Volume 116, Issue 2, Pages 355-360 (February 2009)

30 of 39

ABSTRACT

FULL TEXT

FULL-TEXT PDF (1038 KB)

Age- and Fatigue-related Markers of Human Faces: An Eye-Tracking Study

Presented at: the American Society of Ophthalmic Plastic and Reconstructive Surgery Annual Meeting, November 2006.

Received 10 December 2007; received in revised form 7 October 2008; accepted 7 October 2008. published online 12 December 2008.

Purpose

To investigate the facial cues that are used when making judgments about how old or tired a face appears.

Design

Experimental study.

Participants

Forty-seven subjects: 15 male and 32 female participants, ranging from age 18 to 30 years.

Methods

Forty-eight full-face digital images of “normal-appearing” patients were collected and uploaded to an eye-tracking system. We used an Applied Science Laboratories (Bedford, MA) Eye Tracker device associated with gaze-tracking software to record and calculate the gaze and fixation of the participants' left eye as they viewed images on a computer screen. After seeing each picture, participants were asked to assess the age of the face in the picture by making a selection on a rating scale divided into 5-year intervals; for fatigue judgments we used a rating scale from 1 (not tired) to 7 (most tired).

Main Outcome Measures

The main outcome measure was gaze fixation, as assessed by tracking the eye movements of participants as they viewed full-face digital pictures.

Results

For fatigue judgments, participants spent the most time looking at the eye region (31.81%), then the forehead and the nose regions (14.99% and 14.12%, respectively); in the eye region, participants looked most at the brows (13.1%) and lower lids (9.4%). Participants spent more time looking at the cheeks on faces they rated as least tired than they did on those they rated as most tired (t = 2.079, P<0.05). For age judgments, the eye region (27.22%) and then the forehead (15.71%) and the nose (14.30%) had the highest frequencies of interest; in the eye region, the brows and lower lids also had the highest frequencies of interest (11.40% and 8.90%, respectively). Participants looked more at the brows (t = −2.63, P<0.05) and glabella (t = −3.28, P<0.01) in those faces they rated as looking the oldest.

Conclusions

This study supports the hypothesis that age and fatigue judgments are related to preferential attention toward the eye region. Consequently, these results suggest that aesthetic or functional surgery to the eye region may be one of the most effective interventions in enhancing the appearance of an individual.

Financial Disclosure(s)

The author(s) have no proprietary or commercial interest in any materials discussed in this article.

Available online: December 12, 2008.

Article Outline

• Abstract

• Materials and Methods

• Participants and Main Outcome Measures

• Face Stimuli

• Procedure and Apparatus

• Section 1

• Section 2

• Results

• Discussion

• References

• Copyright

The desire to look younger, less tired, and more appealing are patients' primary motivators for undergoing cosmetic surgery.1 Although some research has been conducted to explain why some people are perceived as old or tired, far more has focused on the positive effects of appearing attractive.2, 3 For example, social psychology research has documented that a youthful and attractive face exerts a strong effect on impressions, making them more positive.4 Attractive individuals have also been found to have advantages in employment settings, to be more likely to be acquitted for a crime, and to be treated less harshly by teachers and peers.5, 6 Furthermore, it has been assumed that positive change in physical appearance for patients will lead to an improvement in their psychologic well-being, including their self-confidence and self-esteem.7, 8 Research has revealed that surgery to enhance appearance produces a positive psychologic effect by improving quality-of-life outcomes; surgical patients reported positive changes in their social lives and interpersonal relationships after surgery.9 Thus, the goal of this research was to investigate the facial cues that are used when making judgments about how old or tired a face appears through an innovative technique: eye tracking.

Eye tracking is a noninvasive method allowing for the measurement of gaze patterns in nearly real-time and has been used in many previous studies.10, 11, 12 For example, eye tracking has been used to investigate emotional expression on faces.12 Eye tracking therefore provides a method for understanding how perceivers process stimuli such as human faces. According to such work, it seems that when making judgments about emotions expressed on the face, perceivers look most at the eyebrows, followed by the mouth and the eyes.13 On the basis of these findings, the current study was designed to use the eye-tracking methodology to investigate where perceivers look when making judgments about the age and fatigue of a face. We hypothesized that the eye region would be the area of the face most used, as indicated by greater fixation, when making age and fatigue judgments. We focused on these judgments because we assumed they would be most related to decisions to undergo plastic surgery.

Materials and Methods

This study obtained approval by the relevant institutional review board.

Participants and Main Outcome Measures

Forty-seven young adults were included in this study: 15 male (31.9%) and 32 female participants (68.1%), ranging from age 18 to 30 years (M = 21.26; standard deviation [SD] = 3.17). Participants were recruited from students in the Boston area and received a monetary stipend or 1 class credit for their participation. All participants signed an informed consent form on their arrival.

Face Stimuli

Forty-eight full-face digital images of “normal appearing” (no major facial lesions or deformities) patients were collected from the Eye Plastic, Orbital, and Aesthetic Surgery image database. Individuals appearing in the pictures consented to the use of their images in research settings. Physicians who took the pictures were well trained so that all the pictures fulfill the same criteria: controlled studio lighting (non-flash), full-face, frontal view, neutral facial expression, limited from the neck to the top of the head, and consistent colored background (blue).

These pictures were then uploaded to the eye-tracking system, so they could be presented to participants as the participants' eyes were tracked. A LookZone, or subunit region of interest, permits the examination of gaze within a prespecified stimulus region. For the current study, several facial LookZones were created by drawing lines that divided the faces of the pictures into smaller areas (i.e., foreheads, eye brows, upper eyelids, lower eyelids, pupils, noses, cheeks, upper lips, lower lips, and chins). Thus, the eye-tracking system could capture and statistically calculate the eye gazes and fixation time within each of these predefined LookZones.

Procedure and Apparatus

The experiment used an Applied Science Laboratories (Bedford, MA) Eye Tracker 504 with Magnetic Head Transmitter device. The eye tracker records the duration and location of the participants' left eye 60 times per second. Visual fixation is defined as a period in which a participant focuses his or her gaze within 1 degree of visual angle on a location for ≥100 ms, within predesignated LookZones. Thus, the total viewing time and percent fixation time within a LookZone can be calculated. Because the ASL 504 calculates gaze on the basis of analysis of the infrared reflection from the pupil and cornea, the technique is noninvasive. Gaze data were recorded while participants looked on the computer screen and stored on the computer hard drive through the GazeTracker software (Eye Response Technologies, Charlottesville, VA), which recorded stimuli, synchronized the data it captured, and was used to analyze these data.

At the start of the experiment, participants were asked to report demographic information, including age, gender, ethnicity, years of education, and marital status. Participants were not asked to report their name. All participants had their vision tested before having their eyes tracked to ensure their vision was adequate to track. Vision was tested using 2 standard measures: a Snellen test of visual acuity and a Pelli-Robson test of contrast sensitivity.

After completing the questionnaires, participants were led to the eye tracker room, seated in a chair in front of the screen, and asked to look at the 20-inch computer screen in front of them. Their seating was adjusted so that their eyes were level with the middle of the computer monitor. The ASL 504 camera rested on a holder below the presentation monitor. The eye tracker was directed toward their left pupil, and a brief (3–5 minutes) calibration was run in which they were asked to look at 17 points on the screen. The purpose of this calibration was to ensure that the tracker was recording within 1 degree visual angle of the appropriate point. After the system was well calibrated, the trials began (Fig 1). Each session was divided into 2 sections.

View Large Image

Figure 1. Eye-tracking process: Experimenter (A) controls the process through 3 monitors (B); participant (C) is sitting in front of the computer screen of the eye-tracker system (D).

Section 1

Participants received the following instructions. “You will see a series of pictures of older individuals that will be shown one by one on the computer screen. Please look at them carefully. Following each picture, you are asked to guess how old they are and then respond by clicking on the scale that will be shown after each picture.” Participants were shown images of 48 older individuals as described previously. Each slide appeared on the screen for 5 seconds. All the photos were shown in the same order to all participants. Following each picture slide, a rating scale was presented on the screen. There was no time limit for participants to make their ratings; the system was set to advance to the next photo whenever the participant clicked the mouse on the rating scale. The scale is shown as in Table 1. At the end of Section 1, participants were asked to relax for a few minutes and then proceeded to Section 2.

Table 1.

Age Rating Scale


Age	Age	Age	Age	Age	Age	Age	Age	Age	Age
<40 y	41–45 y	46–50 y	51–55 y	56–60 y	61–65 y	66–70 y	71–75 y	76–80 y	>80 y

Section 2

Participants were shown the same images as in Section 1 but in a different order than they were in Section 1. However, this new order was also the same for all participants. As in Section 1, each slide contained 1 image and stayed on the screen for 5 seconds, followed by a rating scale. The instructions to participants in this setting were as follows: “You will see a series of pictures of individuals that will be shown one by one on the computer screen. Please look at them carefully and then rate how tired each individual appears by clicking on the scale that will be shown after each picture.” As in Section 1, the system showed the next photo whenever the participant reacted by clicking the mouse on the rating scale. The scale is shown as in Table 2. The direction of participants' eye gaze and their responses to the rating scale were recorded and analyzed by gaze tracker software (Fig 2).

Table 2.

Fatigue Rating Scale


1	2	3	4	5	6	7
Not tired			Somewhat tired			Most tired

View Large Image

Figure 2. Full-face image stimulus with predefined LookZones and gaze points.

Following the experimental session, participants were fully debriefed about the purpose of the study, and the experimenter answered any questions regarding the meaning and procedure of the study.

Results

For fatigue judgments, participants spent the most time looking at the eye region (M = 31.81%, SD = 15.31%), then the forehead (M = 14.99%, SD = 19.02%) and the nose (M = 14.13%, SD = 9.46%); specifically within the eye region, participants looked most at the brows (M = 13.10%, SD = 12.30%) and lower lids (9.40%, SD = 9.56%). For age judgments, participants again looked most at the eye region (M = 27.22%, SD = 10.90%), then less often at the forehead (M = 15.71%, SD = 16.07%) and the nose (M = 14.30%, SD = 9.43%); in the eye region, the brows (M = 11.40%, SD = 8.85%) and lower lids (M = 8.90%, SD = 9.18%) also had the highest frequencies of interest. Thus, for both age and fatigue judgments we noted that the eye region always had the highest percentage of fixation (Fig 3).

View Large Image

Figure 3. Percentage of fixation times on LookZones for age rating and fatigue rating.

To determine whether LookZones differentiated faces rated by participants as older versus younger, a paired-samples t test was used to analyze the data from 2 groups: 24 stimuli that had been rated as younger by participants (age <50 years) versus the other 24 stimuli that had been rated as older (age ≥51 years) (according to participants' age rating, we regrouped the stimuli into 2 small subgroups: 24 stimuli that had the “younger age” and 24 stimuli that had the “older age”; the age of each stimulus had been determined by calculating the mean of total age rating from all participants who rated that stimulus. When we used this methodology, the value of age 50 years appeared as the median that divided the stimuli into 2 equal subgroups. The results showed that there was no significant difference between the 2 groups at the eye region; however, participants looked more at the brows (t = −2.63, P<0.05) and glabella (t = −3.28, P<0.01) in those faces they rated as looking oldest (Table 3). The same analysis was used for the fatigue judgments with 24 stimuli rated ≤3.43 and 24 other stimuli rated ≥3.47 by participants; in this case again there was no significant difference between the 2 groups at the eye region, but participants spent more time looking at the cheek regions on faces they rated as least tired than they did on those they rated as most tired (t = 2.08, P<0.05) (Table 4).

Table 3.

Paired Samples Test for Age Rating


Pairs	t	df
Eye	−0.17	46
Nose	1.85	46
Upper lip	4.23	46
Glabella	−3.28†	46
Forehead	−0.79	46
Cheek	4.92	46
Lower lip	−0.18	46
Pupil	4.43	46
Eyebrow	−2.63⁎	46

The eye region includes the eyelids, eyebrows, and pupils.

⁎

P<0.05.

†

P<0.01.

Table 4.

Paired Samples Test for Fatigue Rating


Pairs	t	df
Eye	−1.70	46
Nose	1.69	46
Upper lip	−0.13	46
Glabella	−0.24	46
Forehead	−0.41	46
Cheek	2.08⁎	46
Lower lip	0.46	46
Pupil	−1.10	46
Eyebrow	−1.01	46

The eye region includes the eyelids, eyebrows, and pupils.

⁎

P<0.05.

For the age rating, the LookZones that attracted gaze when each image first appeared was most often the eye region (46%), then the nose (19.2%), the forehead (13.3%), and the glabella (10.6%) (these numbers reflected the percentage of first fixations that took place within that specific LookZone). For fatigue ratings, the eye region was also the most frequent first gaze point (44.7%), followed by the nose (18%), the forehead (13.7%), and the glabella (12.3%).

To understand further the relative importance of the eye region (brows, upper/lower lids, and interpalpebral zone), we calculated the pixel count within the face and the eye region and their relative areas (using Image J software -http://rsb.info.nih.gov/ij/; accessed July 1, 2006). Results showed that the eye region represented 21% of the face. Although the eye region filled only a small percentage of the face, it attracted a large percentage of gaze fixation. A 1-sample t test where the test value was the actual amount of facial area taken up the by the eyes was conducted. Results showed that there was a significant difference (t = 3.91, P<0.01) between the percentage of fixation in the eye region and the test value, meaning that the eye region received higher visual attention than would be expected based simply on its relative area on the facial images.

Correlations between gaze fixations for age and fatigue ratings were also calculated, as shown in Table 5. There was a statistically significant (P<0.01) positive correlation coefficient for the association between age rating and fatigue rating for the following LookZones: eye (r = 0.42), nose (r = 0.65), glabella (r = 0.43), forehead (r = 0.61), brows (r = 0.42), and pupils (r = 0.50). This finding signals that there is a strong relationship between how facial regions are used for making judgments about age and fatigue.

Table 5.

Correlations between LookZones of Age Rating and Fatigue Rating


Age Rating	Fatigue Rating
Age Rating	Nose	Glabella	Forehead	Cheek	Eye	Pupil	Eyebrow
Nose	0.65†	−0.29	−0.23	0.15	−0.43	−0.06	−0.49
Glabella	0.15	0.43†	0.20	−0.26	−0.19	−0.25	0.089
Forehead	−0.25	0.16	0.61†	−0.37	−.08	−0.33	0.45
Cheek	−0.06	0.00	0.39	0.33⁎	0.14	0.39	−0.28
Eye	−0.32	−0.19	−0.18	0.08	0.42†	0.41	0.04
Pupil	−0.12	−0.17	−0.33	0.26	0.27	0.45†	−0.20
Eyebrow	−0.38	0.02	0.43	−0.38	0.09	−0.15	0.42†

⁎

Correlation is significant at the 0.05 level (2-tailed).

†

Correlation is significant at the 0.01 level (2-tailed).

Discussion

The results of this study supported the principal hypothesis that age and fatigue judgments are related to preferential attention toward the eye region. Both when making judgments about how old and how tired a face appeared, participants spent the most time looking at the eye region. This region was also the area that people focused on “at first sight” for both judgments, which means people tend to use the eye region as a first criteria for their judgments. The brow and the lower lid (including the fat pads of the lower lid) were the most looked at parts of the eye region. Furthermore, a statistically significant (P<0.01) positive correlation between the age rating and the fatigue rating for almost all the predefined LookZones on the face, including the eye zone, indicated that regions were used similar ways for making both types of judgments. We also noted that for age rating, people looked at the brows and glabella in those faces they rated as looking oldest; however, for fatigue rating, they spent more time looking at the cheek region on faces they rated as the least tired than they did on those they rated as most tired. Further research with a larger population is needed to confirm these findings.

Our study also revealed that visual attention toward the eye region was at a significantly higher frequency than would be expected on the basis of the actual facial area taken up by the eye region (P<0.01), further supporting an account in which the eyes are a disproportionately important facial cue when perceivers make age and fatigue judgments.

This is not the first study to investigate the sources of age judgments. For example, Henss14 conducted an experiment to estimate the age of adult men and women on the basis of color head-and-shoulder photographs: Participants were asked to regroup pictures into age groups. Results showed an almost perfect agreement between different groups regarding estimates of the ages of men and women who ranged from their mid 20s to their late 60s.14 The difference between such previous studies and our study is that we have used a real-time measurement of attention to pinpoint precisely where individuals focus to make judgments of facial expressions. Eye tracking is a safe, noninvasive method that can record in nearly real-time eye gaze; this method has been used widely in psychology research and may be expanded to future medical research as well.

Limitations of this study include that judgments were based on viewing static, 2-dimensional images, whereas in real-life subjective assessments of age and fatigue are made on 3-dimensional, dynamic images. A quantitative analysis would have been more challenging using video, because it would have been more difficult to map out precisely the LookZones on moving images. We would also expect that movement of the mouth while talking may result in relatively increased attention directed toward the perioral area. Another limitation is that all participants were young adults; it is possible that results might vary with participants of different ages.

The results of the current study, which demonstrate that the eye region is most important in making age and fatigue judgments, raises the possibility that aesthetic or functional surgery to the eye region may be an efficient, effective intervention in enhancing an individual's perceived attractiveness by possibly reducing how old or tired one appears. To this end, we aim to extend this study to compare before and after ratings of patients who have undergone plastic surgery. These results nonetheless suggest a corollary to the adage, “beauty is in the eye of the beholder,” such that beauty also appears to reside “in the eye of the beholdee.”

References

1. 1Thorpe SJ, Ahmed B, Steer K. Reasons for undergoing cosmetic surgery: retrospective study. Sexualities, Evolution and Gender. 2004;6:75–96.

2. 2Masip J, Garrido E, Herrero C. Facial appearance and impressions of credibility: the effects of facial babyishness and age on person perception. Int J Psychol. 2004;39:276–289.

3. 3Hummert ML, Garstka TA, Shaner JL. Stereotyping of older adults: the role of target facial cues and perceiver characteristics. Psychol Aging. 1997;12:107–114. MEDLINE | CrossRef

4. 4Berry DS. Attractive faces are all not equal: joint effects of facial babyishness and attractiveness on social perception. Pers Soc Psychol Bull. 1991;17:523–531. CrossRef

5. 5Perry BD, Czyzewski DI, Lopez MA, et al. Neuropsychologic impact of facial deformities in children. Clin Plast Surg. 1998;25:587–597. MEDLINE

6. 6Rankin M, Borah G. Perceived functional impact of abnormal facial appearance. Plast Reconstr Surg. 2003;111:2140–2148. MEDLINE | CrossRef

7. 7Grossbart TA, Sarwer DB. Cosmetic surgery: surgical tools-psychosocial goals. Semin Cutan Med Surg. 1999;18:101–111. MEDLINE | CrossRef

8. 8Honigman R, Phillips K, Castle D. A review of psychosocial outcomes for patients seeking cosmetic surgery. Plast Reconstr Surg. 2004;113:1229–1237. MEDLINE | CrossRef

9. 9Rankin M, Borah G, Perry A. Quality-of-life outcomes after cosmetic surgery. Plast Reconstr Surg. 1998;102:2139–2147. MEDLINE | CrossRef

10. 10Isaacowitz DM. The gaze of the optimist. Pers Soc Psychol Bull. 2005;3:407–415. CrossRef

11. 11Isaacowitz DM. An attentional perspective on successful socioemotional aging: theory and preliminary evidence. Res Hum Dev. 2005;3:115–132.

12. 12Isaacowitz DM, Wadlinger HA, Goren D, Wilson HR. Selective preference in visual fixation away from negative images in old age? (An eye tracking study). Psychol Aging. 2006;21:40–48. MEDLINE | CrossRef

13. 13Lundqvist D, Ohman A. Emotion regulates attentions: the relation between facial configurations, facial emotion, and visual attention. Vis Cogn. 2005;12:51–84.

14. 14Henss R. Perceiving age and attractiveness in facial photographs. J Appl Soc Psychol. 1991;21:933–946.

1 Harvard Department of Ophthalmology Boston, Massachusetts

2 Department of Psychology, Brandeis University, Waltham, Massachusetts

3 Eye Plastics Consultants, Brookline, Massachusetts

4 Department of Ophthalmology, University of Tennessee Health Science Center, Memphis, Tennessee

Correspondence: Peter A. D. Rubin, MD, FACS, Eye Plastics Consultants, 44 Washington St., Brookline, MA 02445

Manuscript no. 2007-1578.

Financial Disclosure(s): The author(s) have no proprietary or commercial interest in any materials discussed in this article.

This study acknowledges support from an unrestricted grant awarded to University of Tennessee Health Science Center from Research to Prevent Blindness, Inc.

PII: S0161-6420(08)01027-0

doi:10.1016/j.ophtha.2008.10.007

30 of 39