Eye-Color – the New Phrenology? - PowerPoint Presentation

Slide1

Eye-Color – the New Phrenology?

Taylan S. Ergun, Alla Chavarga, Daniel J. Chi, Stavros P. Hadjisolomou, Kamil Kloskowski, Israel AbramovApplied Vision Institute, Psychology Dept., Brooklyn College/CUNY

We derived “albedo” by summing the reflectance from 450 to 660 nm. Figure 2 shows two individuals’ spectral reflectances, which were the highest and the lowest spectral reflectances between 450 to 660nm. In addition, average reflectances of all individuals is shown in the same graph (Fig. 2). Frequency distribution of albedos indicated a multimodal distribution (Fig. 3). We convert the CIE xyz coordinates to CIE u’ v’ coordinates in order to plot the data points in the CIE 1976 2° Chromaticity Diagram (Fig .4). Figure 5 shows that there are clearly 2 clusters: One follows the black body locus and the other cluster is separate. The larger cluster is comprised mostly of females and the smaller cluster has equally both males and females (Fig 6).In order to represent an individual eye color as a single wavelength, we calculate all individuals’ dominant wavelengths (Fig. 7). The dominant wavelength of a data point is given by the intersection of the ray from Equal Energy White (EE) to the spectrum locus. Dominant wavelength is the single wavelength that, when mixed with Equal Energy White (illuminant that weights all wavelengths equally), is perceived by the observer as identical to the actual reflectance spectrum of the object. The red line is an example of deriving dominant wavelength for one participant; all the other lines pass through the data points of all the other participants. Figure 8a shows frequency of dominant wavelength for all participants. Figure 8b breaks this down by sex. The dominant wavelengths of eye colors are mostly between 580 and 600 nm, but with a small handful between 530 and 580 nm; there is a smaller group with dominant wavelengths less than 490 nm – these radiate towards the spectrum locus in the opposite direction from all the other data points (cf. Fig. 7).Figure 9 shows the scatter plot of albedo vs dominant wavelength. Group 1 (dominant wavelengths > 530 nm) shows a clear correlation between these two iris measures. Group 2 (dominant wavelengths < 490 nm) is also characterized by much lower albedo. We also examined the relationship between iris albedo and albedo of the skin (dermis). Figure 10 shows that there is a moderate positive correlation between iris albedo vs dermis albedo.

We observed 2 clusters on the CIE u’ v’ chromaticity diagram (Fig. 6). The largest cluster (almost all female) followed the chromaticity distribution of black-body radiators; the second cluster is separate, smaller and does not follow black body locus and is comprised of both males and females. The location of a subset of our data points relative to the Black Body locus is surprising and unexplained. The relation between dominant wavelength and albedo (Fig. 9) agrees, more or less, with casual observations that low albedos are associated with eyes that might be termed “brownish,” and those that might be labelled “greenish” have higher albedos. We note that eyes that might correspond to “bluish” also have the highest albedos – it seems there are no eyes that are blue and dark!In this study we used an objective method to measure eye color. Contrary to other studies, we showed that eye color can be measured in a continuous scale instead of labeling them with ordinary color names (green, brown, blue, hazel). There appears to be no clear categorization strategy by color appearance alone. Therefore, psychologists should tread lightly when adopting such a measure. Whereas traditional strategies yield a typical binary (light/dark) or commonly-used five-color system (blue/green/brown/black/hazel), categorization by appearance in a standard color space yields 2 categories not necessarily parallel to previously used scales. Sex differences may exist according to our analyses. The larger cluster in our data set is comprised of females, while the smaller is mixed. We know from previous analyses that sex differences in the visual system exist, both in spatial resolution and color vision (Abramov, Gordon, Feldman, and Chavarga 2012a, b), and are possibly due to the influence of testosterone. It is possible that pre-retinal absorption is similarly influenced by way of sex differences in quantity of melanin, one of the main contributors to phenotypic iris color. Melanin quantity is typically set at about 12 months of age, before which, iris color appearance has been observed to change (Wright, K. W., 1995).Due to its heritability, eye color prevalence is largely determined by geographic location (by way of genetic ancestry). For instance, diversity is greatest in Northern and Eastern Europe, so any observed differences in rates of eye color must consider the representativeness of the subsamples used.

Discussion

Introduction

Eye color varies from person to person. Numerous papers have related an individual’s eye-color to a variety of psychological traits (Table 1). One of the reasons eye color has been considered to be a useful biomarker is its high heritability (98%), and the fact that eye color does not change much over time (

Bito et al., 1997). Contrary to popular belief, eye color is a non-Mendelian trait; it is determined by several genes, each of which determines production of different types of melanin (the pigment that colors the eye). Categorizations according to eye color may be useful particularly in biological studies. For example, researchers have suggested that melanin production in the eye may be associated with norepinephrine and cortisol production, which are biological indices of behavioral inhibition (Kagan et al., 1988).However, many studies use their own arbitrary system of categorizing different eye-colors. There appears to be no widespread, standard, or objective method of measuring eye color, specifically in a way which relates the category to possible melanin quantity or density. In this study, we investigate whether there is an objective method of measuring eye color. We use measures of iris spectral reflectance and translate their distribution in a well-known color space to determine whether they correspond to the nominal judgments of eye color typically observed.Color space is a specific organization of colors. There are several different color spaces for example: RGB (red, green, blue) color space generally used in TV and computer vision applications, and shows the color of light (red, blue, green, yellow etc.), but not its intensity. We use CIE 1976 u’v’ chromaticity space in order to represent our spectral measurements of iris. The first defined color space is The CIE 1931 xyz color space which encompasses all the colors that an average person can experience. CIE 1976 u’v’ is a modification of CIE 1931 xyz to display color differences more accurately by making error zones more uniform. CIE chromaticity diagrams also include Black Body Locus (Planckian Locus) which is the path that the color of black body takes as black body temperature changes. A Black Body spectrum can be used to define the relative color temperature of a light – i.e., the effectiveness of the light source for the human visual system.

Results

Abramov, I., Gordon, J., Feldman, O., &Chavarga, A. (2012). Sex & vision I: spatio-temporal resolution. Biology of Sex Differences, 3(1), 20.Abramov, I., Gordon, J., Feldman, O., &Chavarga, A. (2012). Sex and vision II: color appearance of monochromatic lights. Biology of Sex Differences, 3(1), 21.Bassett, J. F., & Jr., J. M. (2001). Eye color predicts alcohol use in two archival samples. Personality and Individual Differences, 31(4), 535-539. Belfer, I., (2014) Proceedings from APS ’14: 33rd Annual Scientific Meeting of the American Pain SocietyBito, L. Z. (1997). Eye Color Changes Past Early Childhood. Archives of Ophthalmology Arch Ophthalmol, 115(5), 659. Coplan, R. J., Coleman, B., & Rubin, K. H. (1998). Shyness and little boy blue: Iris pigmentation, gender, and social wariness in preschoolers. Dev. Psychobiol. Developmental Psychobiology, 32(1), 37-44.Gardiner, E., & Jackson, C. J. (2010). Eye color Predicts Disagreeableness in North Europeans: Support in Favor of Frost (2006). Current Psychology Curr Psychol, 29(1), 1-9. Gary, A. L. & Glover, J. (1976). Eye color, Sex and Children’s Behavior. Chicago: Nelson-Hall.Gentry, T. A., Polzine, K. M., Wakefield, J. A., 1985. Human genetic markers associated with variation in intellectual abilities and personality. Personality and Individual Differences 6, 111-113. Kagan, J., Reznick, J. S., Snidman, N., Gibbons, J., & Johnson, M. O. (1988). Childhood Derivatives of Inhibition and Lack of Inhibition to the Unfamiliar. Child Development, 59(6), 1580. Markle, A. (1976). Eye Color and Responsiveness To Arousing Stimuli. Perceptual and Motor Skills, 43(1), 127-133. Wright, K. W. (1995). Pediatric ophthalmology and strabismus. St. Louis, MO: Mosby.

References

Numerous papers have related an individual’s eye-color to a variety of psychological traits – e.g., agreeableness, competitiveness, and tendency towards alcoholism (Gardiner & Jackson, 2010; Bassett & Dabbs Jr, 2001). But there is no accepted method of measuring eye-color – each study uses its own arbitrary system of categorizing the different eye-colors. We asked whether there is a clear series of objective categories. Using a scanning

spectroradiometer we measured the spectral reflectance of each participant’s iris across the entire visible spectrum. From this we derive\d: (i) albedo by summing a participant’s reflectances across the spectrum; (ii) color by computing the co-ordinate values for each iris in a standard chromaticity space (CIE 1976, u’ v’). A frequency distribution of albedos indicated a multimodal distribution. On the chromaticity diagram we found three clusters: the largest cluster (almost all female) followed the chromaticity distribution of black-body radiators; a distinct, separate, smaller cluster was equally male and female; the smallest cluster also follows the black-body locus, but is separate from the largest cluster. Our clusters do not correspond to the many categorical descriptions of eye color, such as brown, blue, gray, green, hazel, black (Psychologists beware!).

Abstract

Table 1

Fig. 1

Fig. 2

Fig. 3

Fig. 4

Participants were 14 male and 38 female. Sex was measured as a self-report. All participants had healthy eyes and normal vision, including stereopsis, motion detection, Vernier acuity,

spatio

-temporal contrast sensitivity, and color discrimination

.

We measured the spectral reflectance of each participant’s iris across the entire visible spectrum. This measurement gave us percent of light reflected from iris per wavelength and coordinates of individual’s eye color on CIE 1931 xyz color space. A scanning

spectroradiometer

(Photo Research; model 703A/PC) was used to measure reflectance of a representative area of iris (Fig. 1). In addition, we measured spectral reflectance of skin. Skin (dermis) measurement taken from the inner surface of forearm. This area chosen because is less likely to be heavily suntanned. The instrument measures spectral radiance -- i.e., the energy, per wavelength, that reaches its detector after reflection by the iris; these values can be displayed graphically and converted to assorted standard measures of light. Radiances were measured under standardized conditions: Chin rest + two fluorescent illuminators at either side of head -- measure area of iris that is free of corneal reflections (Fig. 1). Also, measured radiance of neutral white reflector under same conditions; used this to compute

reflectances

of subject’s iris.

Methods

Fig.

2

Fig. 4

Fig. 5

Fig.

6

Fig. 7

Fig. 8a

Fig. 8b

Fig. 10

Fig.

9