficiency, consistent with previous comparisons of the VEP and psychophysics (Siegfried, Tepas, Sperling & Hiss, 1965; Regan, 1970; Siegfried, 1978). This method was then used to measure spectral efficiency in infants. The results indicate that the average infant HFP function is similar to the 10° CIEV0~) function except for an elevation in sensitivity at 420nm. This is consistent with reduced macular pigment screening as a result of using a 5° stimulus (4400520 nm), and with age-related variation in the ocular media
(420 nm). A spectral efficiency function based on the mean infant HFP data was modeled. This modeled function may be useful for equating stimulus luminosity in future infant research.
METHODS
Subjects Three female adults (22, 35, and 41 yr of age), served as observers. All displayed normal color vision when tested with the Neitz anomaloscope, Farnsworth panel D-15 test, and the American Optical HRR pseudoisochromatic plates. Ninety full-term infants were recruited from birth announcements in the local newspaper to participate in this study. Data were successfully collected from 21 2-month-old infants (14 male and 7 female) and 21 4-month-old infants (10 male and 11 female). Each infant was tested within -I-7 days of their 2nd- or 4th-month birthday. Each infant participated in 1-4 testing sessions that lasted approx. 45-60 min. The parents of all babies reported a negative family history of color deficiency, or one that was unlikely to affect the baby (verbal reports suggested that the mothers of two female infants could be heterozygous for a congenital red-green color defect).
Stimulus and apparatus Stimuli. A 5°, circular stimulus was presented as a broadband white light (CIE coordinates x,y = 0.37, 0.40; 4516 K; 0.6 cd/m 2 ~20 td) alternating in 15-Hz squarewave counterphase with one of 13 monochromatic lights (420°660 nm; 20 nm steps). The intensity of the monochromatic light was varied continuously in either an increasing or a decreasing series (30 sec each direction) with a computer-driven neutral density wedge (2 log unit range). Infants were held 32 cm from the stimulus by one of their parents. Adult subjects were positioned 32 cm from the stimulus by use of a forehead and chin rest. No fixation point was used. Optical system. The light source was a 1000-W xenon arc lamp operated at 980W with a d.c. power supply. Spectral lights were produced with a holographic grating monochromator (Instruments SA, Model H-20). Conventional optics were used to collimate and focus the light from two channels to a common focus at a sectored mirror that produced counterphase alternation between the broadband and monochromatic light. The intensity of the stimulus was controlled by calibrated neutral density filters and neutral density wedges. The light from both channels then passed through a field stop to be projected onto a rear-view projection screen inside a dark, electrically-shielded chamber. Radiometric measurements were made with a silicon photodiode and a linear read-out system (United Detector Technology, 81 optometer) that was calibrated against a standard traceable to the National Institute of Standards and Technology. A He-Ne laser (Spectra
INFANT SPECTRAL EFFICIENCY
Physics) and mercury-line interference filters were used to calibrate the monochromator. The spectral bandwidth of the monochromator was 8 nm at half power. The luminance and correlated color temperature of the broadband field were measured with a photometer/spectroradiometer (Photo Research, Model PR703-A). Recording s.vstem. Bipolar VEPs were recorded with SensorMedics EEG electrodes and Beckman earclip electrodes. For the infants, electrodes were placed 1 and 3 cm above the inion, and the ground was placed on the forehead. For the adults, electrodes were placed 2 and 8 cm above the inion and a ground was placed on the ear lobe. Inter-electrode impedance, measured before and after testing with a Grass Impedance Meter (model EZM5B), was ~<5000 Q for the adults and ~<15,000 Q for the infants, typical levels for adults and infants (Dobson, 1976; Moscowitz-Cook, 1978). The EEG signal was led first through an isolation amplifier (Analog Devices, Model 273K) and then into a battery-powered pre-amplifier (Princeton Applied Research, Model 113). The nominal gain of the pre-amplifier was 10,000 and the low and high frequency roll-offs were set at 0.3 and 30.0 Hz, respectively. Both the isolation amplifier and the pre-amplifier were in the shielded chamber with the subject. The amplified and filtered EEG signal was fed into an oscilloscope and then into a vector voltmeter (Ithaco Dynatrac 3, model 393). The vector voltmeter was synchronized to the chopper at 15 Hz, with a 1.25 sec time constant. The vector voltmeter uses the lock-in principle to selectively amplify the phase-locked amplitude of a specific frequency component of the signal. Because the vector voltmeter extracts the VEP in real time, continuous changes in amplitude could be plotted as a function of the intensity of the monochromatic light. We found the fundamental component of the VEP amplitude agreed well with psychophysical measures under our conditions, consistent with previous work by Siegfried et al. (1965; cf. Regan, 1970).
Procedure Psychophysics. Psychophysical HFP data were collected from adult subjects following 10min of dark adaptation. Using the method of adjustment, the observer's task was to vary the intensity of the monochromatic light until flicker was minimized. Psychophysical HFP spectral efficiency was defined by the log reciprocal quanta required to minimize flicker. The wavelength of the stimuli was varied in random order. Each adult participated in three 60-min sessions. Adult VEPs. Testing began after 10min of dark adaptation. During testing, adult subjects were continuously viewed with an infrared illuminating system and video camera to monitor head and eye movements, as well as blinking. If the subject moved or blinked, recording was momentarily interrupted. Evoked responses were recorded as the intensity of the monochromatic light was continuously decreased or increased for 30 sec. VEP-HFP spectral efficiency was defined by the log reciprocal quanta required to produce a minimum amplitude of the VEP. Wavelength was varied in random
1387
order. Each adult subject participated in 3-5 recording sessions that lasted approx. 1.5-2 hr. Infant VEPs. The procedure used to measure infant VEPs was almost identical to the procedure used for the adults. After the electrodes were in position, infants were brought into the shielded chamber and seated on a parent's lap. Prior to testing, infants were dark adapted for 5 min. A control study with one of the adult subjects indicated that there was no observable difference between psychophysical HFP functions obtained after 5 or 10 min of dark adaptation. During testing, infants were viewed continuously on a video monitor so that recording could be stopped whenever the infant was not attending to the stimulus. Unlike adults, it was usually not possible to record responses from infants to all 13 wavelengths in one session. Each infant participated in 1 4 sessions that lasted approx. 30-60 min each. RESULTS Figure 1 shows typical, individual and mean VEP records obtained from adult and infant subjects. Relative VEP amplitude is plotted as a function of the relative log quanta of the monochromatic light. For each adult observer, 4-15 (mean = 8.5) VEP records were obtained at each wavelength, and for each infant observer, 1-5 (mean=2.3) VEP records were obtained at each
Adult (440 nm)
Infant (440 nm) Record 1
Q.
"0
.__= E
< 13_ LLI >
3
.>_ t~ nMean
Vean I
I 1
0
1
Relative Log Quanta F I G U R E 1. Typical VEP records obtained from an infant and adult observer are shown in the left and right panels respectively, Relative amplitude of the VEP response is plotted as a function o f the intensity of the monochromatic light. The mean of each set of three records is shown with the best-fitting parabola. The VEP criterion response was defined by the m i n i m u m of the polynomial function.
1388
MICHELLE L. BIEBER et al.
wavelength. Notice that the individual and mean VEP records have a V or U shape. The minimum amplitude of the VEP of each of these records was found to correspond to the psychophysically-measured point of minimum flicker (i.e. equal sensation luminance between the monochromatic light and standard).
i
Adult spectral efficiency To validate the VEP measure of spectral efficiency, a comparison was made between the psychophysically-obtained H F P functions of each adult observer and their electrophysiologically-obtained H F P functions. Figure 2 presents these functions as well as a mean function for the three adult subjects. Log reciprocal quanta required for the criterion response is plotted as a function of wavelength. The position of each curve along the axis of ordinates is arbitrary. The solid and open symbols represent VEP and psychophysical measures, respectively. The V E P - H F P functions have been normalized to the psychophysical H F P functions using a least-squares criterion. Both of these functions are plotted with Vos' 2 ° V(;L) function (solid curve) and the 10° CIE V(X) function (dashed curve) for comparison. The 2 ° and 10° V(;L) functions have been normalized to the psychophysically obtained H F P functions at 580 nm, where all three functions are expected to have the same shape and are
f
r
[
l
l
,
Data analysis As expected, some records contained noise artifacts (due to abrupt movements or fussy behavior), so an objective method was needed to determine whether a record should be retained for further analysis. After each session, individual records were digitized and fitted with a polynomial regression equation. All records that were fit significantly better by an upward facing parabola compared to a straight line (0t = 0.001) were averaged for each subject and wavelength. Overall, 5-40% of the records had to be eliminated for each subject. While forty percent may sound large, some rather active infants provided only five records, and two had to be eliminated. The mean VEP function for records meeting the criterion was computed at each wavelength. As shown in Fig. 1, a parabola was fitted by the method of least-squares to each of these mean amplitude vs intensity functions, and the log reciprocal quanta required to produce the minimum of the parabola was taken as the H F P criterion response. A parabola was chosen for convenience in locating the minimum of the V-shaped functions. No theoretical significance is attached to this fitting method, although it may be noteworthy that the coefficient of the x 2 parameter varied little with wavelength, implying a relatively fixed shape as a function of wavelength. In addition, based on results of Pokorny, Smith and Lutze (1989), we compared H F P functions for 12 infants and one adult derived from fits to the VEP data in linear rather than log stimulus space. An analysis of variance indicated that there was no significant main effect (F1,13=0.257, P=0.6207) or interaction (F~j.t43--- 1.108, P = 0.3574) between the fitting method and wavelength.
l
MB
t-
v
r~ ,t-
.o iii
,~
vv
I1)
._> I1)
rr
(:7) 0
Mean
o
II 400
I
I 500
I
I
1 I
600
700
Wavelength (nm) FIGURE 2. Log relative efficiency (quanta) is plotted as a function of wavelength (nm). Individual adult functions and the mean adult function are arbitrarily separated by 2 log units. HFP functions obtained electrophysiologicallyand psychophysically are represented by the solid and open symbols respectively. Vos'-modified 2° V(~,) function (solid curves) and the 10° CIE V(~,) function (dashed curves) have been normalized at 580 nm to the psychophysical HFP functions for comparison.
relatively unaffected by variations in ocular media density. As shown in Fig. 2, there is good agreement between the VEP and psychophysical spectral efficiency functions for each observer at most wavelengths, as previously reported for studies with adults using similar methods (Siegfried, et al., 1965; Regan, 1970). The mean adult function is fit slightly better to the 2 ° V(X) function than to the 10° CIE V(X).
Infant spectral efficiency H F P spectral efficiency functions with nine or more wavelengths were obtained from 20 of the 42 infants, using the VEP criterion response. Log reciprocal quanta is plotted for these data as a function of wavelength in Fig. 3. The H F P functions are arbitrarily separated by 1 log unit along the ordinates to better display the data. A least-squares criterion fit for wavelengths ~>520 nm to Vos' modified 2 ° V(X) function (where age-related changes in ocular media density are minimal) was determined for all the infant functions. All remaining infant functions are normalized in the same way unless otherwise stated. This was done to prevent masking an elevation that might be present in the infant data at short
INFANT SPECTRAL EFFICIENCY wavelengths, due to reduced ocular media density (Werner, 1982; Weale, 1988; Hansen & Fulton, 1989) and reduced macular pigment density (Schultze, 1866; Hering, 1885; Bone et al., 1988). The Vos 2 ° V(X) function was used rather than the 10 ° C I E V(k) function because it provided the best fit to the mean adult data. The remaining comparisons using the mean infant data are made to both of these standard functions. There is a close correspondence between most of the infant data sets and the 2 c' V(~,) function. Spectral efficiencies were also measured for 22 of the 42 infants at a subset of the spectral points shown in Fig. 3. These data are consistent with the more complete data sets shown. The infant data are somewhat more variable than the adult data presented in Fig. 2. This is to be expected from the more variable infant behavior during data collection and also because more records were obtained at each wavelength for the adult observers than for the infant observers. The latter factor may be less important than the former. A spectral efficiency function for one adult subject was redetermined using only two randomlyselected VEP records per point (close to the mean number of records obtained from infants). This function looked nearly the same as that determined from the full data set, except for one deviant point. The mean absolute
I
I
I
I
I
I
I
I
2-Month Infants
I
I
~
I
I
I
4-Month Infants
,
•? o
E -1 I:)-
/
0 c-
._o w 0
._> OC: 0 --J
^
)
o- C
yl°/ 400
l
i 500
i
L i 600
i 700
I
Ifl
I
400
500
I
L
600
I
i
700
W a v e l e n g t h (nm)
FIGURE 3. Log relative efficiency(quanta) is plotted for all complete data sets (10 2-month olds and 10 4-month olds) as a function of wavelength (nm). Each data set was normalized to Vos'-modified 2° V(X) function (solid curve) using a least-squares criterion for wavelengths >t 520 nm.
1389
¢.0.0
>" e-
-0.5
.~_ 0
-1.0
lit 11) >
-1.5
a) nr'
-2.0
0 I
I
I
I
L
I
L
400
450
500
550
600
650
700
Wavelength (nrn)
FIGURE 4. Log relative efficiency(quanta) is plotted as a function of wavelength (nm) for the mean infant ((D) and mean adult (0) VEP-HFP functions. Overlapping data for infants and adults are denoted by tD. Both mean functions were normalized to Vos'-modified 2° V(~,) function (solid curve) using a least-squares criterion for wavelengths >~ nm. The 10° CIE V(~,)function (dashed curve) has been normalized at 580 nm to Vos' modified 2 V(X)function.
difference between the two H F P functions was 0.13 log units. The mean spectral efficiency of all infants is shown in Fig. 4 (open circles). Averaging the 2- and 4-month data was justified because there are no systematic differences between the mean functions of the two age groups (average deviation 0.07). There is a close correspondence between the 10 ° V(X) function and the mean infant spectral efficiency function except for a 0.39 log unit elevation in the mean infant function at 420 nm. Also plotted in Fig. 4. is the mean adult VEP spectral efficiency function (solid circles). This function has been similarly normalized to the 10 ° and 2 ° V(~,) functions. As shown in this plot, the adult function is fit better to the 2': V(~,) function except where there is an elevation at 420 and 440 nm. An elevation in sensitivity at short wavelengths from approx. 440-500 nm would be expected for both infants and adults due to the use of a 5 ° field compared to a 2 ° field. With the former, the macular pigment will influence sensitivity less because of its lower density at this retinal eccentricity (Stabell & Stabell, 1980a; Werner, Donnelly & Kliegl, 1987). The macular pigment should not have a significant influence at 420 nm; hence, this elevation is more likely due to lower than average ocular media density in infants and these adults. The elevation in sensitivity at 420 nm observed for infants compared to identically-tested adults and to the 10 ° V(X) function is expected because of even lower ocular media density in infants (Werner, 1982). DISCUSSION H F P spectral efficiency measured with the VEP is similar for both 2- and 4-month-old infants. The mean infant function is essentially the same as the 10°V(~.) function (solid curve), except for an elevation in infant
1390
MICHELLE L. BIEBER et al.
sensitivity at the short wavelength extreme of the spectrum. These results are in close agreement with previous investigations using different methods (Dobson, 1976; Moskowitz-Cook, 1979). Because the previous studies used methods other than HFP, their results may represent the efficiency of different neural mechanisms. The elevation in efficiency at short wavelengths found in this study might, at first glance, be thought to be due to rod and peripheral cone contributions, both of which would enhance sensitivity to short wavelength light (Wooten, Fuld & Spillmann, 1975; Abramov & Gordon, 1977; Stabell & Stabell, 1980b). There is good reason to believe however, that rods did not contribute to the elevation at short wavelengths found in the present study. First, the temporal frequency of the stimulus was 15 Hz, a frequency that rods do not resolve very well under the conditions used in this study (Hecht & Verrijp, 1933). Second, the VEP is usually dominated by signals originating in the fovea (Potts & Nagaya, 1965; Wooten, 1972; Siegfried, 1978). Finally, the field size of the stimulus was 5° . In the newborn eye, the rod-free zone (foveola) is roughly 5.4 °, decreasing to about 2.3 ° at maturity (Banks & Bennett, 1988). Of course, due to unsteady fixation, some contribution from the peripheral cones cannot be ruled out. We argue, instead, that the elevation in efficiency at short wavelengths for infants relative to the 2 ° and 10° standard observers may be explained by developmental changes in the density of the ocular media (420 nm), and reduced macular pigment filtering due to the use of a 5 ° stimulus field size (440--500 nm). Experimental evidence has shown that the density of the lens pigment in the infant eye is low, and increases over the life span (Werner, 1982; Weale, 1988; Hansen & Fulton, 1989), but at every age, there is also considerable individual variation in density that contributes to individual differences in spectral sensitivity (Ruddock, 1972). The macular pigment is also reduced in the infant eye (Schultze, 1866; Hering, 1885) increasing to adult levels over the first few years of life (Bone et al., 1988). At 5° however, macular pigment does not have high density in adults and presumably also not in infants so it is unlikely to contribute to adult-infant differences in spectral efficiency.
*S-conecontributions are difficultto detect in the luminosityfunction based on curve fitting, but can be identified under certain experimental conditions. Additional analyses were conducted to determine whether S-conecontributions might be detectableunder our testing conditions. STEPIT was used to minimize the squared log ratio differencebetweenmean VEP-HFP functionsand the sum of the L-, M- and S-cone fundamentals in the same manner as described above, except S-conecontributions were allowedto vary. For the infant function, STEPITestimatedan S-conescalar of 0.006 and a lens scalar of 0.28. This additional S-coneparameter did not improve the fit to the mean infant VEP function (Fj.~3=0.13994, P>0.05). For a comparably modeled adult function, STEPIT estimated an S-conescalar of 0.006 and a lens scalar of 1.0, exactly equal to that of the standard observer.As with the infant function, allowing S-cones to vary didn't provide a significantlybetter fit to the mean adult VEP function (F~.~3= 1.367, P > 0.05). Giventhe lack of statistical justification, possible S-cone contributions were not included in the modeled infant function.
Modeling the HFP function in the infant The combined signals of the L- and M-cones, in a ratio of approx. 2:1, provides an excellent fit to V00, provided the sensitivities are corrected for the spectrally-selective absorption of the ocular media and macular pigment. In addition to known differences between infants and adults in their preretinal screening pigments, another difference that might contribute to differences in H F P spectral efficiency functions is the length of the cone outer segments. The shorter outer-segments of infant cones (Yuodelis & Hendrickson, 1986) will result in a reduction in photopigment density which is manifested as a narrowing of the shape of the photopigment absorption spectrum (Dartnall, 1962). Sensitivity reductions at the spectral extremes would, therefore, be expected for infants relative to adults. A quantitative evaluation of this effect, assuming maximal differences between infants and adults, revealed that reducing photopigment density from 0.35 (Rushton, 1963; King-Smith, 1971) to 0.001 decreases sensitivity at the spectral extremes only slightly (0.14 and 0.15 log unit reduction in efficiency at 420 and 660 nm respectively). At short wavelengths, this loss in sensitivity would be expected to be more than compensated by the less dense ocular media. At longer wavelengths, reduced photopigment density in infants might be expected to lower sensitivity relative to adults, but it appears not to be detectable in our measurements. Alternatively, the reduction in photopigment density for infants may not be as high as that assumed above. Given our results, possible effects of reduced photopigment density can be ignored for purposes of modeling the mean infant H F P function. The mean infant H F P function was modeled to determine a standard infant spectral efficiency function that may be used to equate chromatic stimuli in terms of luminosity. STEPIT (Chandler, 1965) was used to minimize the squared log ratio difference between the mean infant spectral efficiency function and a sum of Land M-cone fundamentals* (Smith & Pokorny, 1975), no macular pigment screening and a variable multiplicative scalar to adjust ocular media density. The shape of the ocular media density spectrum was assumed to follow the tabled values for the standard observer of Vos (1978b). A 2:1 ratio of L : M cones was assumed based on previous evidence from adult psychophysical studies (Vos & Walraven, 1971; Cicerone & Nerger, 1989). It should be noted, however, that this assumption is not critical as long as the ratio favors L-cones (Pitt, 1935; Nerger, 1988; Kraft & Werner, 1994; cf. Lutze, Cox, Smith & Pokorny, 1990). This implies that, contrary to previous claims (Morrone, Burr, & Fiorentini, 1993), similarity of infant and adult red-green flicker nulls does not necessarily imply similar L : M cone ratios. The STEPIT analysis estimated an ocular media density scalar of 0.51 for the infant, which is in close agreement with previous estimates based on scotopic spectral sensitivity (Werner, 1982). The modeled function is shown in Fig. 5 by the bold solid curve along with the mean infant data (open circles). With the exception of two
INFANT SPECTRAL EFFICIENCY I el. t~
I
I
I
I
I
[
0.0 -0.5
o¢ -
-1.0
"5 -1.5 lit
._> ¢r 0 ..-I
-2.0 -2.5 -3.0
400
I
I
I
I
I
I
450
500
550
600
650
700
Wavelength (nm)
FIGU RE 5. Log relative efficiency is plotted as a function of wavelength (nm). Open circles represent the mean infant HFP function. Error bars denote + 1 SD. The bold solid curve represents the function modeled to the infant data. Vos'-modified 2° V(X)function (solid curve) and the 10c' CIE V(X) (dashed curve) were normalized at 580 nm to the modeled function for comparison.
points (460 and 620 nm), the mean infant data agree well with the modeled infant function. The Vos-modified 2 ° V(k) function (solid curve) and the 10 ° CIE V(~,) function have been normalized at 580 nm to the theoretical infant function for comparison. The modeled infant function fits extremely well to the 10 ° V(X) function except at 420 nm
TABLE 1. Modeled 5° infant spectral efficiency (log quanta) Wavelength (nm)
Efficiency
Wavelength (nm)
400 405 410 415 420 425 430 435 440 445 450 455 460 465 470 475 480 485 490 495 500 505 510 515 520 525 530 535 540 545 550
- 1.754" - 1.624" - 1.494" - 1.363" - 1.233 - 1.171 - 1.110 - 1.048 -0.987 -0.939 -0.890 -0.842 -0.794 -0.726 -0.657 -0.588 -0.519 -0.454 -0.389 -0.324 -0.259 -0.215 -0.171 -0.127 -0.082 -0.063 --0.044 -0.025 -0.006 --0.005 -0.003
555 560 565 570 575 580 585 590 595 600 605 610 615 620 625 630 635 640 645 650 655 660 665 670 675 680 685 690 695 700
Efficiency -0.002 0.000 -0.015 -0.031 -0.046 -0.062 -0.096 -0.130 -0.164 -0.198 -0.252 -0.307 --0.361 -0.415 -0.502 -0.588 -0.674 -0.760 -0.877 -0.994 - 1.111 - 1.228 - 1.370" - 1.511" - 1.653" - 1.795" -- 1.952" --2.109" --2.267* - 2.424*
*Values that are extrapolated beyond the limits of empirical measurement.
139 l
w h e r e r e d u c e d o c u l a r m e d i a filtering is e x p e c t e d to be most pronounced for the wavelength range examined. T h e s p e c t r a l efficiency v a l u e s o b t a i n e d f r o m this a n a l y s i s a r e listed in T a b l e 1 f o r w a v e l e n g t h s b e t w e e n 400 a n d 700nm. A s t e r i s k s d e n o t e t a b l e d v a l u e s t h a t are e x t r a p o l a t e d b e y o n d the s p e c t r a l limits o f the e m p i r i c a l d a t a . W h i l e s p e c t r a l efficiency is s i m i l a r f o r a d u l t s a n d i n f a n t s at m i d d l e a n d l o n g w a v e l e n g t h s , d i f f e r e n c e s at short wavelengths are important. T h e m o d e l e d i n f a n t s p e c t r a l efficiency f u n c t i o n p r o v i d e s a fairly a c c u r a t e d e s c r i p t i o n o f t h e m e a n i n f a n t H F P f u n c t i o n u n d e r s t a n d a r d c o n d i t i o n s o f testing. A s w i t h t h e C I E V(X) f u n c t i o n , this f u n c t i o n m a y be useful in e q u a t i n g s t i m u l i f o r a n a v e r a g e o b s e r v e r . B e c a u s e o f individual differences between observers, however, n e i t h e r the m o d e l e d 5 ° i n f a n t s p e c t r a l efficiency o r the a d u l t C I E V(~,) f u n c t i o n s c a n be e x p e c t e d to a c c u r a t e l y e q u a t e s t i m u l i f o r all i n d i v i d u a l o b s e r v e r s . I n m a n y i n s t a n c e s i n d i v i d u a l s e n s a t i o n l u m i n a n c e will be r e q u i r e d to a c c u r a t e l y e q u a t e s t i m u l i in t e r m s o f l u m i n o s i t y . H o w e v e r , this m o d e l e d s p e c t r a l efficiency f u n c t i o n m i g h t be useful to the r e s e a r c h e r w h o is i n t e r e s t e d in g r o u p d a t a , o r is r e s t r i c t e d b y t i m e o r e q u i p m e n t to d e t e r m i n e individual sensation luminance.
REFERENCES
Abramov, I. & Gordon, J. (1977). Color vision in the peripheral retina. I. Spectral sensitivity. Journal of the Optical Society of America, 67, 195-202. Abramov, I., Gordon, J., Hendrickson, A., Hainline, L., Dobson, V. & LaBossiere, E. (1982). The retina of the newborn human infant. Science, 217, 265-267. Anstis, S., Cavanagh, P., Maurer, D. & Lewis, T. (1986). Early maturation of luminous efficiency for colored stimuli. Investigative Ophthalmology and Visual Science (Suppl.), 27, 264. Banks, M. S. & Bennett, P. J. (1988). Optical and photoreceptor immaturities limit the spatial and chromatic vision of human neonates. Journal of the Optical Society of America A, 5, 2059-2079. Bone, R. A., Landrum, J. T., Fernandez, L. & Tarsis, S. L. (1988). Analysis of macular pigment by HPLC: Retinal distribution and age study. Investigative Ophthalmology and Visual Science, 29, 843-849. Chandler, J. P. (1965). "STEPIT", Quantum chemistry exchange program. Department of Chemistry, Indiana University, Bloomington, IN. Cicerone, C. M. & Nerger, J. L. (1989). The relative numbers of long-wavelength-sensitive to middle-wave-sensitive cones in the human fovea centralis. Vision Research, 29, 115--128. Coblentz, W. W. & Emerson, W. B. (1918). Relative sensibility of the average eye to light of different colors and some practical applications to radiation problems. U. S. Bureau of Standards Bulletin, 14, 167-236. Dartnall, H. J. A. (1962). The photobiology of visual processes. In Davson, H. (Ed.), The eye (Vol. II, pp. 321 533). New York: Academic Press. Dobson, V. (1976). Spectral sensitivity of the 2-month infant as measured by the visually evoked potential. Vision Research, 16, 367-374. Est6vez, O., Spekreijse, H., van den Berg, T. J. T. P. & Cavonius, C. R. (1975). The spectral sensitivities of isolated human color mechanisms determined from contrast evoked potential measurements. Vision Research, 15, 1205-1212, Gibson, K. & Tyndall, E. P. T. (1923). Visibility of radiant energy. U.S. Bureau of Standards Bulletin, 19, 131-191. Hansen, R. M. & Fulton, A. B. (1989). Psychophysical estimates of ocular media density of human infants. Vision Research, 29, 687-690.
1392
MICHELLE L. BIEBER et al.
Hartman, L. W. (1918). Visibility of radiation in the blue end of the visible spectrum. Astrophysical Journal, 47, 83 95. Hecht, S. & Verrijp, C. D. (1933). Intermittent stimulation by light. IV. A theoretical interpretation of the quantitative data of flicker. Journal of General Physiology, 17, 266-286. Hering, E. (1885). Ueber individuelle Verschiedenheiten des Farbensinnes. Lotos, 6, 142-198. Hickey, T. L. & Peduzzi, J. D. (1987). Structure and development of the visual system. In Salapatek, P. & Cohen, L. (Eds), Handbook of infant perception (Vol. 1, pp. 1-42). New York: Academic Press. Hyde, E. P., Forsythe, W, E. & Cady, F. E. (1918). The visibility of radiation. Astrophysical Journal, 48, 65-88. lves, H. E. (1912). Studies in the photometry of lights of different colours: The addition of luminosities of different colour. Philosophical Magazine, 24, 845 853. Judd, D. B. (1951). Report of the U. S. Secretariat Committee on colorimetry and artificial daylight. In CIE Proceedings, Twelfth Session, Stockholm, I, Part 7, pp. 1-60, Bureau Central CIE, Paris. Kaiser, P. K. (1988). Sensation luminance: A new name to distinguish CIE luminance from luminance dependent on an individual's spectral sensitivity. Vision Research, 28, 455-456. King-Smith, P. E. (1971). The optical density of erythrolabe determined by retinal densitometry. Journal of Physiology, 218, 100-101 p. Kraft, J. K. & Werner, J. S. (1994). Spectral efficiency across the life span: Flicker photometry and brightness matching. Journal of the Optical Society of America A, I1, 1213-1221. Ladenheim, B. & Gordon, J. (1986). Heterochromatic flicker photometry in neonates. Investigative Ophthalmology and Visual Science (Suppl.), 27, 76. Lee, J. & Stromeyer, C. F. Ill (1989). Contribution of human short-wave cones to luminance and motion detection. JournalofPhysiology, 413, 563-593. Lennie, P., Pokorny, J. & Smith, V. C. (1993). Luminance. Journal of the Optical Society of America A, 10, 1283-1293. Lutze, M., Cox, N. J., Smith, V. C. & Pokorny, J. (1990). Genetic studies of variation in Rayleigh and photometric matches in normal trichromats. Vision Research, 30, 149-162. Morrone, M. C., Burr, C. A. & Fiorentini, A. (1993). Development of infant contrast sensitivity to chromatic stimuli. Vision Research, 33, 2535-2552. Moscowitz-Cook, A. (1979). The development of photopic spectral sensitivity in infants. Vision Research, 19, 1133 1142. Nerger, J. L. (1988). The relative numbers of long-wavelength-sensitive to middle-wavelength-sensitive cones in the human fovea and parafovea. Ph.D. dissertation, University of California, San Diego, CA. Nusinowitz, S. & Mauer, D. (1992). The luminous efficiency of colored stimuli for the newborn. Investigative Ophthalmology and Visual Science (Suppl.), 33, 714. Padmos, P. & Norren, D. V. (1972). The vector voltmeter as a tool to measure electroretinogram spectral sensitivity and dark adaptation.
Im,estigative Ophthalmology, 11,783-788. Peeples, D. R. & Teller, D, (1978). White-adapted photopic spectral sensitivity in human infants. Vision Research, 18, 49-53. Pitt, F. H. G. (1935).)(IV. Characteristics of dichromatic vision. Special Report Series, No. 200. London: Medical Research Council. Pokorny, J., Smith, V. C. & Lutze, M. (1989). Heterochromatic modulation photometry. Journal of the Optical Society of America, 6, 1618 1623. Potts, A. M. & Nagaya, T. (1965). Studies on the visual evoked response. 1. The use of the 0.06 degree red target for evaluation of foveal function. Investigative Opththalmology, 4, 303 309. Regan, D. (1970). Evoked potential and psychophysical correlates of changes in stimulus colour and intensity. Vision Research, 10, 163 177. Ruddock, K. H. (1972). Light transmission through the ocular media and macular pigment and its significance for psychophysical
investigation. In Jameson, D. & Hurvich, L. M. (Eds), Handbook ~/ sensoo, physiology Vol. VII/4, Visual po,chophysics. Berlin: Springer. Rushton, W. A. H. (1963). The density of chlorolabe in the foveal cones of the protanope. Journal of Physiology, 168, 360-373. Schultze, M. (1866). Ueber den gelben Fleck der Retina, seinen Einfluss auf normales Sehen und auf Farbenblindheit. Bonn: von Max Cohen & Sohn. Siegfried, J. B. (1978). VECP: Its spectral sensitivity. In Armington, J. C., Krauskopf, J. & Wooten, B. R. (Eds), Visualpsvchophysics and physiology (pp. 257 266). New York: Academic Press. Siegfried, J. B., Tepas, D. I., Sperling, H. G. & Hiss, R. H. (1965). Evoked brain potential correlates of psychophysical responses: Heterochromatic flicker photometry. Science, 149, 321-323. Smith, V. C. & Pokorny, J. (1975). Spectral sensitivity of the foveal cone photopigments between 400 and 500nm. Vision Research, 15~ 161-171. Speranskaya, N. I. (1959). Determination of spectrum color coordinates for twenty-seven normal observers. Optics and Spectroscopy, 7, 424428. Stabell, U. & Stabell, B. (1980a). Variation in the density of the macular pigmentation and in short-wave cone sensitivity with eccentricity. Journal of the Optical Society of America, 70, 706-711, Stabell, B. & Stabell, U. (1980b). Spectral sensitivity in the far peripheral retina. Journal of the Optical Society of America, 70, 959-963. Stiles, W. S. & Burch, J. M. (1959). N, P. L. colour-matching investigation: final report. Optica Acta, 6, 1-26. Stockman, A., MacLeod, D. 1. A. & DePriest, D. D. (1987). The temporal properties of the human short-wave photoreceptors and their associated pathways. Vision Research, 31, 189 208. Teller, D. & Lindsey, D. T. (1989). Motion nulls for white versus isochromatic gratings in infants and adults. Journal of the Optical Society of America, A, 12, 1945-1954. Verriest, G. (1971). L'influence de l'fige sur les fonctions visuelles de l'homme. Bulletin De L'Academie Royale De Medicine De Belgique, I1,527 577. Vos, J. J. (1978a). Colorimetric and photometric properties of a 2' fundamental observer. Color Research and Application, 3, 125-128. Vos, J. J. (1978b). Tabulated characteristics of a proposed 2' fundamental observer. Institute for Perception, Soesterberg, The Netherlands. Vos, J. J. & Walraven, P. L. (1971). On the derivation of the foveal receptor primaries. Vision Research, I1,799-818. Vos, J. J., Est6vez, O. & Walraven, P. L. (1990). Improved color fundamentals offer a new view on photometric additivity. Vision Research, 30, 937 943. Weale, R. A. (1988). Age and the transmittance of the human crystalline lens. Journal o[' Physiology, 395, 577-587. Werner, J. S. (1982). Development of scotopic sensitivity and the absorption spectrum of the human ocular media. Journal of the Optical Society of America, 5, 247 258. Werner, J. S., Donnelly, S. K. & Kliegl, R. (1987). Aging and human macular pigment density: Appended with translations from the work of Max Schultze and Ewald Hering. Vision Research, 27, 257-268. Wooten, B. R. (1972). Photopic and scotopic contributions to the human visually evoked cortical potential. Vision Research, 6, 1647-1660. Wooten, B. R., Fuld, K. & Spillmann, L. (1975). Photopic spectral sensitivity of the peripheral retina. Journal of the Optical Society of America, 65, 334-342. Yuodelis, C. & Hendrickson, A. 0986). A qualitative and quantitative analysis of the human fovea during development. Vision Research, 26, 847-855.
Acknowledgements--Supported by NICHD grant HD19143. We thank Karen Fox, Catie Johnston, Susan McEvoy, Chris Monson, Patrick McFadden, Jamie Phillips and Kim Atkin for assistance in testing infants.