Basic color terms use by aged observers: lens aging and perceptual compensation.
BCTs are terms that are consistently used in a specific language to identify color categories (Berlin & Kay, 1969; Kaiser & Boynton, 1996; Schirillo, 2001). That is, a BCT is a monolexemic word used consistently by most of the speakers of a language to name a set of colors sharing some perceptual characteristics. Several works (Boynton & Olson, 1987; Lin, Luo, MacDonald, & Tarrant, 2001a, 2001b; Sturges & Whitfield, 1995) showed that English has 11 BCTs. A recent work (Lillo, Moreira, Vitini, & Martin, 2007) showed that English and Spanish BCTs are colorimetrically similar. That is, each English BCT has a Spanish equivalent (both are used for naming similar parts of the color space of the Commission Internationale de l'Eclairage [CIE]). BCT use is based on two perceptual properties: similarity and differentiability. Colors named with the same BCT occupy proximal positions in CIE color space. There are no "gaps" between the stimuli belonging to a BCT (CIE proximity implies perceptual similarity, Hunt & Pointer, 2011). Stimuli named with different BCTs occupy different parts in CIE color space and are perceptually differentiable.
According to clinical nomenclature (Birch, 2001) the cones more responsive to long (L cones), medium (M cones), or short (S cones) wavelengths are called 'proto', 'deutera', or 'trita' cones. Tritan errors appear when an individual confuses stimuli only differing because of tritacone activity. Such an observer cannot use consistently different BCTs to name stimuli that he or she perceives without any differences.
A classic work (Verriest, 1963; see also Knoblauch et al., 1987) studied age changes in the capacity to differentiate hues (ages between 10 and 64 years). Results indicated that tritan errors increased from 20-24 years and become more frequent for older people. As will be shown, some ocular aging changes are related to this result.
Two age-related ocular changes (Schieber, 2006; Weale, 1986) produce tritan errors. Pupil size reduction (senile pupil meiosis; Winn, Whitaker, Elliott, & Phillips, 1994) is the first change. Such a reduction decreases the amount of light reaching the retina (retina illuminance). Knoblauch et al. (1987) showed that retinal illuminance reduction causes tritan errors even in young people.
Lens brunescence is another age-related ocular change. Several works (Mellerio, 1987; Norren & Vos, 1974; Pokorny, Smith, & Lutze, 1987; Weale, 1986, 1988) have shown that lens brunescence reduces lens transmittance, especially for short wavelengths (which are more relevant for tritacones). Studies in which spectral sensitivity function was measured using flicker photometry (Kraft & Werner, 1994; Sagawa & Takahashi, 2001) or similar procedures (Lillo & Moreira, 2005) found sensitivity reductions proportional to lens transmittance changes due to aging.
According to what is called Model A (filtering without compensation) in this paper, aged people's color perception (and BCT use) is similar to that of young people when responding to the distorted stimulation created by an aged lens. Because there are filters (Obama, Ikeda, Sagawa, & Shinoda, 2005) and mathematical algorithms (Pokorny et al., 1987) that precisely mimic lens brunescence effects, Model A predictions can be easily tested.
Lindsey and Brown (2002) used Model A to explain the existence of what is colloquially called "grue" ("gr" from green, "ue" from blue; Franklin, Clifford, Williamson, & Davies, 2005; Hardy, Frederick, Kay, & Werner, 2005; Kay, Berlin, Maffi, Merrifield, & Cook, 2010; Lindsey & Brown 2009; Regier, Kay, & Khetarpal, 2009; Roberson, Davidoff, Davies, & Shapiro, 2005). That is, the BCT used in many equatorial languages to name stimuli that in English (or Spanish) are blues [azules] or greens [verdes]. According to Lindsey and Brown (2002), grue appears because the acceleration in lens brunescence produced by the exposure to high UV radiation levels changes color perception for many people living in equatorial countries. Consequently, many speakers should have the same chromatic experience (green) when seeing some stimuli that normal people differentiate (green or blue). Supporting this interpretation, Lindsay and Brown found that, when young English speakers named stimuli that mimic lens brunescence effects, they reduced the use of blue and increased the use of green.
Lindsey and Brown's (2002) aging simulation was mathematically consistent with a valid model of lens brunescence (Pokorny et al., 1987). However, it is very questionable to assume that aged people's color vision must be similar to that of young people when responding to a retinal stimulation that mimics the one received by aged people. As Hardy et al. (2005) indicated, aged people have always lived with their optics and can develop chromatic compensation processes (Neitz, Carroll, Yamauchi, Neitz, & Williams, 2002; see also Delahunt, Webster, Ma, & Werner, 2004; Webster, Halen, Meyers, Winkler, & Werner, 2010). Developing such processes requires exposure to distorted stimulation for much longer time periods than the exposures used in Lindsey and Brown's experiment.
Hardy et al. (2005) replicated Lindsey and Brown's (2002) results with young observers (they also used transformed stimuli to mimic lens aging effects). But when real aged observers (normal, according to clinical standards) performed the same naming task, they were more similar to young people than to the simulated old observers (young people responding to transformed stimuli). From these results, Hardy et al. concluded the existence of a perceptual compensatory mechanism. Such a mechanism could also explain the following pattern of results. First, age does not change the wavelengths that produce yellow or blue unique colors (Schefrin & Werner, 1990). Second, age does not change the chromatic coordinates corresponding to achromatic stimuli (Werner & Schefrin, 1993). Third, although aging reduces purity discrimination (Kraft & Werner, 1999a), old people perceive saturation levels very similarly to young people (Kraft & Werner, 1999b). Forth, although cataract surgery dramatically changes chromatic coordinates for achromaticity (Delahunt et al., 2004), most of this change disappear after some months.
Model B (white normalization) provides a plausible explanation for the results described in the previous paragraph. Model B assumes that aged people compensate lens aging effects by using a von Kries-type compensation mechanism (Hunt & Pointer, 2011), similar to the one frequently proposed to explain color constancy (Foster et al., 1997). In agreement with Neitz et al. (2002) see also Wuerger, Xiao, Fu, and Karatzas, 2010, this mechanism operates at cortical level, adjusting the relative weight of each cone type to obtain responses of similar magnitude for the stimuli normally seen as white. In other words, to compensate the reduction in energy reaching tritacones produced by lens brunescence (and the relative increase in the energy reaching protocones), the compensation process increases tritacone response weighting (and reduces the weighting for the protocones).
In addition to Models A (filtering without compensation) and B (filtering with compensation based on white normalization), our research also considered a third model. According to Model C (tritan lines), aged observers must confuse the stimuli (and the BCTs used to name them) located on the same tritan line. Here we evaluate Model C for two reasons. First, as previously indicated, aging is associated with an increase in tritan errors. Second, aging reduces the energy acting on tritacones and, consequently, their functionality. But the main limitation of Model C is to assume, wrongly, that transmittance lens changes only affect tritacone responses (and not proto and deuteracone responses).
As in our previous research related to dichromats' BCT use (Lillo et al., 2012; Moreira et al., 2012), we used two search tasks similar to the ones described in Berlin and Kay's (1969) seminal study. Both studies used a stimulus set, from which participants had to point to: (a) all the stimuli that could be included in each BCT (mapping task), and (b) the best representative of each BCT (best representative task). The stimulus set (see Appendix 1, for a complete description) fulfilled the following two criteria: (a) it included a reduced number of stimuli (so simultaneous presentation would be possible) and (b) all the BCTs were accurately represented. Due to the nature of the stimulus set used, our study is the first to provide a global view on how aged people use BCTs (previous studies only provide partial information). Additionally, the current study complements our previous work (Lillo et al., 2012; Moreira et al., 2012) on dichromats' (protanopes and deuteranopes) use of BCTs (for a complete description of both studies see Moreira, 2010).
As previously stated, the aims of the study were descriptive and explicative. On the descriptive side, our main goal was to provide a global view on how aged people use BCTs and how such use differs from young people (some differences were expected considering the changes produced by the aging process in the visual system). On the explicative side, our main goal was to evaluate which one of the three outlined models was the most accurate for predicting use of BCTs and, consequently, which model contained the most relevant variables.
To get a global view on how aged people use BCTs, the results provided by the mapping task were used to perform several multidimensional scalings (MDS). Such analysis enables the comparison of global dimensions that describe BCT use in aged and young people. The results of the mapping task were also used to specify categorical naming errors (i.e., which BCTs were more frequently confused: green and blue? blue and black? etc.), and BCT use frequency (i.e., do aged people present a reduction for blue and an increase for green use as suggested by Model A?). Finally, the descriptive analysis also included comparison of aged people's performance on both search tasks. It was expected that, as with dichromats (Lillo et al., 2012; see also Lillo, Davies, Collado, Ponte, & Vitini, 2001), the best representative task would produce less errors because of the greater psychophysical specificity of this type of stimulus (a best representative occupies a specific localization in CIE color spaces, but each BCT can be used to name larger parts of such spaces).
To explain aged people's use of BCTs (one of the main goals of our research) we compared the empirical results provided by the mapping task with the predictions of the three models. More specifically, for each BCT, we evaluated whether or not correct and erroneous responses appeared for stimuli located on similar confusion tritan lines (Model C prediction). We also evaluated whether these responses were similar to those expected in young people when responding to stimuli that mimic the effects of lens brunescence, either considering (Model A) or not considering the effects derived from the performance of a von Kries compensation mechanism (Model B prediction).
Forty-five observers took part in the experiment. They belonged to the following groups: normal young (n = 15, age range 19-24, mean age = 21.4 years), normal aged (n = 15, age range 71-87, mean age = 81.33 years), and tritanomalous aged (n = 15, age range 74-98, mean age = 82.5 years). All participants were tested for color vision using Ishihara (1996) Pseudo-Isochromatic color plates, the City University Color Vision Test (CUCVT; Fletcher, 1980) and the Lanthony (1985) test. No participant produced protan or deuteran errors in any test. Diagnosis of tritanomaly was based on responses provided by CUCVT and Lanthony tests. People with two or more errors in at least one of these tests were considered tritanomalous. All participants collaborated voluntarily in the research and could stop their participation in any moment.
Materials and Stimuli
From previous results (Lillo et al., 2007), a set of 102 stimuli from the NCS color atlas (Scandinavian Color Institute, 1997) was chosen that included: (a) each BCT prototype, (b) "boundary-stimuli" between BCTs defined by previous use of combined terms, such as red-purple, and (c) stimuli halfway along the line in CIE L*u*v* between a prototype and each relevant boundary color. The set was presented on a grey background (S-5000 N, L* = 50), with a small gap between adjacent stimuli. Stimuli were viewed from a distance of 50 cm, each stimulus was 4[degrees] square, and the entire display was 64.42[degrees] x 33.08[degrees]. Illuminance was approximately 250 lux, and correlated color temperature 5500[degrees] K. Such illumination was obtained using incandescent lamps and color temperature corrective filters. Illuminance and color temperature were measured using a Minolta CL-200 luxo-colorimeter.
Color vision diagnosis was performed the day before the search tasks. Young observers' color vision diagnosis and experimental tasks were performed in our laboratory. Color vision diagnosis and experimental tasks with old observers were conducted in their place of residence. The order in which the mapping task and the best representative task were performed was counter balanced. BCT order was randomised across observers.
Observers performed two different searching tasks in natural binocularly viewing conditions. In the mapping task observers were asked to point to all of the stimuli nameable with each of the eleven BCTs. In the best representative task observers were asked to point to the corresponding focal colors. Stimuli were presented simultaneously. In both tasks the BCTs were indicated sequentially (in random order) by the experimenter, who registered the data in twelve previously printed sheets (one for each BCT in the mapping task and another for the best representative task).
Half of the observers performed the mapping task first and then the best representative task. The other half performed the tasks in the opposite order.
The performance of the two tasks took about two hours as maximum. Most of aged people could not perform the two tasks continuously, so they were allowed to rest several times during the experiment. Data were later codified in a digital database to be processed.
It was considered that there was a link between two BCTs if one or more stimuli were pointed to as belonging to both BCTs. Two strategies were considered to quantify link strength. The first was similar to the one described in Bonnardel (2006) and can be resumed as follows: for each participant an 11 x 11 confusion matrix was created with cell entry being one if the corresponding link between BCTs existed and zero otherwise. Individual matrices were added to derive a frequency confusion matrix for each group, which informed about the number of observers who showed every possible link between BCTs. The second strategy also begun with the identification of links between BCTs, but the quantification of the strength of the links was carried out in the manner specified below.
From both strategies an 11 x 11 confusion matrix was created for each group. These matrices were used as input for MDS analyses. The second strategy provided better fit values and very similar results to the first one. Because of such similarity, our commentaries will be restricted to the second strategy, from which matrices presented in Tables 1.A-1.C were derived (Table 1.A, normal young; B, normal aged; C, tritanomalous aged).
In order to illustrate the second strategy used to quantify the strength of the links between BCTs, we will focus on the first row of Table 1.A, which shows the links of the BCT white by the young normal trichromats group. For each stimulus the number of observers who used every BCT was computed, obtaining a 102 x 11 matrix. Then we focused on those samples pointed to when looking for whites and constructed a new 102 x 11 matrix that contained the distribution of the use of white. Some of these samples were also pointed to when looking for other BCT. From the distribution of the use of the BCT white we derived the percentage of use of white (75.6%) and of any other BCT (for example, 2.3% of the distribution of the use of white was also pointed to when looking for greens, 3.5% when looking for yellows, etc.). The percentages of each row percents add up to 100%.
As previously indicated, the matrices of Table 1 were used to perform non-metric MDS (1). Tridimensional (3D) solutions were forced, obtaining good normalized raw stress values for the three groups (normal young, .012, normal aged, .007, tritanomalous aged, .002). Figure 1 shows that these solutions were easily interpretable in terms of opponent mechanisms for the two normal groups (2). Their first dimension (D1) had red and green at its extremes (r = .95 between both D1 dimensions). The young group's second dimension (D2) and the normal aged group's third dimension (D3) had yellow and blue at their extremes (r = .79). Lastly, the young group's third dimension (D3) and the normal aged group's second dimension (D2) were very similar (r = .77) lightness dimensions (white and black at the extremes). The remaining correlations between normal young and aged groups' dimensions were always lower than r = .22 and, consequently, not relevant.
[FIGURE 1 OMITTED]
For the aged tritanomalous group, a bidimensional solution (Figure 2) provided easily interpretable dimensions and a good normalized raw stress value (.021). These two dimensions can be related to, respectively, the red-green (D1) and the white-black mechanisms (D2). These two dimensions were very highly related to their equivalents in the aged normal group (r = .93 between the two D1s, r = .94 between the two D2s).
Summing up, for the three groups, the first dimension (the most important one) was a red-green one. The second dimension was different for young (a yellow-blue dimension) and old (a lightness dimension) observers. Lastly, the importance of the yellow-blue dimension varied depending on the observer group. It was D2 for the normal young group and D3 for the normal aged group. There was no clear yellow-blue dimension for the tritanomalous aged group.
[FIGURE 2 OMITTED]
Hits and Errors
Table 2 shows correct (hits) and incorrect (errors) response percentages for both aged groups (only percentages over 3% appear). The criterion used to define hits and errors was as follows: A hit was computed whenever an aged observer pointed to a stimulus using the same BCT as the young group. If the BCTs used did not coincide, an error was computed. It was considered that a given stimulus belonged to the BCT more frequently used by the young group when pointing to such stimulus (see Bonnardel, 2006).
Table 2 diagonal shows the percentage of hits for each BCT for each aged group. Such percentages were always larger for the normal aged (no brackets) than for the tritanomalous aged group (in brackets). A Mann-Whitney U test indicated that the global percentage of hits was significantly higher (U = 35, p < .005) in the normal aged group (M = 87.0%, n =15) than in the tritanomalous aged group (M = 68.6%, n = 15). Significant differences were also observed when comparisons were performed separately only for the following BCTs: black (U = 50.5, p < .01), blue (U = 57.5, p < .05), brown (U = 39.5, p < .005), orange (U = 53 p < .01), purple (U = 37, p < .01), and grey (U = 62, p < .05). Table 2 diagonal values indicate that the hit range in the normal aged group (from 92.5% to 73.1%, [DELTA] = 19.4%) was smaller than in the tritanomalous aged group (from 87.9% to 44.9%, [DELTA] = 43%). The main cause of this difference was the low percentages for some BCTs in the tritanomalous aged group.
Table 2 shows two further important results. First, some specific errors only appeared in the tritanomalous aged group. For example, the normal aged group never pointed to blues when looking for greens, but this error appeared in the aged tritanomalous group (4.7%). Second, there were some asymmetries between categories. For example, in the aged tritanomalous group, they pointed to 30.9% of the green stimuli when looking for blues but, on the other hand, only selected 4.7% of the blues when looking for greens.
Figure 3 shows the mean pointing frequency (number of stimuli pointed to by each observer) for each BCT. For example, the three bars at the left indicate that when looking for white exemplars, observers selected a number of stimuli near to 5 (each bar represents one of the three groups). A series of Kruskal-Wallis nonparametric analyses of variance indicated that significant group differences appeared only for the following BCTs: black, [chi square](2,45) = 10.244, p < .01; green, [chi square] (2,45) = 8.261, p < .05; and purple, [chi square] (2,45) = 7.828, p < .05.
[FIGURE 3 OMITTED]
The application of a series of Mann-Whitney U analyses indicated that: (a) there were no significant differences between the two aged groups; (b) both aged groups pointed to more stimuli that the normal young group when looking for black (normal aged vs. normal young, U = 62.5, p < .05; tritanomalous aged vs. normal young: U = 45, p < .005), and to fewer stimuli than the normal young group when looking for purple (normal aged vs. normal young, U = 53, p < . 05; tritanomalous aged vs. normal young: U = 57, p < .05); and (c) the tritanomalous aged group pointed to fewer stimuli when looking for green than did young observers (tritanomalous aged vs. normal young: U = 40, p < .05).
Tritan Lines (Model C)
Figures 4 (normal) and 5 (tritanomalous) graphically represent the selections of both aged groups for two BCTs (A. Red, B. Blue), as compared with young observers. Crosses indicate normal young observers' selections. Circles show responses in the aged group (normal or tritanomalous). Size indicates frequency (the larger, the more frequent the selection). The spatial coincidence of crosses and circles corresponds to hits (stimuli pointed to both by young and aged observers). Circles alone are errors (stimuli only pointed to by aged observers).
[FIGURE 4 OMITTED]
[FIGURE 5 OMITTED]
Each CIE u'v' chromaticity diagram included in Figures 4 and 5 shows two tritan lines. The solid line corresponds to the hit median. The dashed line corresponds to the empirical error median. Tritan lines are defined using two points. A point corresponds to a hypothetical stimulus than only activates tritacones (and not proto and deuteracones). Its coordinates were u' = .257; v' = .000, according to Regan, Reffin and Mollon (1994). The other point corresponds to the stimulus median. Each tritan line allows the specification of what we call the "tritan angle" (the one formed with the abscissa axis). For example, it can be seen in Figure 4.B that, for red, the error median angle is approximately 90[degrees].
According to Model C predictions, hit and error tritan lines (and tritan angles) must be similar. As Figures 4 and 5 indicate, this prediction is not confirmed for the two BCTs represented. Confirming the graphic impression, a series of Mann Whitney U-tests indicated that there were significant differences between tritan angles (p < .05) of hits and errors for 9 of the 11 BCTs (black, red, yellow, blue, brown, pink, orange, purple, and grey).
Filtering without Compensation (Model A)
According to Model A, aged people's color perception (and BCT use) is similar to that expected in young people when responding to the distorted stimulation created by an aged lens. To evaluate this prediction, the following strategy was used. Firstly, for each stimulus, the transformation produced by an aged lens (chromatic coordinates and lightness transformation) was calculated. Secondly, young people's expected response for such transformed stimuli was determined (according to Model A, this was the response expected in aged people).
[FIGURE 6 OMITTED]
To transform the original coordinates (u', v') into the ones that mimic the effect of an aged lens ([u'.sub.f], [v'.sub.f]), we used Pokorny et al.'s (1987) aging model, assuming an age of 88.47 years (the mean age of the tritanomalous group). Figure 6 shows the effect of this transformation. On the left (6.A), the original stimulus chromatic coordinates are shown. On the right (6.B) are those corresponding to the transformed stimuli.
Lens aging also changes stimulus lightness. To estimate transformed lightness ([L*.sub.f]), the following equation was used:
[L*.sub.f] = 116 [([Y.sub.f]/[Y.sub.fn]).sup.1/3] - 16
This equation is similar to the one usually used to compute L* (standard lightness, see Hunt & Pointer, 2011) but now, the luminance factor is computed from filtered luminances ([Y.sub.f] and [Y.sub.fn]) instead of from standard ones (Y and [Y.sub.n]).
From transformed lightness ([L*.sub.f]) and chromatic coordinates ([u'.sub.f], [v'.sub.f]), it was possible to compute [u*.sub.f] and [v*.sub.f] in the same way that [* and v* are usually computed from u', v' and L* (see Hunt & Pointer, 2011). Later, we determined which stimulus belonging to the original set had L*u*v* values nearest to the [L*.sub.f], [u*.sub.f] and [v*.sub.f] values of each transformed stimulus. Table 3 (Model A predictions) shows the results of these calculations for each BCT.
Table 3 predicts no hits for six BCTs (white, black, blue, pink, purple and grey). This means that none of the stimuli that the young observers pointed to as members of these BCTs have transformed trichromatic values ([L*.sub.f], [u*.sub.f], and [v*.sub.f]) similar to the ones (L*, u*, and v*) usually associated with such categories. Table 3 also makes very specific predictions related to aged people's expected errors. For example, the second row indicates that it predicted that all the expected selections when looking for black stimuli must be errors and that these errors correspond to stimuli that are blues for young people.
Comparison between, on the one hand, Table 3 diagonal percentages and, on the other, the Table 2 equivalents (aged groups' empirical responses) indicates that, with the only exception of green, Model A predicted lower percentages of hits than the actual empirical hits. A series of goodness-of-fit tests showed that, except for green, Model A predicted higher error percentages than the empirical errors for all the BCTs, where such analyses were possible (red, yellow, brown, and orange, p < .001). Readers are reminded that the model did not predict any selections for white, blue, pink, and purple. No hits were predicted for black and grey. No errors were predicted for green.
Two columns in Table 4 (labeled "[R.sup.2] Model A") show the proportion of empirical variance explained by Model A for both aged groups, considering the whole set of selections ("Global", [R.sup.2] = .14 and .10) and each BCT separately. There are no values for the BCTs white, blue, pink, and purple because Model A predicts no hits for these categories (Table 3). It can be observed that, except for green, only minor proportions of empirical variance were explained.
Filtering with Compensation (Model B)
Model B (white normalization) assumes that aged people compensate lens aging effects by using a von Kries-type compensation mechanism. Such a mechanism adjusts the relative weight of each cone type to obtain responses of similar magnitude for white stimuli. For the predictions of Model B, we used the following procedure. Firstly, the responses of the three cone types (L, M, and S) for each stimulus were computed using the following equations:
L = [K.sub.L] [700.summation over (400)] [L.sub.e] ([lambda]) [L.sub.2] ([lambda]) [DELTA][lambda]
M = [K.sub.M] [700.summation over (400)] [L.sub.e] ([lambda]) [M.sub.2] ([lambda]) [DELTA][lambda]
S = [K.sub.S] [700.summation over (400)] [L.sub.e] ([lambda]) [S.sub.2] ([lambda]) [DELTA][lambda]
L ([lambda]), M ([lambda]), and S ([lambda]) are the fundamentals of, respectively, proto (L), deutera (M), and trita (S) cones, according to Stockman and Sharpe (2000). [L.sub.e] ([lambda]) is the spectral radiance, and [K.sub.L], [K.sub.M] and [K.sub.S] are constants adjusted to rescale the relative cone responses. This rescaling made the L:M:S relation equal to 1:1:1 for the reference white. Subsequently, the LMS values were recalculated using the following equations:
[L.sub.f] = [K.sub.Lf] [700.summation over (400)] T([lambda])[L.sub.e] ([lambda]) [L.sub.2] ([lambda]) [DELTA][lambda]
[M.sub.f] = [K.sub.Mf] [700.summation over (400)] T([lambda])[L.sub.e] ([lambda]) [M.sub.2] ([lambda]) [DELTA][lambda]
[S.sub.f] = [K.sub.Sf] [700.summation over (400)] T([lambda])[L.sub.e] ([lambda]) [S.sub.2] ([lambda]) [DELTA][lambda]
T([lambda]) is the extra filtering produced by the lens aging, and [K.sub.Lf], [K.sub.Mf], and [K.sub.Sf] are constants adjusted to rescale the L:M:S (again 1:1:1 for the reference white).
Table 5 shows Model B predictions for each BCT. A quick visual comparison with the empirical data shown in Table 2 reveals that Model B predicted higher percentages of hits than actual empirical hits. Confirming this impression, a series of goodness-of-fit tests indicated that, except for orange, Model B significantly (p < .001) predicted more hits for all the BCTs in the tritanomalous aged group. However, Model B only predicted significantly more responses for four BCTs for the normal aged group (no aphakic observers): black and red (p < .001), blue (p < .001), and purple (p < .005).
Two columns in Table 4 (labeled "[R.sup.2] Model B") show the proportion of variance explained by Model B for the aged groups. This information is provided both for the full set of responses ("Global", [R.sup.2] = .91 and .97) and for each BCT. Two other columns (labeled [DELTA] [R.sup.2]) present the increase in the proportion of explained variance when using Model B instead of Model A. This increase was remarkable both at the global level ("Global", [DELTA] [R.sup.2] = .77 and .87) and for most of the BCTs separately (green was the only exception).
Best Representative Searching Task
Except for brown and purple (93%, both BCTs), the normal young group obtained 100% of hits for the BCTs in this task. That is, for nine BCTs, young observers always chose as the best representatives of a BCT stimuli that were correct responses according to the criteria defined from the mapping task criterion. Another important result was that the stimulus pointed to as the best representative of each BCT in the young group was very frequently pointed to as belonging to this category in the mapping task for both aged groups. More specifically, the percentages of selections were (for the normal, and the tritanomalous aged group, respectively): White (100, 100), black (100, 93), red (93, 93), green (100, 100), yellow (100, 100), blue (100, 80), brown (100, 93), pink (80, 80), orange (100, 80), purple (60, 47), grey (87, 60).
As Table 6 diagonal indicates that normal aged observers obtained 100% of hits for 8 BCTs (all except for black, brown, and grey, data without brackets) but in the tritanomalous aged were more accurate than the tritanomalous aged group. Mann-Whitney U tests confirmed the previous statement at a global level (p < .05 for the comparison between each normal group and the tritanomalous aged group). When the comparison was performed for each BCT, a series of chi square analyses showed the same pattern of significant differences for the BCTs blue, orange, and grey.
Comparing the diagonals from Tables 2 and 6, it can be observed that, with only one exception (tritanomalous grey), the percentages of hits in the best representative task were higher than their equivalents in the mapping task. A series of Wilcoxon tests confirmed this impression for the normal aged group (medians: 100 vs. 87.01, Z = -3.18, p < .001), and for the tritanomalous aged group (medians: 90.91 vs. 68.04, Z = -3.35,p < .001).
Aged people, especially the tritanomalous, do not use BCTs as younger people do. Differences appeared in the nature and relevance of the dimensions resulting from the links between BCTs (Figures 1 and 2) and in the errors they made in the pointing tasks.
Although some aged people make tritan errors when responding to standard diagnostic tests (Birch, 2001), it appears inaccurate to call them tritanomalous when considering how they use BCTs. Specifically, it was observed that, in contrast to the prediction of Model C, hit and error tritan lines were significantly different for 9 BCTs (all except for white and green). This result can be explained considering that lens aging changes retinal stimulation for the three cone types and, consequently, not only for tritacones (S cones). As our results indicate, BCT use depends on the information provided by the three cone types, and BCT confusions cannot be explained only considering the reduction of tritacone responses due to lens aging.
Model A only explained a reduced proportion of the mapping task variance (Table 4) because it predicted (Table 3) less accurate performances than the real ones. Consequently, it is a mistake to think that aged people's chromatic perception (and BCT use) is similar to that expected in young people when responding to stimuli transformed by an aged lens. Similarity between the predictions of Model B (Table 5) and empirical results (Table 3) clearly reveal the working of a von Kries-type compensation mechanism (similar to the one described by Neitz et al., 2002). Model B explained 91% of the empirical variance for aged tritanomalous and 97% for normal non-aphakic aged observers. Despite its accuracy, this model predicted (Table 5) significantly better performances than the empirical ones for ten BCTs in the tritanomalous aged group. There were only four BCTs with superior performances for normal aged people.
Our research can be considered an extension of Hardy et al.'s (2005), in which the analysis was extended to all the BCTs. Similarly to Hardy et al., we observed that aged observers' BCT use is more similar to young people's use than to the predictions of Model A. Nevertheless, aged observers made some important naming errors. This fact is in accordance with the idea, advanced by Kraft and Werner (1994, p. 1220), of considering a perfect compensation of lens aging effects impossible. Such a "Perfect compensation would require a function that is the exact inverse of the spectral absorption of the ocular media. Precise wavelength information is lost in the transduction process, however, so that an arbitrary spectral modification function cannot be constructed, and the visual system cannot exactly compensate for lenticular senescence", Kraft and Werner (1994, p. 1220).
Two results suggest that lens aging produces a reduction in the blue-yellow mechanism functionality. Firstly, MDS analysis (Figures 1 and 2) showed that the relevance of the blue-yellow dimension varied depending on the observers' group. For young people, it was the second dimension and for the normal aged group, it was the third dimension, but it was not well defined for the tritanomalous group. Secondly, the three BCTs with the lower percentages of hits (Table 2) were black, blue, and purple, the categories whose use may be more dependent on the information provided by the tritacones to the blue-yellow mechanism.
Lindsey and Brown (2002) hypothesized that lens aging must produce important reductions in the use of some BCTs (the ones, such as blue and purple, associated with the detection of energy in short wavelengths) and a corresponding increase in others (such as green and black). According to this hypothesis, the 'grue' term appears because many stimuli that are blues for young people became greens for aged people. This prediction agrees with the changes in the chromaticity diagrams associated with lens aging (Lindsey and Brown, Figure 3; this paper, Figure 6) and with the prediction of not using blue in the mapping task (Table 3). However, as Figure 3 shows, there were few significant differences in BCT use frequency between young and aged people. In fact, such differences only partially agree with Lindsay and Brown's predictions. As they expected, both aged groups classified more stimuli as blacks and fewer stimuli as purple compared to the young observers. However, the aged tritanomalous observers selected fewer stimuli as greens compared to young observers (and there were no significant differences between any pair of groups for blue).
Figure 2 shows that aged people, especially the tritanomalous, have difficulty differentiating between some categories. For example, when tritanomalous looked for blues, they pointed to a significant percentage of greens (30.9%). A similar pattern appeared between other pairs of BCTs (black and blue, brown and purple, purple and pink, etc.). Such difficulties could be related to the existence of some BCTs that, like 'grue', are sometimes called "composites" (Jameson, 2005b; Kay, Berlin, Maffi, & Merrifield, 1997; Kay et al., 2010; Lindsey & Brown, 2009).
Lindsey and Brown's (2002) Figure 2 shows that there is a relationship between latitude and the number of languages that include 'grue' (the closer to the equator, the higher the probability). However, this correlation is not perfect and consequently, other factors could be related to the existence of 'grue'. For example, the "distribution, frequency and relative importance of different perceptual groups within a society" (Jameson, 2005a, p. 323). That is, many languages spoken near the equator appear in low-developed societies where: (a) high exposure to ultraviolet radiations and low frequency of cataract surgery could produce an important proportion of aged tritanomalous (who have difficulties differentiating some stimuli); (b) these people could play a relevant social role; (c) for these people, 'grue' would facilitate social communication. This kind of explanation, nowadays very speculative, could also be used with regard to the other important errors appearing in Table 2 (i.e., composite BCTs could derive from the confusion of blue and green, blue and black, brown and purple, etc.).
Table 2 can be used to select colors according to Universal Design premises (Vanderheiden, 2006). According to this conceptual framework, objects should be designed to promote accurate use in the largest possible number of potential users. For example, when designing a political map, colors should facilitate the perceptual differentiation of different countries. Considering aged people's confusion of BCTs, the use of some pairs of colors may be problematic and should be avoided. More specifically, neighbouring countries that must be differentiated (like France and Germany) should not be printed in colors belonging to BCTs that could be confused (e.g., some blues and some greens). If, for some reason, this is unavoidable, the best alternative would be to use young observers' best representatives of BCTs. As indicated, these stimuli were accurately pointed to in the mapping task in all three groups. Consequently, they were not confused with stimuli belonging to other BCTs.
To end, it must be emphasised that our research provided the first global analysis on aged people's use of BCTs both at the descriptive and explicative levels. The findings show a remarkable similarity between aged and young people's use of BCTs, which indicates some kind of neural compensation (white normalisation) for the effects of the ocular changes produced by the aging process. Such compensation was not perfect, especially for the tritanomalous, and aged people showed important peculiarities in their use of [deleted 'the'] BCTs. Among these peculiarities were [deleted 'the'] confusions between some specific colour terms (i.e. [deleted 'there were'] significant errors between blue and green or black) and a reduction in the importance of the information provided by the yellow-blue mechanism
Stimuli used in the searching tasks. First column indicates each stimulus NCS notation. Other columns inform about luminance factor (Y/Yn), chromatic co-ordinates (u', v') and the most frequent (mode) Basic Colour Term used for naming a stimulus in the Trichromatic Normal (TN) group. NCS Y/[Y.sub.n] u' v' Mode Notation (TN) S 0505-Y 61,87 0,197 0,493 White S 0510-Y 58,49 0,200 0,500 Yellow S 0580-Y 47,09 0,234 0,554 Yellow S 1030-Y 47,02 0,211 0,521 Yellow S 1070-Y 39,94 0,231 0,548 Yellow S 2005-Y 42,11 0,195 0,492 Grey S 2060-Y 28,75 0,227 0,543 Yellow S 2502-Y 47,82 0,195 0,488 Grey S 6030-Y 7,70 0,225 0,531 Brown S 7020-Y 5,21 0,223 0,522 Brown S 7502-Y 9,36 0,194 0,488 Grey S 8502-Y 4,77 0,195 0,490 Black S 4050-Y20R 13,66 0,256 0,536 Brown S 6030-Y30R 7,69 0,250 0,525 Brown S 0585-Y50R 23,17 0,333 0,538 Orange S 1070-Y50R 24,40 0,304 0,531 Orange S 6005-Y50R 16,50 0,204 0,492 Brown S 6010-Y50R 10,70 0,218 0,499 Brown S 6020-Y50R 8,66 0,240 0,510 Brown S 7010-Y50R 7,07 0,221 0,501 Brown S 0560-Y60R 31,52 0,289 0,520 Orange S 0570-Y60R 26,01 0,309 0,525 Orange S 1020-Y60R 41,00 0,225 0,501 Orange S 2060-Y60R 17,98 0,300 0,521 Orange S 3020-Y60R 24,46 0,228 0,502 Brown S 3050-Y60R 14,52 0,288 0,519 Brown S 4030-Y60R 14,68 0,257 0,510 Brown S 5030-Y60R 9,46 0,267 0,513 Brown S 5040-Y60R 7,16 0,294 0,520 Brown S 7020-Y60R 4,26 0,260 0,513 Brown S 0585-Y70R 16,98 0,381 0,528 Orange S 1580-Y80R 10,55 0,383 0,518 Red S 0580-Y90R 14,85 0,381 0,511 Red S 1580-Y90R 8,98 0,404 0,513 Red S 3060-Y90R 8,50 0,339 0,505 Red S 6030-Y90R 4,50 0,281 0,500 Brown S 0540-R 33,97 0,254 0,492 Pink S 1030-R 34,59 0,233 0,485 Pink S 1070-R 14,49 0,353 0,497 Red S 8502-R 3,25 0,211 0,494 Black S 2070-R10B 7,55 0,383 0,489 Red S 0505-R20B 51,92 0,201 0,485 White S 0510-R20B 53,76 0,204 0,483 Pink S 2020-R20B 30,07 0,216 0,478 Pink S 2060-R20B 12,21 0,318 0,470 Pink S 6020-R20B 6,26 0,245 0,477 Purple S 0510-R30B 53,60 0,203 0,481 Pink S 1040-R30B 31,54 0,235 0,464 Pink S 2005-R30B 38,05 0,197 0,481 Grey S 2010-R30B 35,91 0,202 0,479 Pink S 2020-R30B 32,01 0,211 0,474 Pink S 3020-R30B 21,66 0,215 0,475 Pink/Purple S 3040-R30B 14,40 0,253 0,463 Pink S 4040-R30B 8,56 0,273 0,462 Purple S 3040-R40B 14,46 0,241 0,452 Purple S 3050-R40B 9,63 0,267 0,437 Purple S 3020-R50B 24,14 0,202 0,467 Purple S 3030-R50B 19,50 0,212 0,455 Purple S 3050-R50B 11,10 0,239 0,426 Purple S 4010-R50B 20,44 0,198 0,474 Grey S 5030-R50B 8,33 0,222 0,444 Purple S 7010-R50B 6,43 0,199 0,469 Purple/Grey S 7020-R50B 3,32 0,222 0,444 Purple S 4030-R60B 12,38 0,197 0,442 Purple S 4040-R60B 8,92 0,203 0,422 Purple S 2060-R70B 13,68 0,180 0,398 Blue S 2060-R80B 15,12 0,156 0,395 Blue S 0510-R90B 54,28 0,187 0,478 White S 2060-R90B 14,69 0,142 0,404 Blue S 4010-R90B 20,56 0,184 0,473 Grey S 4020-R90B 17,08 0,175 0,461 Grey S 4550-R90B 4,62 0,137 0,381 Blue S 7020-R90B 3,51 0,160 0,441 Blue S 8010-R90B 2,93 0,171 0,456 Blue S 5502-B 19,91 0,186 0,477 Grey S 8502-B 4,14 0,183 0,474 Black S 4030-B30G 16,52 0,152 0,466 Green S 4050-B30G 10,42 0,121 0,451 Green S 2060-B50G 19,51 0,117 0,470 Green S 3555-B80G 11,22 0,109 0,488 Green S 0505-G10Y 59,75 0,193 0,488 White S 0520-G10Y 54,74 0,179 0,498 Green S 3060-G10Y 13,24 0,127 0,530 Green S 4010-G10Y 21,38 0,179 0,489 Green S 4020-G10Y 19,12 0,169 0,497 Green S 5040-G10Y 8,06 0,139 0,519 Green S 7020-G10Y 4,79 0,157 0,508 Green S 8010-G10Y 3,31 0,168 0,498 Black S 8502-G10Y 3,31 0,168 0,498 Black S 4005-G20Y 23,43 0,186 0,486 Grey S 2050-G30Y 24,80 0,168 0,528 Green S 0530-G50Y 53,68 0,191 0,518 Green S 4550-G50Y 10,77 0,184 0,543 Green S 1075-G80Y 37,07 0,209 0,554 Yellow S 2070-G80Y 26,71 0,211 0,554 Yellow S 1075-G90Y 38,57 0,222 0,553 Yellow S 0500-N 81,40 0,194 0,484 White S 1000-N 71,71 0,194 0,484 White S 1500-N 63,39 0,194 0,485 Grey S 5500-N 21,22 0,191 0,482 Grey S 7500-N 10,17 0,190 0,481 Grey S 9000-N 2,51 0,186 0,481 Black
Received January 26, 2011
Revision received September 30, 2011
Accepted November 4, 2011
Berlin, B., & Kay, P. (1969). Basic color terms: Their universality and evolution. Berkeley, CA: University of California Press.
Birch, J. (2001). Diagnosis of defective color vision. Oxford, England: Butterworth-Heinemann.
Bonnardel, V. (2006). Color naming and categorization in inherited color vision deficiencies. Visual Neuroscience, 23, 637-643. http://dx.doi.org/10.1017/S0952523806233558
Boynton, R. M., & Olson, C. X. (1987). Locating basic colors in the OSA space. Color Research and Application, 12, 94-105. http://dx.doi.org/10.1002/col.5080120209
Delahunt, P., Webster, M. A., Ma, L., & Werner, J. S. (2004). Color appearance changes after cataract surgery reveal a long-term chromatic adaptation mechanism. Visual Neuroscience, 21, 301-307.
Fletcher, R J. (1980). The City University Color Vision Test (2nd Ed.). London, England: Keeler.
Foster, D. H., Nascimento, S. M. C., Craven, B. J., Linnell, K. J., Cornelissen, F. W., & Brenner, E. (1997). Four issues concerning colour constancy and relational colour constancy. Vision Research, 37, 1341-1345. http://dx.doi.org/10.1016/S0042-6989(96)00285-4
Franklin, A., Clifford, A., Williamson, E., & Davies, I. R. L. (2005). Color term knowledge does not affect categorical perception of color in toddlers. Journal of Experimental Child Psychology, 90, 114-141. http://dx.doi.org/10.1016/jjecp. 2004.10.001
Hardy, J. L., Frederick, C. M., Kay, P., & Werner, J. S. (2005). Color naming, lens aging, and grue: What the optics of the aging eye can teach us about color language. Psychological Science, 16, 321-327. http://dx.doi.org/10.1111/j.0956-7976.2005.01534.x
Hunt, R. W. G., & Pointer, M. R. (2011). Measuring colour (4th Ed.). Chichester, England: John Wiley & Sons, Ltd. http://dx.doi.org/10.1002/9781119975595
Ishihara,M. D. (1996). Ishihara's tests for colour-deficiency: 24 Plates Edition. Tokyo, Japan: Kanehara & Co., Ltd.
Jameson, K. A. (2005a). Culture and cognition: What is universal about the representation of color experience? Journal of Cognition and Culture, 5, 293-348. http://dx.doi.org/10.1163/156853705774648527
Jameson, K. A. (2005b). Why GRUE? An interpoint-distance model analysis of composite color categories. Cross-Cultural Research 39, 159-204. http://dx.doi.org/10.1177/1069397104273766
Kaiser, P. K., & Boynton, R. M. (1996). Human color vision (2nd Ed.). Washington, DC: Optical Society of America.
Kay, P., Berlin, B., Maffi, L., & Merrifield, W. R. (1997). Color naming across languages. In C. L. Hardin & L. Maffi (Eds.), Color categories in thought and language (pp. 21-56). Cambridge, England: Cambridge University Press. http://dx.doi.org/10.1017/CBO9780511519819.002
Kay, P., Berlin, B., Maffi, L., Merrifield, W. R., & Cook, R. (2010). World color survey. Stanford, CA: Center for the Study of Language and Information (CSLI).
Knoblauch, K., Saunders, F., Kusuda, M., Hynes, R., Podgor, M., Higgins, K. E., & de Monasterio, F. M. (1987). Age and iluminance effects in the Farnsworth-Munsell 100-hue Test. Applied Optics, 26, 1441-1448. http://dx.doi.org/10.1364/AO.26.001441
Kraft, J. M., & Werner, J. S. (1999a). Aging and the saturation of colors. 1. Colorimetric purity discrimination. Journal of the Optical Society of America, A, 16, 223-230. http://dx.doi.org/10.13 64/JOSAA.16.000223
Kraft, J. M., & Werner, J. S. (1999b). Aging and the saturation of colors. 2. Scaling of color appearance. Journal of the Optical Society of America, A, 16, 231-235. http://dx.doi.org/10.1364/JOSAA.16.000231
Kraft, J. M., & Werner, J. S. (1994). Spectral efficiency across the life span: Flicker photometry and brightness matching. Journal of the Optical Society of America, A, 11, 1213-1221. http://dx.doi.org/10.1364/JOSAA.11.001213
Lanthony, P. (1985). Album tritan. Paris, France: Luneau Ophtalmologie.
Lillo, J., Davies, I., Collado, J., Ponte, E., & Vitini, I. (2001). Colour naming by colour blind children. Anuario de Psicologia, 32, 5-24.
Lillo, J., & Moreira, H. (2005). Relative luminance and figure-background segmentation problems: Using AMLA to avoid non-discernible stimulus pairs in common and color blind observers. Psicologica, 26, 189-207.
Lillo, J., Moreira, H., Alvaro, L., & Davies, I. (2012). Use of Basic Color Terms by red-green dichromats: 1. General description. Color Research and Application. Manuscript submitted for publication.
Lillo, J., Moreira, H., Vitini, I., & Martin, J. (2007). Locating basic Spanish color categories in CIE L*u*v* Space: Identification, lightness segregation and correspondence with English equivalents. Psicologica, 28, 21-54.
Lindsey, D. T., & Brown, A. M. (2002). Color naming and the phototoxic effects of sunlight on the eye. Psychological Science, 13, 506-512. http://dx.doi.org/10.1111/1467-9280.00489
Lindsey, D. T., & Brown A. M. (2009). World color survey color naming reveals universal motifs and their within-language diversity. Proceedings of the National Academy of Sciences of the United States of America of the USA, 106, 19785-19790.
Lin, H., Luo, M. R., MacDonald, L. W., & Tarrant, A. W. S. (2001a). A cross-cultural colour-naming study. Part I: Using an unconstrained method. Color Research and Application, 26, 40-60. http://dx.doi.org/10.1002/1520-6378(200102)26:1 <40::AID-COL5>3.0.CO;2-X
Lin, H., Luo, M. R., MacDonald, L. W., & Tarrant, A. W. S. (2001b). A cross-cultural colour-naming study. Part II: Using a constrained method. Color Research and Application, 26, 193-208. http://dx.doi.org/10.1002/col.1017.abs
Mellerio, J. (1987). Yellowing of the human lens: Nuclear and cortical contributions. Vision Research, 27, 1581-1587. http://dx.doi.org/10.1016/0042-6989(87)90166-0
Moreira, H. (2010). Uso de terminos de color basicos en daltonicos dicromatas y personas de edad avanzada [Basic color terms use by red-green dichromats and by aged people]. (Doctoral thesis). Universidad Complutense, Madrid, Espana.
Moreira, H., Lillo, J., Alvaro, L., & Davies, I. (2012). Use of basic color terms by red-green dichromats: 2. Models. Color Research and Application. Manuscript submitted for publication.
Neitz, J., Carroll, J., Yamauchi, Y., Neitz, M., & Williams, D. R. (2002). Color perception is mediated by a plastic neural mechanism that is adjustable in adults. Neuron, 35, 783-792. http://dx.doi.org/10.1016/S0896-6273(02)00818-8
Norren, D. V., & Vos, J. J. (1974). Spectral transmission of the human ocular media. Vision Research, 14, 1237-1244. http://dx.doi.org/10.1016/0042-6989(74)90222-3
Obama, T., Ikeda, M., Sagawa, K., & Shinoda, H. (2005). Range of similar colours with and without cataract experiencing goggles. In J. L. Nieves & J. Hernandez-Andres (Eds.), Proceedings book of AIC Colour 05 (pp. 13-16). Granada, Espana: Comite Espanol del Color.
Pokorny, J., Smith, V., & Lutze, M. (1987). Aging of the human lens. Applied Optics, 26, 1437-1440. http://dx.doi.org/10.1364/ AO.26.001437
Regan, B. C., Reffin, J. P., & Mollon J. D. (1994). Luminance noise and the rapid determination of discrimination ellipses in colour deficiency. Vision Research. 34, 1279-1299. http://dx.doi.org/10.1016/0042-6989(94)90203-8
Regier, T. Kay, P., & Khetarpal, N. (2009). Color naming and the shape of color space. Language, 85, 884-892.
Roberson, D., Davidoff, J., Davies, I. R. L., & Shapiro, L. (2005). Color categories in Himba: Evidence for the cultural relativity hypothesis. Cognitive Psychology, 50, 378-411. http://dx.doi.org/10.1016/j.cogpsych.2004.10.001
Sagawa, K., & Takahashi, Y. (2001). Spectral luminous efficiency as a function of age. Journal of the Optical Society of America A, 18, 2659-2667. http://dx.doi.org/10.1364/JOSAA.18.002659
Schefrin, B. E., & Werner, J. S. (1990). Loci of spectral unique hues throughout the life span. Journal of the Optical Society of America A, 7, 305-311. http://dx.doi.org/10.1364/JOSAA. 7.000305
Schieber, F. (2006). Vision and aging. In J. E. Birren & K. Warner Schaie (Eds.), Handbook of the psychology of aging (6th Ed., pp. 129-161). San Diego, USA: Elsevier Academic Press. http://dx.doi.org/10.1016/B978-012101264-9/50010-0
Schirillo, J. A. (2001). Tutorial on the importance of color in language and culture. Color Research and Application, 26, 179-192. http://dx.doi.org/10.1002/col.1016
Scandinavian Color Institute (1997). NCS Index Edition 2. Stockholm, Sweden: Scandinavian Color Institute.
Stockman, A., & Sharpe, L. T. (2000). Spectral sensitivities of the middle- and long-wavelength sensitive cones derived from measurements in observers of known genotype. Vision Research, 40, 1711-1737. http://dx.doi.org/10.1016/S00426989(00)00021-3
Sturges, J., & Whitfield, T. W. A. (1995). Locating basic colours in the Munsell space. Color Research and Application, 20, 364-376. http://dx.doi.org/10.1002/col.5080200605
Vanderheiden, G. C. (2006). Design for people with functional limitations. In Gavriel Salvendy (Ed.), Handbook of human factors and ergonomics (3rd Ed., pp. 1387-1417). New York, NY: Wiley. http://dx.doi.org/10.1002/0470048204.ch53
Verriest, G. (1963). Further studies on acquired deficiency of color discrimination. Journal of the Optical Society of America A, 53, 185-197. http://dx.doi.org/10.1364/JOSA.53.000185
Weale, R. A. (1986). Aging and vision. Vision Research, 26, 1507-1512. http://dx.doi.org/10.1016/0042-6989(86)90170-7
Weale, R. A. (1988). Age and the transmittance of the human crystalline lens. Journal of physiology, 395, 577-587.
Webster, M. A., Halen, K., Meyers, A. J., Winkler, P., & Werner, J. S. (2010). Color appearance and compensation in the near periphery. Proceedings of the Royal Society B: Biological Sciences, 277, 1817-1825. http://dx.doi.org/10.1098/rspb.2009.1832
Werner, J. S., & Schefrin, B. E. (1993). Loci of achromatic points throughout the life span. Journal of the Optical Society of America A, 10, 1509-1516. http://dx.doi.org/10.1364/JOSAA.10.001509
Winn, B., Whitaker, D., Elliott, D. B., & Phillips, N. J. (1994). Factors affecting light-adapted pupil size in normal human subjects. Investigative Ophthalmology and Visual Science, 35, 1132-1137.
Wuerger, S., Xiao, K., Fu, C., & Karatzas, D. (2010). Colour-opponent mechanisms are not affected by age-related chromatic sensitivity changes. Ophthalmic and Physiological Optics, 30, 653-659. http://dx.doi.org/10.1111/j.1475-1313.2010.00744.x
(1) MDS analyses were performed using Proxscal (DTSS, Faculty of Social and Behavioral Sciences, Leiden University, The Netherlands), SPSS v 15.0. Proxcal minimizes the normalized raw stress value, a measure of departure from good fit that ranges from 0 to 1: the smaller this value, the better the fit.
(2) We have not used INDSCAL analysis because the focus of this paper is on differences between groups (not on individual differences), and we wanted specific solutions for each group. To compare quantitatively the similarity of the dimensions obtained for different groups we used Spearman correlation coefficient, a useful way to facilitate the interpretation of the dimensions obtained in MDS solutions.
Julio Lillo, Humberto Moreira, Leticia Perez del Tio, Leticia Alvaro, and Maria del Carmen Duran
Universidad Complutense de Madrid (Spain)
This work was supported by Project 2008-04 166/PSIC, from the Spanish Ministerio de Ciencia y Tecnologia and by FPI Grant UCM (BE48/09).
Correspondence concerning this article should be addressed to Julio Lillo. Facultad de Psicologia. Campus de Somosaguas. 28223 Madrid (Spain). Phone:+34-913943198. E-mail: email@example.com
Table 1 Mapping task results 1.A. Normal young. W Bla R Gre Y Blu Br Pi W 75.6 2.3 3.5 2.3 4.6 Bla 88.9 2.2 3.3 1.1 R 83.3 5.9 Gre 1.0 1.0 89.7 4.1 1.6 Y 2.7 7.3 85.5 4.6 Bin 2.2 3.3 3.3 86.8 Br 0.5 2.4 93.3 Pi 3.2 4.9 76.3 O 6.4 4.1 2.1 Pu 0.5 2.6 1.1 9.0 Gry 6.5 1.9 3.2 1.3 2.6 O Pu Gry W 11.6 Bla 1.1 3.3 R 6.0 4.8 Gre 2.6 Y Bin 2.2 2.2 Br 1.9 1.9 Pi 1.6 13.9 O 87.5 Pu 85.7 1.1 Gry 1.3 83.2 1.B. Normal aged. W Bla R Gre Y Blu Br Pi W 83.5 5.5 3.3 5.5 Bla 75.0 5.4 4.1 9.3 R 69.2 0.8 0.8 2.4 11.5 Gre 3.8 0.6 78.7 4.0 4.0 2.1 0.6 Y 5.3 7.5 85.1 Blu 3.0 5.1 1.0 7.1 68.7 1.0 2.0 Br 6.0 1.6 2.0 0.5 81.9 1.1 Pi 3.8 10.8 0.8 1.5 1.5 58.1 O 9.6 2.3 5.4 Pu 0.6 5.6 1.2 3.5 1.8 15.0 Gry 1.4 4.5 0.7 6.1 1.4 4.2 5.9 0.7 O Pu Gry W 2.2 Bla 0.8 5.3 R 6.8 7.7 0.8 Gre 1.2 5.0 Y 2.1 Blu 6.1 6.1 Br 1.1 1.6 4.4 Pi 3.6 19.3 0.8 O 80.7 2.0 Pu 1.1 69.5 1.8 Gry 2.1 73.2 1.C. Tritanomalous aged. W Bla R Gre Y Blu Br Pi W 84.4 2.8 0.9 5.6 Bla 53.6 4.5 0.4 7.9 15.3 R 56.7 1.3 4.0 14.4 Gre 4.7 58.9 4.2 14.8 3.8 Y 2.3 0.6 1.5 6.5 65.2 4.3 3.2 3.1 Blu 9.1 16.4 3.0 47.7 7.2 0.6 Br 0.3 8.9 1.6 2.2 1.1 3.6 52.6 1.4 Pi 4.0 13.8 2.6 0.7 3.2 49.7 O 13.69 2.94 8.27 0.76 2.80 14.32 Pu 0.7 8.3 5.1 2.8 4.6 21.1 6.09 Gry 1.9 5.4 0.4 5.9 2.4 5.6 13.6 0.8 O Pu Gry W 1.9 4.5 Bla 11.7 6.7 R 12.6 10.3 0.7 Gre 2.0 4.2 7.6 Y 8.7 4.8 Blu 0.6 7.4 8.0 Br 1.1 17.3 9.9 Pi 12.7 11.9 1.3 O 53.7 3.51 Pu 1.6 39.4 10.3 Gry 11.7 52.4 Note. Each row indicates pointing percentages for a BCT and group. A: Normal young. B: Normal aged. C: Tritanomalous aged. W=White. Bla=Black. R=Red. Gre=Green. Y=Yellow. Blu=Blue. Br=Brown. Pi=Pink. O=Orange. Pu=Purple. Gry=Grey. Table 2 Hits Percentages (Diagonal) and Error distribution for Normal Aged (without Brackets) and Tritanomalous (in Brackets) Groups in the Mapping Task W Bla R Gre Y Blu W 89.5 5.2 (76.1) (4.5) (3.3) Bla 76.8 4.4 11.6 (58.3) (9.5) (11.7) R 78.2 (74.4) Gre 3.8 92 (87.9) (4.9) (4.7) Y 4.6 92.5 (6.9) (82.0) Blu 4.7 11.5 75 (7.7) (3.9) (49.5) Br 8 (9.3) Pi 5.6 (12.5) O (13.3) (10.9) Pu (11.6) (3.4) (3.9) Gry 4.8 5.7 (5.5) (8.8) Br Pi O Pu Gry W 3.5 (13.4) Bla 3.6 (10.4) (5.9) (4.2) R 7.6 10.5 (8.1) (7.5) (10.0) Gre Y (4.1) (4.7) Blu 5.4 (3.4) (7.1) Br 89.2 (64.9) (15.9) (4.3) Pi 77.5 8.4 5.3 (64.5) (14.8) (4.4) O 97.3 (5.1) (66.7) Pu 4.1 19.6 73.1 (22.0) (9.2) (44.9) Gry 4.3 82.1 (9.7) (60.3) Note. W=White. Bla=Black. R=Red. Gre=Green. Y=Yellow. Blu=Blue. Br=Brown. Pi=Pink. O=Orange. Pu=Purple. Gry=Grey. Table 3 Mapping Task Hit Percentages (Diagonal) and Error distribution according to Model A Predictions (without Brackets) W Bla R Gre Y Blu W 0.0 (76.1) (4.5) (3.3) Bla 0.0 100 (58.3) (9.5) (11.7) R 44.2 (74.4) Gre 100 (87.9) (4.9) (4.7) Y 25.3 20.0 29.6 (6.9) (82.0) Blu 0.0 (7.7) (30.9) (49.5) Br 19.0 6.7 8.6 (9.3) Pi (12.5) O 12.4 9.8 (13.3) (10.9) Pu (11.6) (3.4) (3.9) Gry 100 (5.5) (8.8) Br Pi O Pu Gry W (13.4) Bla (0.4) (4.2) R .9 7.2 26.7 (7.5) 0.0 (8.1) Gre Y 23.9 (4.1) (4.7) Blu (3.4) (7.1) Br 28.5 2.8 5.0 (64.9) (15.9) (4.3) Pi 0.0 (64.5) (14.8) (4.4) O 7.8 24.9 23.6 0.52 (5.1) (66.7) Pu 0.0 (22.0) (9.2) (44.9) Gry 0.0 (9.7) (8.5) (60.3) Note. For comparison, the aged tritanomalous group's empirical results appear in brackets. W= White. Bla=Black. R=Red. Gre=Green. Y=Yellow. Blu=Blue. Br=Brown. Pi=Pink. O=Orange. Pu=Purple. Gry=Grey. Table 4 Proportion of variance ([R.sup.2]) Explained by Models A and B in the Mapping Task A. Tritanomalous aged BCT [R.sup.2] [R.sup.2] [DELTA] Model A Model B [R.sup.2] White -- 0.98 -- Black 0.00 0.94 0.94 Red 0.77 0.97 0.20 Green 1.00 1.00 0.00 Yellow 0.35 0.99 0.64 Blue -- 0.71 -- Brown 0.61 0.93 0.32 Pink -- 0.91 -- Orange 0.32 0.96 0.64 Purple -- 0.75 -- Grey 0.02 0.96 0.94 Global 0.14 0.91 0.77 B. Normal Aged (no aphakic observers) BCT [R.sup.2] [R.sup.2] [DELTA] Model A Model B [R.sup.2] White -- 0.99 -- Black 0.02 0.96 0.94 Red 0.75 0.95 0.20 Green 1.00 1.00 0.00 Yellow 0.33 1.00 0.67 Blue -- 0.98 -- Brown 0.41 1.00 0.59 Pink -- 1.00 -- Orange 0.25 0.99 0.74 Purple -- 0.99 -- Grey 0.01 0.98 -- Global 0.10 0.97 0.87 Note. [DELTA] [R.sup.2] column informs about the increase in this proportion derived from using model B instead of Model A. Table 5 Mapping task Hit Percentages (Diagonal) and Error Distribution according to Model B Predictions W Bla R Gre Y Blu Br W 87.7 Bla 93.8 3.5 R 95,7 Gre 94.4 3.3 Y 98.2 Blu 3.6 3.5 90.0 Br 93.9 Pi O 13.7 9.1 Pu 8.6 3.6 Gry 4.9 Pi O Pu Gry W 3.9 6.9 Bla R 3.1 Gre Y Blu Br Pi 86.2 12.3 O 74.8 Pu 4.5 83.0 Gry 89.2 Note. For comparison, inside brackets appear the aged tritanomalous group empirical results. W=White. Bla=Black. R=Red. Gre=Green. Y=Yellow. Blu=Blue. Br=Brown. Pi=Pink. O=Orange. Pu=Purple. Gry=Grey. Table 6 Hit percentages (Diagonal) and Error Distribution for Normal Aged (without Brackets) and Tritanomalous (in Brackets) Groups in the Best Representative Searching Task W Bla R Gre Y Blu W 100 (80) (8.6) Bla 80 20 (100) R 100 (86.7) Gre 100 (100) Y 100 (100) Blu 100 (19.9) (73.3) Br 6.7 (7.3) Pi (5.7) O (6.8) (14.6) Pu (13.3) (3.3) (7.1) (12.4) Gry 6.7 (15.0) (8.1) Br Pi O Pu Gry W (8.6) Bla R (13.3) Gre Y Blu (6.8) Br 93.3 (80) (5.4) (7.3) Pi 100 (80) (4.4) O 100 (78.6) Pu 100 (5.2) (4.8) (53.3) Gry 93.3 (8.1) (15.1) (53.3) Note. W=White. Bla=Black. R=Red. Gre=Green. Y=Yellow. Blu=Blue. Br=Brown. Pi=Pink. O=Orange. Pu=Purple. Gry=Grey.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||terminos basicos de color; articulo en ingles|
|Author:||Lillo, Julio; Moreira, Humberto; Perez del Tio, Leticia; Alvaro, Leticia; del Carmen Duran, Maria|
|Publication:||Spanish Journal of Psychology|
|Date:||Nov 1, 2012|
|Previous Article:||Effects of mercury chloride (Hg[Cl.sub.2]) on Betta splendens aggressive display.|
|Next Article:||Working Memory Capacity and Individual Differences in the Making of Reinstatement and Elaborative Inferences.|