SEARCH

SEARCH BY CITATION

Keywords:

  • hair color;
  • physical damage;
  • visual attention;
  • age, health;
  • attractiveness

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Female hair color is thought to influence physical attractiveness, and although there is some evidence for this assertion, research has yet not addressed the question if and how physical damaging affects the perception of female hair color. Here we investigate whether people are sensitive (in terms of visual attention and age, health and attractiveness perception) to subtle differences in hair images of natural and colored hair before and after physical damaging. We tracked the eye-gaze of 50 men and 50 women aged 31–50 years whilst they viewed randomized pairs of images of 20 natural and 20 colored hair tresses, each pair displaying the same tress before and after controlled cuticle damage. The hair images were then rated for perceived health, attractiveness, and age. Undamaged versions of natural and colored hair were perceived as significantly younger, healthier, and more attractive than corresponding damaged versions. Visual attention to images of undamaged colored hair was significantly higher compared with their damaged counterparts, while in natural hair, the opposite pattern was found. We argue that the divergence in visual attention to undamaged colored female hair and damaged natural female hair and associated ratings is due to differences in social perception and discuss the source of apparent visual difference between undamaged and damaged hair.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Face research typically controls for the effect of hair, assuming that hair color and / or style may affect judgment. With the aid of modern digital imaging and image manipulation technology, it is relatively simple to remove hair as a variable and present face shape, feature placement, and skin quality only to panellists when asking them to evaluate age, health, and attractiveness. However, this approach is (by its very nature) ignorant of the effect of hair and, thus, captures only a part of the vast natural variation in human appearance. The paucity of literature relating to the social perception of hair points appears to confirm this feature as a largely ignored aspect of human physical appearance.

The few available studies in this area have concentrated almost exclusively on the perceived attractiveness of hair color. For example, Rich and Cash[1] investigated whether blond women are over-represented in print media sampled over a period of some four decades. It was found that the percentage of blondes in magazines (e.g., Ladies’ Home Journal, Vogue, and Playboy) was higher than the incidence of blondes in a White female sample, suggesting that blonde is, somehow, an “ideal” of feminine beauty (although there were some temporal shifts that could not entirely be explained by this data). Frost[2] postulated the existence of a “rare-colour advantage” with regard to female hair. According to this view, those with a hair color of lower incidence in a given population should be perceived as more attractive than those with a hair color with greater incidence. More recently, Sorokowski[3] showed that men judged images of women with (digitally enhanced) blonde hair significantly younger than the same images with brown hair (particularly in women around aged 30). This result is, however, in contrast to the findings of Swami and Furnham,[4] who reported that blondes were rated as less physically attractive and more promiscuous than brunettes, a result that was replicated in two follow-up studies, including an international investigation.[5, 6]

In line with evolutionary psychology theory, Hinsz, Matz and Patience[7] argued that female hair might well signal reproductive potential, because men are attracted to women who are physically young, healthy, and attractive and that well-groomed, “good-looking” hair may signal these precise parameters. Etcoff[8] proposed that hair which is healthy, shiny, and strong signals overall physical health and, conversely, that hair which lacks these signals may be perceived to have been damaged through illness. In light of this, Hinsz et al.[7] hypothesized that age-related characteristics of female hair will cause a woman to be perceived as less attractive because of the direct link between age and reproductive potential. The authors found some support for their assertion, as their data revealed a significant correlation between hair length and age, indicating that younger women have longer hair.[9] In addition, hair quality was correlated with female physical health, leading the authors to conclude that hair length and quality indeed signal aspects of youth and health, indicating reproductive potential.

Hinsz et al.[7] proposed various directions that future research on hair quality and perception could take, to address questions that remained unanswered by their study. One of these questions concerns female hair color and why women feel the need to dye and tint their hair. Most certainly, this question cannot be addressed by a single study. In this present study, however, we embrace the assertion that there has been negligence in the consideration of certain salient physical features in attractiveness research (such as hair[1]), and hypothesize that the quality of women's natural and colored hair affects male visual attention and perception of age, health, and attractiveness. To our knowledge, there has been no research that has considered the effect of condition on the perception of hair (natural or colored) and whether this is a significant cue, which drives the visual attention of others.

The available peer-reviewed literature presents somewhat inconsistent findings with regard to hair color and preference in a social context. Hence, we consider it likely that, regardless of whether a frequency-dependent preference for a certain hair color exists or not, hair condition (independent of color) affects our perception (in terms of social attribution) and attracts visual attention differentially. In the present study, therefore, we set out to investigate this hypothesis by using digital images of female hair tresses in both natural and colored condition, before and after they underwent a controlled degree of physical damage, and studied male visual attention (via eye-tracking technology) and corresponding perception of age, health, and attractiveness. We hypothesized that images of undamaged natural and colored hair would receive higher visual attention and a more positive assessment than their damaged counterparts.

Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Participants

Our participants comprised 50 men and 50 women aged 30–51 years (mean age = 39.3, SD = 7.3), recruited from the local population in Reading (UK). Half of the group (25 men, 25 women; mean age = 40.0, SD = 7.3) participated in the experiment on perception of natural (uncolored) hair and the other half (25 men, 25 women; mean age = 38.5, SD = 7.3) participated in the experiment on perception of images of colored (dyed) hair. The participants were all screened for normal visual acuity, and none wore glasses or contact lenses (self-reported). All participants were asked to complete two tasks with the same stimulus material; (i) an eye-tracking experiment; and (ii) a rating study. They were fully debriefed at the end of the experiment and received a compensation of 15 GBP.

Stimuli

We used four different sets of hair images: natural (uncolored) hair (Set 1), which was subjected to a standardized damage cycle (see below) and then imaged again (Set 2), and natural hair that was first colored (Set 3) and then also subjected to a standardized damage cycle before imaging again (Set 4). Sets 1 and 2 comprised 20 human hair samples (untreated single head, formed into tresses, 22 × 2.5 cm; Fig. 1) that were purchased from three different suppliers (Kerling Int., Backnang, Germany; International Hair Importers & Products Inc., Glendale, NY, USA; Hugo Royer International Ltd., Hampton, UK) to represent a broad range of naturally occurring hair shades. No further modification was made to the hair tresses of Set 1. For Set 2, the dry hair tresses were treated in a multi-stage combing wheel (500 cycles) to simulate the mechanical damage suffered by hair in everyday grooming procedures.[10] The level of damage was monitored and compared with pre-damage values by measuring changes in shine contrast.[11] Sets 3 and 4 originated from untreated, mixed hair samples, which were purchased from one supplier (Kerling Int., Backnang, Germany), and formed into tresses, as for Sets 1 and 2. These hair tresses were treated with commercial colorant products (Koleston Perfect, Wella, and Nice'n Easy, Clairol) matched to the range of starting shade levels. The mixed colorant was thoroughly massaged into the hair (4 g of product per g of hair) and then left for 30 min (Koleston) or 25 min (Nice'n Easy) in a temperature-controlled oven set to 30 °C. The hair was then completely rinsed and the hair shampooed and conditioned as per the colorant product instructions, and finally dried. Set 4 received the same standardized damage cycle as Set 2.

image

Figure 1. Screen-capture of a scene showing the image of hair in natural condition and its damaged counterpart, as presented in the eye-tracking experiment. The rectangles around the images depict the areas of interest, which were defined for all hair pairs presented. A blank image with a fixation cross (as shown) preceded the presentation of each scene.

Download figure to PowerPoint

The four sets of hair tresses were imaged in a custom, enclosed light-box (Verivide, Leicester, UK), using a fluorescent D65 source simulator. A Nikon D1X digital camera fitted with a Micro-Nikkor 105 mm F/2.8 lens was calibrated against a Gretag-MacBeth (now X-Rite, Grand Rapids, MI) Mini Colour Checker chart. The camera was set up in such way that (a) grey balance was maintained across all the grey scale samples and (b) tonal curve, F-Stop and exposure were selected to ensure that white chip color values were in the range R = G = B = 240 ± 2 and black chip color values were in the range R = G = B = 35 ± 2. Full resolution (3008 × 1960 pixels), uncompressed 24-bit TIFF images were recorded and then cropped to size for presentation. Hair tresses were mounted (forced aligned) to ensure that the shine band was prominent and visible in situ using a proprietary curved mount (with a curvature approximating that of an average adult human head).

With these images, two sets of presentation slides were created by combining each image of undamaged hair with its damaged counterpart on one slide (Fig. 1). This produced a set of 20 slides with pairs of natural (Set 1) and natural damaged (Set 2) hair images and another set of 20 slides showing pairs of colored (Set 3) and colored damaged (Set 4) hair images. Undamaged and damaged hair images were randomized and balanced with regard to the side of presentation (left or right). Another two sets of image pairs were then created using the same combinations, only this time the side of presentation was reversed.

Eye-tracking set-up

Eye-tracking methodology was employed for measuring participants’ visual attention to images of undamaged and damaged natural and colored hair when presented together in the form of pairs on a calibrated computer monitor. Eye-tracking was performed using the Tobii T60 system (Tobii Technology, Danderyd, Sweden), a stand-alone binocular pupil-tracking system for recording eye movement. This system consisted of a 17’’ TFT screen at a resolution of 1280 × 1024 pixels with a built-in camera that used near infrared (NIR) light for tracking eye-gaze by detecting the contrast of the pupil against the surrounding sclera (‘Pupil Centre Corneal Reflection’, PCCR). Pupil contrast is increased in this setup because it absorbs NIR wavelengths, whereas the sclera is a highly optically turbid structure, reflecting these wavelengths. Thus, the centre of the pupil can be determined and tracked readily, allowing the mapping of gaze direction at a sampling rate of 60 Hz, providing a tracking resolution of <0.5° and gaze position accuracy of <0.3°.

Raw data from the tracking process were analyzed using tobii studio software (Tobii Technology, Danderyd, Sweden). For each screen showing one image pair, two rectangular Areas of Interest (AOI) were defined, which covered the entire undamaged or damaged hair image, respectively (Fig. 1). Visual attention data for these AOIs can be expressed by several means, although fixation count and cumulative dwell time have been reported as the most useful.[12, 13] Fixation count reflects the number of times an individual's gaze is directed into a particular AOI (in this case, that drawn over the damaged or the undamaged hair image), whereas cumulative dwell time represents the sum of the time in milliseconds (ms) a participant directs his/her gaze into the respective AOI. In the present experiment, a “fixation” was defined as having a minimum dwell time of 100 ms.

Eye tracking procedure

All participants were instructed not to move during the course of the experiment. Participants were not given any specific instruction prior to the experiment, nor were they told how the stimuli varied, as such information could have resulted in perceptual expectation and, thus, altered focal attention.[13, 14] A nine-point-calibration procedure preceded the experiment, where the participant was instructed to focus on a target point, which then automatically moved to different locations on the screen after a minimum fixation of 400 ms. Following calibration, the experiment started without delay with the presentation of the first scene. Participants were assigned either to the natural / natural damaged or the colored / colored damaged sets of hair images and viewed two versions of the 20 image pairs in random presentation (versions differing in the left / right position of the undamaged / damaged hair images). A blank stimulus with a fixation cross in the centre of the screen was presented prior to each stimulus scene to ensure a constant starting position for each participant. Each stimulus scene was presented to the participants for five seconds, thus the total time taken for presentation of all scenes was 3 min and 20 s. For the statistical analysis, differences in mean dwell time and number of fixations between undamaged and damaged hair images were calculated.

Rating study

All participants viewed the same images of hair pairs as in the eye-tracking tracking task, i.e., each pair showing an image of Set 1 and Set 2 (natural / natural damaged) and the same for Set 3 and Set 4 (“colored / colored damaged) on one screen. They were presented on color-corrected TFT monitors (LaCie 324, LaCie Inc, Portland, USA, Paris, France), set to a resolution of 1900 × 1200 pixels at 32-bit (“true color”) color depth at 6500K, a Gamma of 2.2 and constant luminance. The LaCie blue eye calibrator was used to determine this color profile, which was set as the standard profile for all ratings. Hair image pairs were presented sequentially and in blocks on the screen using medialab software (Empirisoft Corp., New York, NY, USA), and the order of appearance within each block was randomized between participants. Within each block, participants were asked to select the hair image, which they considered to be younger, healthier, and more attractive, respectively. Difference values of summarized assessments of undamaged vs. damaged versions of hair were used for the statistical analysis.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

A one-sample Kolmogorov-Smirnov goodness-of-fit test indicated no significant deviation from normal distribution of the visual attention measures and assessment data (natural / natural damaged hair: all < 0.97, all > 0.05; colored /colored damaged hair: all < 1.34, all < 0.05). A one-sample t-test (against the value of zero, i.e., the assumption of no difference between undamaged and damaged hair images) revealed a significant negative deviation for dwell time and number of fixations in natural /natural damaged hair (dwell time: = −3.76, P < 0.001; fixations: = −4.16, < 0.001; both two-tailed), that is, participants gave higher visual attention to damaged natural hair than to the undamaged counterparts. In contrast, for colored /colored damaged hair, there was a significant positive deviation from zero of dwell time (= 3.89, < 0.001, two-tailed), and the number of fixations followed this pattern, but just failed to reach significance (= 2.03, = 0.056, two-tailed). In other words, undamaged colored hair received longer dwell time and a higher number of fixations than the damaged counterparts (Table 1).

Table 1. Mean values (± SD) of visual attention measures for natural and colored hair in undamaged condition and damaged condition
 Dwell timeFixations
UndamagedDamagedUndamagedDamaged
Natural1377.9 (103.3)1547.1 (136.1)4.8 (.4)5.4 (.5)
Colored1596.2 (86.9)1484.3 (83.1)5.1 (.3)4.9 (.4)

With regard to participants’ assessments of hair images, there was a significant positive deviation from zero (i.e., the assumption of no difference between age, health and attractiveness assessments for undamaged and damaged version of hair) between undamaged and damaged versions of hair on all three measures, this being the case for both natural and colored hair (natural hair: age = 3.80, < 0.001; health = 4.39, < 0.001; attractiveness = 4.48, < 0.001; colored hair: age = 2.89, < 0.01; health = 3.39, < 0.01; attractiveness = 3.28, < 0.01). Thus, undamaged versions of images of natural and colored hair were perceived as significantly younger, healthier, and more attractive than the corresponding damaged counterparts.

Considering associations between visual attention and perception measures, we found significant negative correlations of both dwell time and the number of fixations and participants’ age, health and attractiveness assessments of natural hair. That is, natural damaged hair images received significantly higher visual attention than their undamaged counterparts, although they were perceived less positively (dwell time: age = −0.57, health = −0.60, attractiveness = −0.62, all < 0.01, two-tailed; fixations: age = −0.64, < 0.01, two-tailed, health = −0.69, P < 0.001, two-tailed, attractiveness = −0.67, P < 0.001, two-tailed). In colored hair images, there were no significant associations of visual attention measures and participants’ perceptions (all < 0.21, all > 0.05, two-tailed).

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

We predicted that undamaged versions of images of natural and colored hair would receive higher visual attention and a more positive assessment (in terms of age, health and attractiveness) than damaged counterparts. Our data support this hypothesis for colored hair, but not for natural hair, as participants gave higher visual attention to images of undamaged colored hair and to those of damaged natural hair. Furthermore, age, health, and attractiveness assessments showed that higher visual attention for damaged natural hair was significantly negatively associated with age, health, and attractiveness perception, whereas in colored hair, it was not. This indicates that there are two means of receiving attention through hair color / condition, differing in the “quality” of mechanism explained by age, health and attractiveness judgments, which are clearly in favor of undamaged colored hair (“good attention”) and damaged natural hair (“bad attention”). To our knowledge, this is the first study that links visual attention to isolated images of hair (and, thus, independent of hair style) with key attributes in human social perception (age, health, and attractiveness), as argued by the evolutionary psychology framework.

Evolutionary psychology studies of human attractiveness have shown that certain physical features of the face and body are associated with perception of age and health, and that people tend to judge those features as attractive when expressed by young and healthy individuals.[15, 16] According to the evolutionary psychology logic, this is particularly so for women, given that age and health are strongly linked to female fertility. While evidence for this assertion is relatively strong in the case of faces and bodies, it is rather scarce with regard to hair in this context. Studies that have investigated the social perception of hair have focused almost exclusively on the hypothesis of a bias toward the representation preference of blond women, arguing that hair color present at relatively low frequency in a population is perceived as more attractive than that present at a higher frequency.[1, 2] In the present study, we did not consider the “rare advantage hypothesis”, but rather studied the proximate effect of people's ability to discern relatively subtle visual cues of hair damage from undamaged hair.

In explaining the social relevance of the observations in this current study, it is important to consider visual attention measures in combination with corresponding preference data. Our data suggest that high visual attention can be driven by either undamaged colored hair or damaged natural hair, but also that attributions to either (in terms of age, health, and attractiveness) differ significantly. Damaged natural hair was strongly negatively linked to age, health, and attractiveness assessments, while we did not find a significant association of visual attention measures with these attributions in colored hair.

Our experimental setup based on perception of cropped images of hair tresses taken in a custom, enclosed light-box, negates the possibility that hairstyle or length could be the cause of the effects noted, as reported in other studies.[7, 9] We argue instead that the standardized physical damage cycle affected hair optical properties and, thus, hair appearance, and that these subtle visual cues are both noticeable to the human eye and affect judgments of age, health, and attractiveness. The optical properties of hair are fairly well characterized from older work on fibre optics,[17] the pioneering work of Stamm et al.,[18, 19] through to a variety of modern-day investigations.[20-22] Light interacting with hair fibres creates complex visible patterns through reflection (the “shine band”), absorption (the “chroma band”), and scattering (“diffused light”), and, in combination, they represent the very specific visual appearance of human hair.[23]

The specular reflection of hair, resulting from illumination with incidental light, consists of a series of discrete micro-reflections of individual hair fibres that correspond to their structural elements. The predominant structural elements affecting light transport are layers of shingle-like, over-lapping cuticle cells which, in healthy and undamaged hair, lie flat and aligned on the hair shaft surface, presenting a regular, repeating topography and a corresponding tight, bright shine band.[21] If, however, the cuticle is damaged physically by either frictional forces (for example, brushing or combing) or chemistry (for example, high pH oxidation such as might be experienced in a session of hair coloration), irregularity introduced to both the shape and orientation of the cuticle results in a significant increase in the ratio of diffuse to specular reflection. In simple terms, apparent shine is reduced which, in consequence, may be perceived as less healthy.[24] In this present study, we used a relatively low number of cycles of a comb-wheel to create a tightly controlled degree of physical damage and a corresponding modest increase in diffuse reflection. Interestingly, from our data, it seems evident that people are sensitive to these subtle variations in the optical properties of hair, given that they were able to discern between damaged and undamaged hair just through visual inspection of two-dimensional digital images of hair swatches on a computer screen. It is clear, therefore, that limiting surface damage in natural (uncolored) hair (e.g., by training in combing / brushing techniques, limiting the use of heated appliances, etc.) and in colored hair (e.g., as before and, additionally, by reducing oxidative stress to the cuticle structure) has significant, measureable benefit in driving perception of youth, health, and attractiveness in the owner.

In conclusion, it seems that: (i) hair with minimal surface damage is perceived as younger, healthier, and more attractive than hair with a modest level of controlled physical surface damage (whether natural or colored) and that; (ii) there are different ways to receive attention to one's hair - “good attention” (via undamaged colored hair) and “bad attention” (via damaged natural hair), with consequences relating to key attributes in human social perception. What could be responsible for this apparent differential perceptual response to surface damage in natural and colored hair? From our data, damage to the surface of natural hair results in greater visual attention, although the reason for this directed visual attention is not clear. Why, however, is the same phenomenon not observed for the colored hair? It is known that, generally, the penetration of colorant chromophores with a high extinction coefficient into the hair shaft results in a net reduction in diffuse reflection and a consequent increase in apparent shine/luster.[25] To a certain extent, therefore, coloring may be able to “offset” the increase in diffuse reflection due to surface damage, making more prominent the attractiveness “reward”[26] presented by the perceived health in the shine band of undamaged hair.

The experimental setup used this study is, of course, not present in nature (i.e., people viewing undamaged and damaged versions of the same hair at the same time). However, it nevertheless exploits a perceptual mechanism that applies to a real-life situation, in which people varying in hair coloration (natural vs. colored) and condition (undamaged vs. damaged) interact with one another. Our data suggest that well-maintained natural and colored hair in particular has a positive effect on how people perceive each other in terms of age, health, and attractiveness. The strength of the effect of hair coloration/condition in relation to hairstyle and facial morphology presents intriguing questions to be addressed in future studies.

Acknowledgments

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

This project was funded by Procter & Gamble Company, Cincinnati, USA and preparation of the article was supported by a grant of the German Science Foundation (DFG), number FI 1450/4-1, awarded to Bernhard Fink.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  • 1
    Rich MK, Cash TF. The American image of beauty – media representations of hair color for four decades. Sex Roles 1993; 29: 11324.
  • 2
    Frost P. European hair and eye color - A case of frequency-dependent sexual selection? Evol Hum Behav 2006; 27: 85103.
  • 3
    Sorokowski P. Attractiveness of blonde women in evolutionary perspective: studies with two Polish samples. Percept Mot Skills 2008; 106: 73744.
  • 4
    Swami V, Furnham A. Unattractive, promiscuous, and heavy drinkers: perceptions of women with tattoos. Body Image 2007; 4: 34352.
  • 5
    Swami V, Furnham A, Joshi K. The influence of skin tone, hair length and hair colour on ratings of women's physical attractiveness, health and fertility. Scand J Psychol 2008; 49: 42937.
  • 6
    Swami V, Rozmus-Wrzesinska M, Voracek M et al. The influence of skin tone, body weight and hair colour on perceptions of women's attractiveness, health and fertility: a cross-cultural investigation. J Evol Psychol 2008; 6: 32141.
  • 7
    Hinsz VB, Matz DC, Patience RA. Does women's hair signal reproductive potential? J Exp Soc Psychol 2001; 37: 16672.
  • 8
    Etcoff N. Survival of the prettiest. New York: Doubleday, 1999.
  • 9
    Mesko N, Bereczkei T. Hairstyle as an adaptive means of displaying phenotypic quality. Hum Nat 2004; 15: 25170.
  • 10
    Haake H-M, Marten S, Seipel W, Eisfeld W. Hair breakage - How to measure and counteract. J Cosmet Sci 2009; 60: 14351.
  • 11
    Sendelbach G, Haller M, Springob C, Schmidt H. Hair lustre measurements via image analysis and its use for cosmetically altered hair. Proceeding of the 12th International Hair Science Symposium of the DWI. Germany: Heidelberg; 2001.
  • 12
    Duchowski AT. A breadth-first survey of eye-tracking applications. Behav Res Methods 2002; 34: 45570.
  • 13
    Duchowski AT. Eye tracking methodology: theory and practice. London: Springer Verlag; 2003.
  • 14
    Yarbus AL. Eye movements and vision. New York, NY: Plenum Press; 1967.
  • 15
    Fink B, Penton-Voak IS. Evolutionary Psychology of Facial Attractiveness. Curr Dir Psychol Sci 2002; 11: 1548.
    Direct Link:
  • 16
    Grammer K, Fink B, Møller AP, Thornhill R. Darwinian Aesthetics: sexual Selection and the Biology of Beauty. Biol Rev Camb Philos Soc 2003; 78: 385407.
  • 17
    Lundberg JL. Light scattering from large fibers at normal incidence. J Colloid Interface Sci 1969; 29: 56585.
  • 18
    Stamm RF, Garcia ML, Fuchs JJ. The optical properties of human hair I. Fundamental considerations and goniophotometer curves. J Soc Cosmet Chem 1977a; 2: 57199.
  • 19
    Stamm RF, Garcia ML, Fuchs JJ. The optical properties of human hair II. The luster of hair fibres. J Soc Cosmet Chem 1977b; 28: 6019.
  • 20
    Kharin A, Varghese B, Verhagen R, Uzunbajakava N. Optical properties of the medulla and the cortex of human scalp hair. J Biomed Opt 2009; 14: 2435.
  • 21
    McMullen R, Jachowicz J. Optical properties of hair: effect of treatments on luster as quantified by image analysis. J Cosmet Sci 2003; 54: 33551.
  • 22
    McMullen R, Jachowicz J. Optical properties of hair - detailed examination of specular reflection patterns in various hair types. J Cosmet Sci 2004; 55: 2947.
  • 23
    Lefaudeux N, Lechocinski N, Clemenceau P, Breugnot S. New luster formula for the characterization of hair tresses using polarization imaging. J Cosmet Sci 2009; 60: 15369.
  • 24
    Sinclair RD. Healthy hair: what is it? J Investig Dermatol Symp Proc 2007; 12: 25.
  • 25
    Keis K, Ramprasad KR, Kamath YK. Effect of hair colour on luster. J Cosmet Sci 2004; 55: 42336.
  • 26
    Aharon I, Etcoff N, Ariely D et al. Beautiful faces have variable reward value: fMRI and behavioral evidence. Neuron 2001; 32: 53751.