See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/229696263

Color memory in children

Article in Color Research & Application · October 2008

DOI: 10.1002/col.20433

CITATIONS READS
6 424

3 authors, including:

Joaquín Pérez-Carpinell Mario Trottini

University of Valencia University of Alicante
19 PUBLICATIONS 268 CITATIONS 39 PUBLICATIONS 435 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Joaquín Pérez-Carpinell on 03 November 2021.

The user has requested enhancement of the downloaded file.

Color Memory in Elderly Adults

Joaquı́n Pérez-Carpinell,1*
Vicente J. Camps,2 Mario Trottini,3
Carmen M. Pérez-Baylach4
1
Depto. de Óptica, Facultad de Fı́sica, Universitat de València, 46100 Burjassot, Spain

2
Depto. de Óptica, E. U. O, Universidad de Alicante, 03080 Alicante, Spain

3
Depto. de Estadı́stica e Investigación Operativa, Facultad de Matemáticas, Universidad de Alicante, 03080 Alicante,
Spain

4
Hospital Clı́nico Universitario, 46010 Valencia, Spain

Received 29 July 2004; revised 31 January 2006; accepted 17 March 2006

Abstract: The methods of simultaneous and successive color (i.e., to judge the ripeness of a fruit, matching clothes to
matching have been compared for a set of five Munsell color make an outfit, etc.), we use the successive or memory color
samples by 50 older adult observers, 25 men and 25 women matching method. However, when we compare pairs of
(ranging in age from 64 to 80 years). From comparison color samples side-by-side, we use the simultaneous color
between this population and one of 50 younger adult observ- matching method.
ers, 25 men and 25 women (in the 20 –27 age range), we can Comparison by color memory is a much more common
deduce, in general, the following: (a) In the elderly adults the task than simultaneous color matching, as this latter is rarely
mean CIELAB total color difference (⌬E*ab) in simultaneous done unless we place both samples so close together that we
color matching is lower than the ⌬E*ab by memory color can observe them in the same area of the fovea.
matching. (b) While younger adults matched well the color of Situations where a remembered color has to be repro-
all the reference tests, the elderly adults matched poorly both duced or a present color has to be judged in relation to a
greens and orange. (c) Younger adults remember the original remembered color are common. Some people have an ex-
color better than do the older adults (P ⫽ 0.007), depending on cellent immediate color memory whereas others do not.
gender and delay time. (d) Although with simultaneous match- After observing a color, some observers are able to retain it
ing, the observer’s gender does not determine significant dif- in their memory and identify it fairly closely after a rela-
ferences, by memory, men matched the color of reference test tively short time, whereas others may remember it incor-
more poorly than did women (P ⫽ 0.0), independently of age, rectly or even forget it completely.
color and delay time, especially for bluish green, violet, and In the literature there is a lot of research related to color
pink. © 2006 Wiley Periodicals, Inc. Col Res Appl, 31, 458 – 467, 2006; memory.1–11 Reviews can be found where both methods of
Published online in Wiley InterScience (www.interscience.wiley.com). DOI color matching have been studied and compared. However,
10.1002/col.20258 we have not found in the perceptual literature, any research
Key words: color memory; simultaneous color matching; which explicitly studies color memory in elderly observers.
elderly adults; gender differences; old–young adults differ- This is despite the fact that life expectancy is still increas-
ences ing, and that the average age of the population in North
America and Europe is approximately 80, with half of the
population reaching the age of 65, and the number of people
INTRODUCTION
aged 65 to 85 in developing countries will increase over the
In daily life, when we compare separate color samples in next 35 years by 50% to 90%.
time, trying to match a given color with a remembered one However, there are many reports of losses in visual
functions associated with nonpathological aging. We found
*Correspondence to: Joaquı́n Pérez-Carpinell (e-mail: joaquin.p.carpinel@
numerous studies on age-related changes in color vision.
uv.es) Lakowski12 indicates that colors are perceived and discrim-
© 2006 Wiley Periodicals, Inc. inated most accurately between the ages of 16 and 35. After

458 COLOR research and application

age 55 to 60 years, there is deterioration in the ability for The spectral saturation discrimination function is well
fine color discrimination, which is greater when the discrim- documented for young adults,23–26 whereas there are very
ination is in the blue–yellow or violet/blue– green axis and few systematic studies of this function and aging.27 Kraft
less if the discrimination is in the red– green axis. and Werner28 demonstrate that with aging there are losses in
These losses in discrimination have been widely studied colorimetric purity discrimination sensitivity. The differ-
using different color tests such as the Farnsworth–Munsell ence in discrimination with retinal illumination from 10 to
100 hue test,13–15 where presumably the reduction in the 250 trolands was relatively small for young observers but
amount of chromatic information available, a consequence greater for elderly observers, which indicates that they suf-
of aging, accentuates the difficulties in carrying out the fer from a decrease in this function at low light levels.
discrimination tasks which this test requires. In the literature, there are numerous references that indi-
These systematic changes in color vision, due to aging, cate physiological differences between men and women,
are similar to the effect produced in a young person putting which can influence the perception and memory of the
a yellow filter in front of the eye.16 Blue light dims more color. It is established that between 15% and 20% of Cau-
than the light of longer wavelengths, which displaces the casian women possess the genetic potential to express four
appearance of white towards yellow. There is a relative classes of retinal photopigments,29,30 which can lead to
darkening of blue-colored objects and a general tendency of differences in color perception comparing to normal trichro-
the rest of the colors toward chromaticities of long wave- mat women.30 Other studies indicated that men and women
lengths. differ in the physiological substrate of color vision. Ham-
Werner and Steele17 report that lenticular senescence pro- mond et al.31 found that in women the average level of
duces a greater reduction in the sensitivity of color vision macular pigment (which filters short-wavelength light) was
mechanisms linked to short wavelengths. This study also sug- lower than in men. Women showed also greater responsive-
gests that about 30% to 40% of the sensitivity losses for SWS ness to long wavelengths in color discrimination.32 Besides,
cones (440 –500 nm) were due to age-related changes in col- in a previous article, Pérez-Carpinell et al.,7 with young
oration of the ocular media, the remaining loss ascribable to adults, found that by memory, when only considering the
receptoral and/or postreceptoral changes.17,18 average for all colors and delay times used in the experi-
Boettner and Woltner19 report that the transmission of ment, men matched mean color more poorly than women
visible radiation through the vitreous and aqueous humors (P ⬍ 0.05).
does not vary with age. They also report that while the These findings of difference in color vision related with
transmission of infrared radiation through the eye, as a aging and gender motivated us to examine whether and to
whole, is independent of age, ultraviolet transmission tends what extent these factors also affect color memory. In
to decrease with age. particular, in this study, we obtain and compare variations
Knoblauch et al.15 report that, using the Farnsworth– of CIELAB total color difference, ⌬E*ab, and variations of
Munsell 100 hue test, the effect of aging on visual function the components of color difference (CIELAB lightness dif-
is similar to that obtained in young observers when light ference, ⌬L*, CIELAB chroma difference, ⌬C*ab, and
levels decrease. Thus, observers between the ages of 20 and CIELAB hue difference, ⌬H*ab), between the reference test
40 at 5.7 lux show the same behavior as 50 to 60-year-old and the adjacent matching sample (delay time interval 0 s)
observers between 18 and 57 lux, and similar to 70-year- and between the reference test and the memory matching
olds at 180 lux. Thus color vision in the elderly adults can sample for different delay time intervals (15 s, 15 min, and
be improved by using higher light levels while avoiding 24 h) in two different population ages (elderly and young
glare to which they are more sensitive. adults), and for both genders. The study addresses three
In addition, Lakowski12 and Verriest et al.,20 by simulat- main research questions: (1) Are there any age-related color
ing the effects of old age with neutral density filters in front memory differences between elderly and young adults? If
of young adult eyes, showed that senile changes in foveal so, do they depend on gender and/or delay time? (2) Are
color vision are mainly due to the progressive yellowing of there any gender-related color memory difference? If so, do
the nucleus of the lens. Weale21 indicates that this effect is they depend of age and/or delay time? (3) Does delay time
increased by senile miosis, common in elderly adults, which influence color memory? If so, does the influence depend on
accentuates the color change produced by a yellowed lens, gender and/or age?
producing a decrease in retinal illumination, which is
greater than that would be anticipated if only the relative
MATERIALS AND METHODS
pupillary area is considered.
The effects of senile miosis and lenticular absorption and The experimental set-up consisted of a Macbeth cabinet
scatter combine to elevate detection thresholds and impair containing the test instrument, a grey cardboard circle panel
color recognition. These effects are particularly significant on which the comparison chips were mounted.
when ambient luminance passes to lower photopic or upper We selected five samples to be remembered, reference
mesopic levels. Weale22 indicates that due to senile miosis tests from a Munsell Book of Color (glossy). The reference
and crystalline yellowing, approximately only a third of the tests used, all with the same Munsell value and the corre-
light available to the retina of a 20-year-old observer is sponding Munsell notation, were violet 10PB 5/4, bluish
present in the retina of a 60-year-old. green 10G 5/8, pink 10RP 5/6, yellow green 5GY 5/6, and

Volume 31, Number 6, December 2006 459

 ⫽ –2.5, b* ⫽ ⫺2.4, under illuminant D65 simulator) of 30 cm
in diameter with a concealed identification. The observer was
deprived of any spatial reference that would lead to a special
predisposition to select a concrete color chip, because the chips
were randomly distributed, and all the observers were naı̈ve
regarding the experimental set up.
The observation took place inside the Macbeth cabinet
under an illuminant D65 simulator. The luminance at the
bottom of the cabinet was approximately 300 cd/m2, the
CIE geometry of illumination and viewing for reflection
0°/45°, and the observation distance of about 40 cm.
We started the study with a population of 108 older adult
observers, men and women, ranging in age from 64 to 82
years. Observers were not experienced in psychophysics
and were naı̈ve to the hypotheses of the experiment. These
subjects lived independently in the community and were
recruited from a Social Center for Retired People.
All the subjects underwent thorough medical, ocular, and
color vision examinations by a physician, one of the authors
CMP-B, and an optometrist, professor at the School of
Optometry in the University of Valencia.
Subjects were included in the study only if they had no
history of diabetes or other systemic diseases, and they were
not regularly taking medications that are known to have a
significant effect on color vision. In addition, all subjects
reported a negative history of ocular disease or surgery and
were found to have good ocular health. Normal color vision,
as assessed by a Ishihara pseudoisochromatic plates test, a
Farnsworth D-15 color vision test and a Standard pseudo-
FIG. 1. Chromaticity coordinates in CIE(1976) a*b* plane isochromatic plates part 2 (SSP2) test, was also a prerequi-
(up) and CIE(1976) L*C* plane (down) of the bluish green site for inclusion.
reference test (F) and the related comparison chips (⫻). As a result of these examinations our initial population
was reduced to 50 older adult observers, 25 men (ranging in
age from 64 to 80 years, mean age ⫽ 69.5 years) and 25
orange 5YR 5/8. To obtain the color coordinates L*, a*, b* women (ranging in age from 64 to 77 years, mean age ⫽
of these tests, we used a MiniScan MS/B 2000S. 69.1 years).
For the study of each reference test, we used 20 compar- Most of the observers had a middle class standard of
ison color chips, also chosen from a Munsell Book of Color living, and in terms of education, 88% had a college degree,
(glossy). One of the comparison color chips was equal to the and 12% graduated from high school or some form of
reference test, while the 19 remaining ones were chosen as
further education.
follows: First, we chose the Munsell chip equal to our
We also used a control group of 50 young adults in good
reference test and eight neighboring chips on the same page.
general health, 25 men and 25 women, selected from uni-
From earlier and later pages, we chose five or six chips,
versity students, in the 20 to 27 age range (mean age 23.5
respectively, with the same Munsell chroma and Munsell
years), and all were given the Ishihara test to check that they
value as the reference test and those with a higher and lower
Munsell chroma and Munsell value (for example and for the were normal from a color vision viewpoint.
yellow green reference test, the 20 comparison color chips Before beginning our experiment, we explained to the
used were: 5GY 5/4, 5GY 5/6, 5GY 5/8, 5GY 6/4, 5GY 6/6, observers the aim of the study, the procedure and that, to
5GY 6/8, 5GY 4/4, 5GY 4/6, 5GY 4/8, 2.5GY 5/4, 2.5GY complete the study, the observer had to be available for a
5/6, 2.5GY 5/8, 2.5GY 6/6, 2.5GY 4/6, 7.5GY 5/4, 7.5GY second session at 24 hours after the first session.
5/6, 7.5GY 5/8, 7.5GY 6/4, 7.5GY 6/6, and 7.5GY 4/6). The observers adapted to laboratory illumination, at
Figure 1 shows, for example, for bluish green, the distribu- about 90 cd/m2 for 5 min, before each session, and then sat
tion in CIE(1976) a*b* and CIE(1976) L*C* plans of the 20 in front of the Macbeth cabinet. In each session, for suc-
comparison chips and the corresponding reference test. cessive and simultaneous color matching, the observer took
They made a total of 20 comparison color chips, 17 mm by as much time as they wished to make their match. In the
20 mm, for each reference test color. The comparison chips same session, each observer performed the following tasks
were arranged in a ring on a grey circle panel (L* ⫽ 75.2, a* with the five color reference tests.

460 COLOR research and application

TABLE I. Variation of mean CIELAB color difference (⌬E*ab) between the mean selected chip by simultaneous
matching/successive matching and the test of reference, for each color.
Simultaneous By memory
0s ⫾SD 15 s ⫾SD 15 min ⫾SD 24 h ⫾SD

Men
Violet 0.5 1.7 9.1 3.3 9.2 2.7 9.6 3.7
Bluish green 5.3 3.9 10.8 3.1 11.3 2.0 11.0 3.5
Pink 1.2 2.8 9.2 3.9 10.1 4.2 9.9 3.9
Yellow green 5.0 6.2 10.0 4.7 11.7 3.5 10.8 3.5
Orange 3.4 5.4 12.9 3.7 12.7 4.5 12.7 4.4
Women
Violet 0.5 2.1 8.8 4.1 7.3 3.6 7.9 4.0
Bluish green 5.0 4.9 10.7 2.7 10.1 4.0 8.8 4.3
Pink 1.3 1.4 7.0 4.3 7.0 4.8 8.0 5.1
Yellow green 2.0 4.7 11.7 3.4 11.8 4.2 11.9 4.9
Orange 2.7 4.5 13.2 3.2 11.5 4.1 11.5 4.1

Mean values and standard deviations for each group of OLDER ADULT observers, men and women, and delay times.

Successive or Memory Color Matching Measurements the panel cover for 3 min. The observer always performed
(Delay Times: 15 s, 15 min, and 24 h) the same operation with the reference tests. As always,
between reference tests the observer looked at the grey
A covered grey panel was placed inside the cabinet and the
cover for 2 min.
observer was given one of the five reference tests and asked
Mechanics in each of the two sessions were as follows: in
to look at it for 5 s, paying close attention in order to
the first practice session, memory color matching session,
memorize it. In our investigation, only a single memory
the observer looked at the reference test for 5 s, and then
color test was used in this matching; a different mechanism
selected, at 15 s and 15 min, for each of the five reference
of color memory might work when several colors have to be
tests, the comparison chip that most resembled the memo-
stored and recalled.33
rized reference test. In the second practice session, memory
After that time, we took out the reference test and waited
color matching and simultaneous color matching session,
for 15 s while the observer looked at the grey cover of the
respectively, the subject, for each of the five reference tests,
panel and we then removed it, asking the observer to rec-
chose at 24 h the comparison chip most equal to the mem-
ognize the comparison test equal to the memorized one. We
orized reference test, and then, selected the comparison chip
noted the selected chip and how long the matching took for
that most resembled the reference test side-by-side.
each original color test.
The observer was asked to recognize the reference test
among the comparison chips after 15 min and 24 h, and,
again, the selected chip was then noted. In all the cases (15 RESULTS
s, 15 min, and 24 h), the observer was not allowed to see the
reference test. After 3 min adaptation to the panel cover, the We address the three research questions of interest, de-
observer then performed the simultaneous color matching scribed in the Introduction, by testing for the equality of
session. mean values of the observed color memory differences in
the four groups obtained crossing age and gender (young
adult men, young adult women, older adult men, and older
Simultaneous or Perceptual Color Matching
adult women).
Measurements (Delay Time: 0 s)
We used Repeated Measures Analysis of Variance in
Again the covered panel was put inside the cabinet and at order to model the correlation between the repeated mea-
the same time, the observer was given the reference test. sures (at the different time delays). The associated spheric-
The cover was removed and he was asked to select from ity assumption was verified. The results of the analysis are
among the comparison chips the one which most resembled discussed in the next section. For those cases where a
the reference test. We advised the observer to put the statistically significant difference was observed, we quanti-
reference test close to the comparison chips in order to fied the magnitude of the difference using the 95% confi-
perform an accurate side-by-side color matching. dence interval, IC95.
This procedure was also used with the five reference tests. Tables I–IV show the experimental results for mean
Between consecutive reference tests, the observer looked at values of ⌬E*ab, ⌬L*, ⌬C*ab, and ⌬H*ab, respectively,
the panel cover for 2 min. We noted the selected chip and between the selected sample by simultaneous/by memory
how long it took for each of the colors. and the reference test for each color and the older adult
Previously and as in the previous sessions, the observer group, men and women.
adapted to laboratory illumination for 5 min and, after, to Similarly, Tables V–VIII show the mean CIELAB dif-

Volume 31, Number 6, December 2006 461

TABLE II. Variation of mean CIELAB lightness difference (⌬L*) between the mean selected chip by simultaneous
matching/successive matching and the test of reference, for each color.
Simultaneous By memory
0s ⫾SD 15 s ⫾SD 15 min ⫾SD 24 h ⫾SD

Men
Violet 0.1 0.2 1.3 6.1 3.5 5.5 3.2 7.1
Bluish green 0.3 0.5 0.0 7.4 0.4 5.7 ⫺1.0 7.4
Pink 0.7 2.4 2.8 4.6 1.0 5.4 2.8 6.1
Yellow green ⫺0.5 1.9 ⫺0.6 4.7 0.3 6.1 0.3 4.6
Orange 0.0 0.1 1.0 5.2 ⫺0.2 4.0 ⫺0.1 5.5
Women
Violet ⫺0.4 2.1 ⫺1.2 5.5 0.5 5.4 ⫺0.7 5.8
Bluish green ⫺0.3 2.2 ⫺2.0 4.5 ⫺0.5 6.4 ⫺2.6 4.6
Pink 0.3 0.4 0.7 4.7 0.4 5.9 0.4 6.4
Yellow green ⫺0.1 0.1 ⫺0.7 3.5 ⫺0.6 4.7 ⫺1.0 5.7
Orange ⫺0.3 1.8 0.3 3.6 0.3 4.6 ⫺0.1 5.8

Mean values and standard deviations for each group of OLDER ADULT observers, men and women, and delay times.

ferences corresponding to the young adult group, men and pends on the studied color. The principal differences are
women. established in the bluish green (P ⫽ 0.001 and IC95 ⫽ [2.3,
5.4]) orange (P ⫽ 0.001 and IC95 ⫽ [0.9, 4.9]), and yellow
green (P ⫽ 0.008 and IC95 ⫽ [1.0, 6.8]).
ANALYSIS
For bluish green, older men match the original test more
We want to point out that as the CIELAB space is not chromatically than do young men (P ⫽ 0.001 and IC95 ⫽
perfectly uniform34,35 neither the different reference tests [1.2, 6.3]), and older women choose color chips more blue
nor the corresponding color dimensions can be strictly com- than that of the original (P ⫽ 0.015 and IC95 ⫽ [0.2, 2]). For
pared. yellow green, older men select the original chroma more
poorly than do young men (P ⫽ 0.01 and IC95 ⫽ [1.0, 6.5]).
For orange, significant differences are not obtained in the
Which Colors Show the Most Significant Old and
perceptual attributes. Lightness of reference tests does not
Young Adults Shift?
depend on age.
As pointed out in the Introduction, it should be noted that in Successive or Memory Color Matching: In general, inde-
the perceptual literature we have not found any research pendent of the analyzed color, young adults remember color
which explicitly studies color memory in elderly adults. of reference tests better than elderly adults do (P ⫽ 0.007
From Tables I–VIII, we concluded for each color refer- and IC95 ⫽ [0.2, 1.2]), although the magnitude of these
ence test the following. differences depend on the gender and delay time.
Simultaneous Color Matching: We found that, indepen- Violet: With short delay times, old women remember the
dent of gender, the elderly adults match color of reference original test more chromatically than do young women (P ⫽
tests more poorly than do young adults (P ⫽ 0.0 and IC95 ⫽ 0.014 and IC95 ⫽ [0.7,4.7]). Nevertheless, this behavior is
[0.9, 2.1]), although the magnitude of the difference de- not observed in the rest of memory times. At 15 min, older

TABLE III. Variation of mean CIELAB chroma difference (⌬C*ab) between the mean selected chip by simulta-
neous matching/successive matching and the test of reference, for each color.
Simultaneous By memory
0s ⫾SD 15 s ⫾SD 15 min ⫾SD 24 h ⫾SD

Men
Violet 0.3 1.6 5.8 4.1 4.4 5.0 5.3 4.1
Bluish green 3.4 5.0 3.8 6.6 2.0 9.3 4.7 6.7
Pink 0.1 0.3 6.4 4.7 5.8 7.1 4.2 6.6
Yellow green 4.7 6.0 1.6 7.8 3.1 8.4 2.2 8.0
Orange 1.3 5.3 5.8 10.1 7.4 9.4 5.9 9.9
Women
Violet 0.1 0.3 6.0 4.6 3.3 4.7 3.0 5.7
Bluish green 0.5 6.3 4.4 8.1 ⫺0.1 8.4 ⫺0.7 7.6
Pink 0.1 0.2 4.2 4.6 1.8 5.0 1.9 6.5
Yellow green 2.0 4.6 ⫺1.6 9.6 ⫺1.1 9.9 ⫺0.8 10.0
Orange ⫺0.3 4.4 4.9 11.1 1.2 9.9 0.3 9.5

Mean values and standard deviations for each group of OLDER ADULT observers, men and women, and delay times.

462 COLOR research and application

TABLE IV. Variation of mean CIELAB hue difference (⌬H*ab) between the mean selected chip by simultaneous
matching/successive matching and the test of reference, for each color.
Simultaneous By memory
0s ⫾SD 15 s ⫾SD 15 min ⫾SD 24 h ⫾SD

Men
Violet ⫺0.1 0.6 0.6 2.4 0.7 2.5 0.6 1.9
Bluish green 0.3 2.7 0.8 4.2 0.7 3.7 1.1 3.5
Pink ⫺0.4 1.7 0.7 3.2 1.6 2.8 ⫺0.2 3.0
Yellow green ⫺0.6 0.8 4.3 4.6 3.2 5.2 3.6 5.6
Orange 1.2 3.1 0.0 4.7 0.2 5.0 ⫺0.3 4.7
Women
Violet 0.1 0.5 1.1 2.3 1.4 2.1 0.5 2.3
Bluish green 1.5 1.8 0.8 3.7 0.2 3.4 0.4 3.6
Pink ⫺1.2 1.4 ⫺1.4 2.5 ⫺0.3 3.1 0.8 2.3
Yellow green ⫺0.3 0.6 4.3 5.3 3.3 5.5 2.9 5.4
Orange 0.7 2.1 ⫺0.9 5.3 ⫺3.3 4.8 ⫺2.3 5.0

Mean values and standard deviations for each group of OLDER ADULT observers, men and women, and delay times.

women remember the original test tending to red (P ⫽ 0.014 Chroma is the component of color difference with the
and IC95 ⫽ [0.3, 2.4]), whereas young women did not show highest contribution to the total color difference. Elderly
a definite tendency. In lightness, older men tend to remem- adults remember the original test more chromatically than
ber the original test as lighter than do young men throughout do young adults at 15 s (P ⫽ 0.02 and IC95 ⫽ [1.4, 5.8]) and
the three memory delay times, but these differences are not at 15 min (P ⫽ 0.012 and IC95 ⫽ [0.7, 5.7]). Nevertheless,
significant. this difference is not observed at 24 h. We found no statis-
Bluish Green: At 15 s, young adults, independent of tical differences for hue. We found that young women tend
gender, remember the original test as more blue than do to remember the original test as lighter than do older
elderly adults (P ⫽ 0.041 and IC95 ⫽ [0.2, 2.1]). Similar women, this difference being significant at 15 s (P ⫽ 0.02
behavior is observed for the rest of delay times although and IC95 ⫽ [0.6, 6.0]).
not in a significant way. In addition, we found that older Yellow Green: Older women remember the original test
women tend to remember the original test as darker than more poorly than do young women at 15 s (P ⫽ 0.023 and
do young women at 15 s and 24 h (P ⫽ 0.02 and IC95 ⫽ IC95 ⫽ [0.4, 5.4]); this behavior is also observed at 24 h but
[0.6, 5.8], and P ⫽ 0.013, and IC95 ⫽ [0.8,6.3], respec- not in a significant way. In chroma, independent of the three
tively); these differences are not observed so clearly at 15 delay times, young women remember the original test more
min. Finally, we observe at 15 s and 15 min that older chromatically, and older women less chromatically (P ⫽
men remember the original chroma qualitatively poorer 0.007 and IC95 ⫽ [1.0, 6.3]). In addition, we found that
than do the young men. independent of gender and delay time, although older adults
Pink: We found that elderly adults remember original remember the original hue test tending to green, young
color more poorly than do young adults throughout the three adults remember it tending to yellow (P ⫽ 0.0 and IC95 ⫽
delay times considered, although the difference is only [3.6, 6.7]). In lightness, young men tend to remember the
significant at 15 min (P ⫽ 0.04 and IC95 ⫽ [0.2, 5.0]). original test as darker than do older adults, who did not

TABLE V. Variation of mean CIELAB color difference (⌬E*ab) between the mean selected chip by simultaneous
matching/successive matching and the test of reference, for each color.
Simultaneous By memory
0s ⫾SD 15 s ⫾SD 15 min ⫾SD 24 h ⫾SD

Men
Violet 0.0 0.0 7.7 4.7 8.2 4.0 8.7 4.6
Bluish green 1.1 2.9 10.7 2.5 11.4 1.4 10.6 3.0
Pink 0.8 1.3 7.8 4.5 7.6 4.2 8.8 4.5
Yellow green 1.0 3.5 11.6 4.4 11.2 4.3 10.6 4.7
Orange 0.0 0.0 12.7 5.2 10.9 5.5 10.0 5.3
Women
Violet 0.0 0.0 7.4 4.6 6.9 4.4 7.1 4.9
Bluish green 1.5 3.5 10.2 3.7 10.3 1.8 10.4 2.6
Pink 0.6 1.2 7.7 5.0 6.7 4.8 7.3 5.2
Yellow green 0.5 2.5 8.8 5.1 13.2 4.1 9.3 6.1
Orange 0.3 1.5 9.8 5.3 10.6 6.6 12.7 5.9

Mean values and standard deviations for each group of YOUNG ADULT observers, men and women, and delay times.

Volume 31, Number 6, December 2006 463

TABLE VI. Variation of mean CIELAB lightness difference (⌬L*) between the mean selected chip by simultaneous
matching/successive matching and the test of reference, for each color.
Simultaneous By memory
0s ⫾SD 15 s ⫾SD 15 min ⫾SD 24 h ⫾SD

Men
Violet 0.0 0.0 ⫺0.4 5.6 0.4 6.3 ⫺0.1 7.3
Bluish green ⫺0.1 0.1 1.1 3.5 ⫺0.4 5.7 ⫺0.9 5.2
Pink 0.2 0.3 1.5 4.3 1.2 5.8 0.0 5.5
Yellow green 0.0 0.1 ⫺1.7 4.6 ⫺2.0 3.9 ⫺1.2 4.2
Orange 0.0 0.0 ⫺0.3 2.9 0.7 2.8 0.3 3.4
Women
Violet 0.0 0.0 0.6 6.9 ⫺0.1 6.4 2.0 6.7
Bluish green 0.0 0.3 1.2 4.6 0.6 3.7 0.9 5.1
Pink 0.2 0.3 4.0 4.9 1.9 3.7 3.3 5.6
Yellow green 0.0 0.1 1.0 3.4 ⫺0.1 5.0 ⫺0.5 3.4
Orange 0.0 0.0 0.6 2.8 1.0 3.3 0.6 2.8

Mean values and standard deviations for each group of YOUNG ADULT observers, men and women, and delay times.

show a definite behavior. Nevertheless, these differences are CIELAB total color difference or in the different compo-
not significant. nents of color difference.
Orange: For this color we only found noticeable differ- Successive or Memory Color Matching: In general, men
ences for women. Older women remember the color of remember more poorly than do women (P ⫽ 0.004 and IC95
reference test more poorly than do young women at 15 s ⫽ [0.26, 1.31]), independent of age, color, and delay time,
(P ⫽ 0.009 and IC95 ⫽ [0.9, 5.9]). For the rest of delay since the different interactions are not statistically signifi-
times there were no significant differences. At 15 min and cant.
24 h, young women remember the original chroma more Violet: Men remember the color of the original test more
chromatically (P ⫽ 0.02 and IC95 ⫽ [1.0,11.5], and P ⫽ poorly than do women, independent of age and delay time
0.04 and IC95 ⫽ [2.7, 13.8], respectively), whereas older (P ⫽ 0.05 and IC95 ⫽ [0.1, 2.4]). This difference cannot be
women did not show a definite tendency. With respect to explained from perceptual attributes because they only con-
hue and lightness, we did not obtain a statistically signifi- tribute to lightness and chroma, although not for all delay
cant difference. times. We found that young men remember the original test
more chromatically than do young women at 15 s (P ⫽ 0.04
and IC95 ⫽ [0.1, 4.9]), and older men remember more
Are There any Gender-Related Color Memory
poorly than do older women at 15 min and 24 h, but these
Difference?
differences are not significant. In addition, although older
From Tables I–VIII, we concluded for each color reference men remember the original test as lighter, independent of
test the following. delay times, older women tend to remember as darker or not
Simultaneous Color Matching: We did not find any sig- show a definite tendency; the difference was only statisti-
nificant difference depending on gender either in the mean cally significant at 24 h (P ⫽ 0.04 and IC95 ⫽ [0.1, 7.5]).

TABLE VII. Variation of mean CIELAB chroma difference (⌬C*ab) between the mean selected chip by simulta-
neous matching/successive matching and the test of reference, for each color.
Simultaneous By memory
0s ⫾SD 15 s ⫾SD 15 min ⫾SD 24 h ⫾SD

Men
Violet 0.0 0.0 5.4 4.2 3.9 5.2 3.0 5.8
Bluish green ⫺0.7 2.4 1.8 9.2 0.8 9.4 0.6 9.0
Pink 0.1 0.1 3.0 6.9 2.5 5.2 4.2 6.7
Yellow green 1.0 3.4 3.2 10.0 3.7 9.3 3.0 9.0
Orange 0.0 0.0 9.5 8.3 6.1 9.0 3.7 8.7
Women
Violet 0.0 0.0 2.9 4.2 2.4 4.5 3.0 4.0
Bluish green 0.2 3.6 3.8 7.9 4.1 8.1 1.3 8.8
Pink 0.1 0.1 0.4 6.2 ⫺1.4 6.8 0.9 5.8
Yellow green 0.5 2.5 5.7 6.7 5.2 10.3 4.7 8.5
Orange 0.2 0.9 5.9 7.2 7.4 8.6 8.6 9.9

Mean values and standard deviations for each group of YOUNG ADULT observers, men and women, and delay times.

464 COLOR research and application

TABLE VIII. Variation of mean CIELAB hue difference (⌬H*ab) between the mean selected chip by simultaneous
matching/successive matching and the test of reference, for each color.
Simultaneous By memory
0s ⫾SD 15 s ⫾SD 15 min ⫾SD 24 h ⫾SD

Men
Violet 0.0 0.0 1.5 1.8 0.8 1.7 1.1 1.5
Bluish green 0.7 1.7 3.2 3.5 1.1. 3.7 1.7 3.7
Pink ⫺0.8 1.3 ⫺0.5 2.3 0.4 2.9 0.4 2.8
Yellow green ⫺0.1 0.4 0.5 4.9 ⫺0.7 5.2 ⫺2.0 5.1
Orange 0.0 0.0 ⫺0.5 4.9 ⫺0.4 5.1 1.7 5.2
Women
Violet 0.0 0.0 0.8 1.5 0.1 1.5 0.7 1.6
Bluish green 0.4 1.2 2.8 3.5 1.2 3.8 1.4 3.3
Pink ⫺0.5 1.1 ⫺0.2 2.7 ⫺0.4 2.1 ⫺0.1 2.5
Yellow green ⫺0.1 0.3 ⫺1.7 3.6 ⫺3.1 5.3 ⫺2.6 3.9
Orange 0.2 1.2 ⫺1.6 5.4 ⫺0.4 4.2 0.3 4.5

Mean values and standard deviations for each group of YOUNG ADULT observers, men and women, and delay times.

Bluish Green: In general, men remember more poorly [0.7, 6.3]), this behavior is also observed for the rest of
than do women, independent of age and delay time (P ⫽ delay times.
0.019 and IC95 ⫽ [0.2, 1.6]). This behavior cannot be
explained from perceptual attributes because there is only a
significant difference in chroma for elderly adults and at Does Delay Time Influence Color Memory?
24 h, where older men remember more poorly than do older In general, we found that in the elderly adults, as with young
women (P ⫽ 0.01 and IC95 ⫽ [1.3, 9.4]). In lightness, older adults, with all reference tests the mean ⌬E*ab in simulta-
women remember darker than do older men, but these neous matching is lower than each of the three mean ⌬E*ab
differences are not statistically significant. in memory matching, corresponding to the three different
Pink: Men remember the original test more poorly than do delay time intervals used (P ⬍ 0.05). In addition, we find no
women (P ⫽ 0.02 and IC95 ⫽ [0.2, 2.9]), independent of statistically significant differences among these three mean
gender and delay time. The greatest contribution to the ⌬E*ab.
CIELAB total color difference is related to chroma. Men
remember the chroma original more poorly than do women
(P ⫽ 0.002 and IC95 ⫽ [0.8, 3.4]), especially at 15 min and DISCUSSION
24 h. Hue also establishes differences in the elderly adults: at
How Can We Justify the Differences Found Between
15 s, women remember the original test tending to red and men
Older and Young Adults?
did not show a definite tendency (P ⫽ 0.013 and IC95 ⫽ [0.5,
3.7]), and at 15 min, men remember it tending to red and We have not found a correlation between behaviors corre-
women did not show any tendency (P ⫽ 0.03 and IC95 ⫽ [0.3, sponding to both color matching methods, for any of our
3.6]). In lightness, young women remember at 24 h the original reference tests. In yellow green, for example, we found that
test as lighter than do men (P ⫽ 0.041 and IC95 ⫽ [0.1, 6.4]); while in simultaneous matching there were noticeable
older men tend to remember the original test as lighter than do chroma differences in men, in successive matching there
older women, but this difference is not significant. were differences for hue and chroma in women.
Yellow Green: For this color, we observe a few gender- In simultaneous matching, we found that the elderly
related color differences. In young adults, men remember adults show more difficulties to match the chroma and the
the reference test more poorly than do women at 15 s (P ⫽ hue of reference tests, which confirms partly the studies that
0.044 and IC95 ⫽ [0.01, 5.5]), but this tendency is not indicate that the elderly adults have more problems in the
observed for the rest of the times. This difference is basi- discrimination of saturation and hue.13–15,27
cally because that at 15 s young men remember the original On the other hand, justifying the results found by color
test as darker (with P ⫽ 0.02 and IC95 ⫽ [0.4, 5.1]); this memory matching is more complicated because we have not
tendency also observed for the rest of delay times. In hue, found any studies in the perceptual literature on elderly
young women remember the sample tending to yellow over adult-related color memory differences. We think that the
three delay times, whereas in men this occurred only at 24 h, color memory matching differences we found in old–young
but these differences are not significant. adults may, at least in part, be explained by chromatic
Orange: In elderly adults, men remember original discrimination differences related to age and that we have
chroma more poorly than do women at 15 min and at 24 h already mentioned in the Introduction to the present study.
(P ⫽ 0.028 and IC95 ⫽ [0.7, 11.7], and P ⫽ 0.048 and IC95 In general, as the radiance levels of the chips are held
⫽ [0.4, 11.1], respectively. In hue, at 15 min, old women constant for all observers, the elderly adults experience a
remember this sample tending to red (P ⫽ 0.015 and IC95 ⫽ greater reduction in retinal illumination15,21,22 due to senile

Volume 31, Number 6, December 2006 465

miosis and lenticular yellowing.12,21,22 In chroma, differ- difficulty matching chroma and hue. In general, lightness
ences can be due to the greater difficulties experienced by and hue of reference tests are quite well matched inde-
elderly adults in discriminating saturation,27,28 and also it pendent of age.
should also be taken into account that old–young adult 3. By memory, independent of the color analyzed, young
differences are greater with desaturated test stimuli,23–26 as adults remember better than do older adults (P ⫽ 0.007),
occurred for violet (pex ⫽ 0.19), bluish green (pex ⫽ 0.25), although it depends on gender and delay time. Chroma
and pink (pex ⫽ 0.20). and hue are the perceptual attributes that better explain
Differences found in the hue memory could be justified the color differences obtained.
by the loss in the capacity to discriminate hue with age.13–15 4. Our results show that in simultaneous color matching the
In addition, as expected from lenticular senescence, the loss gender of elderly adults did not show significant differ-
in sensitivity for each mechanism was greater at shorter ences.
wavelengths,17 which might explain our hue memory results 5. In general, men remember the color of reference test
for bluish green and the violet. more poorly than do women (P ⫽ 0.0), independent of
For violet and bluish green, it should be remembered that age, color, and delay time, especially for bluish green,
the elderly experience a deterioration in color discrimina- violet, and pink. The ⌬L* and ⌬H*ab contributions to
tion ability, which affects violet/blue– green discrimina- gender-related CIELAB total color differences are small.
tion12 in particular. Also, Verriest et al.20 reported system- On the contrary, chroma is the perceptual color attribute
atic changes in responses which were pronounced in the that better explains gender differences (P ⫽ 0.0), al-
blue–violet range and this is attributed to a progressively though that always depends on the reference test consid-
higher lenticular absorbance. ered.

How Can We Justify the Differences Found Due to ACKNOWLEDGMENTS

Gender?
We would like to thank the City Council of Burjassot
From our results, we cannot establish a correlation between (Valencia), and especially Miss Ana Sevilla, for the facili-
simultaneous and successive matching, since in simulta- ties provided for this study and their assistance in recruiting
neous matching we have not found differences depending subjects from the ‘Villa del Pilar’ Social Center for Retired
on gender, whereas by memory matching, men remember People.
more poorly than do women for pink and violet. We also would like to thank Dr. J.C. Soriano, for his
Our results confirm and extend those we obtained in a helpful comments, F. Olmos, optometrist, for the eye ex-
previous work,7 where we found that, in general, young men amination and refraction of the elderly group, and all those
adults remembered more poorly than did young women subjects who acted as observers, without them our research
adults. would not be possible.
We want to indicate the possibility that color memory
differences in the present article with both gender groups, as 1. Hamwi V, Landis C. Memory for color. J Psychol 1956;39:183–194.
suggested in the literature with simultaneous color match- 2. Burnham RW, Clark JR. A test of hue memory. J Appl Psychol
ing, may explain the different behavior we found: (a) for 1955;39:164 –172.
pink and orange, because women show greater responsive- 3. Newhall SM, Burnham RW, Clark JR. Comparison of successive with
simultaneous color matching. J Opt Soc Am 1957;47:43–56.
ness to the long wavelength segment of the spectrum in a 4. Bartleson CJ. Memory colors of familiar objects. J Opt Soc Am
color discrimination task;36 (b) for violet and bluish green, 1960;50:73–77.
because the average level of macular pigment (which selec- 5. Siple P, Springer RM. Memory and preference for the colors of
tively filters short-wavelength light) is significantly lower in objects. Percept Psychophys 1983;34:363–370.
women.31 6. Jin EW, Shevell SK. Color memory and color constancy. J Opt Soc
Am A 1996;13:1981–1991.
7. Pérez-Carpinell J, Baldovı́ R, de Fez MD, Castro J. Color memory
matching: Time effect and other factors. Color Res Appl 1998;23:
CONCLUSIONS 234 –247.
8. Pérez-Carpinell J, de Fez MD, Baldovı́ R, Soriano JC. Familiar objects
1. In the elderly adults we found that the mean ⌬E*ab by and memory color. Color Res Appl 1998;23:416 – 427.
memory color matching corresponding to each delay 9. Amano K, Uchikawa K, Kuriki I. Characteristics of color memory for
natural scenes. J Opt Soc Am A 2002;19:1501–1514.
time is higher than that of the mean ⌬E*ab in simulta-
10. Pérez-Carpinell J, Camps VJ, Diaz JA. Color memory in protanomals
neous color matching. If we consider the average of the and deuteranomals: Matching time effect. Color Res Appl 2003;28:
three memory delay times, we obtain a significant dif- 151–154.
ference between simultaneous and memory matching 11. Bodrogi P, Tarczali T. Investigation of colour memory. In: MacDonald
(P ⬍ 0.001). LW, Luo MR, editors. Colour Image Science. Exploiting Digital
Media. Chichester, England: Wiley; 2002. p 23– 46.
2. With simultaneous matching, we found that while young
12. Lakowski R. Age and colour vision. Adv Sci 1958;15:231–236.
adults matched the color of all original tests well, the 13. Verriest G. Further studies on acquired deficiency of color discrimi-
elderly adults matched poorly both greens (P ⬍ 0.05) nation. J Opt Soc Am 1963;53:185–195.
and orange, probably because older men have more 14. Smith VC, Pokorny J, Pass AS. Color-axis determination on the

466 COLOR research and application

 Farnsworth–Munsell 100-hue test. Am J Ophthalmol 1985;100:176 – sensitivities based on saturation discrimination. Color Res Appl 1991;
182. 16:289 –296.
15. Knoblauch K, Saunders F, Kusuda M, Hynes R, Podgor M, Higgens 27. Cooper BA, Ward M, Gowland CA, McIntosh JM. The use of the
KE, deMosasterio M. Age and illuminance effects in the Farnsworth– lanthony new color test in determining the effects of aging on color
Munsell 100-hue test. Appl Opt 1987;26:1441–1448. vision. J Gerontol 1991;46:320 –324.
16. Warburton FL. Variations in normal colour vision in relation to prac- 28. Kraft JM, Werner JS. Aging and the saturation of colors. 1. Colori-
tical colour matching. Proc Phys Soc B 1954;67:477– 484. metric purity discrimination. J Opt Soc Am A 1999;16:223–230.
17. Werner JS, Steele VG. Sensitivity of human foveal color mechanisms 29. Sharpe LT, Stockman A, Jaegle H, Nathans J. Opsin genes, cone
throughout the life span. J Opt Soc Am A 1988;5:2122–2130. photopigments, color vision and color blindness. In: Gegenfurtner K,
18. Weale RA. Age and the transmittance of the human crystalline lens. Sharpe LT, editors. Color Vision: From Genes to Perception. Cam-
J Physiol 1988;395:577–587. bridge: Cambridge University Press; 1999. p 3–50.
30. Jameson KA, Highnote SM, Wassermann LM. Richer color experi-
19. Boettner EA, Wolter JR. Transmission of the ocular media. Invest
ence in observers with multiple photopigment opsin genes. Psychon
Ophthalmol 1962;1:776 –783.
Bull Rev 2001;8:244 –261.
20. Verriest G, Vandevyvere R, Vanderdonck R. Nouvelles recherches se
31. Hammond BR, Jr., Curran-Celentano J, Judd S, Judd K, Krinsky NI,
rapportant à l’influence du gendere et de l’age sur discrimination
Wooten BR, Snodderly DM. Sex differences in macular pigment
chromatique, ainsi qu’à la signification du test 100 hue de Fansworth
optical density: Relation to serum and dietary patterns. Vision Res
Munsell. Rev Opt Théor Instrum 1962;41:499 –509.
1997;36:2001–2012.
21. Weale R. Retinal illumination and age. Trans Illuminating Eng Soc 32. MacGuinness D, Lewis I. Sex differences in visual persistence: Experi-
1961;26:95–100. ments in the Ganzfield and afterimages. Perception 1976;5:295–301.
22. Weale R. The Aging Eye. London: Lewis HK; 1963. 33. Uchikawa K, Shinoda H. Influence of basic color categories on color
23. Pinckers A. Color vision and age. Ophthalmologica 1980;181:23–30. memory discrimination. Color Res Appl 1996;21:430 – 439.
24. Bowman KJ, Collins MJ, Henry CJ. The effect of age on performance 34. Robertson AR. The CIE 1976 color-difference formulae. Color Res
on the panel D-15 and the desaturated D-15: A quantitative evaluation. Appl 1977;2:7–11.
In: Verriest G, editor. Colour Vision Deficiencies. The Hague, The 35. Pointer MR. A comparison of the CIE 1976 colors. Color Res Appl
Netherlands: Junk; 1984. 1981;6:108 –118.
25. Kaiser PK, Comerfeld JP, Bodinger DM. Saturation of spectral lights. 36. Eisner A, Fleming SA, Klein ML, Mauldin WM. Sensitivities in older
J Opt Soc Am 1976;66:818 – 826. eyes with good acuity: Cross-sectional norms. Invest Ophthalmol
26. Kimura E. Effects of luminance level on the saturation function- Vision Sci 1987;28:1824 –1831.

Volume 31, Number 6, December 2006 467

View publication stats