Eyes on the bodies: An eye tracking study on deployment of visual attention among females with body dissatisfaction

Highlights

• This study investigated attentional bias among women with BD.

• Event-related and epoch-related analyses of eye movements were conducted.

• Women with BD showed sustained maintenance biases on thin and fat body images.

Abstract

Visual attentional bias has important functions during the appearance social comparisons. However, for the limitations of experimental paradigms or analysis methods in previous studies, the time course of attentional bias to thin and fat body images among women with body dissatisfaction (BD) has still been unclear. In using free reviewing task combined with eye movement tracking, and based on event-related analyses of the critical first eye movement events, as well as epoch-related analyses of gaze durations, the current study investigated different attentional bias components to body shape/part images during 15 s presentation time among 34 high BD and 34 non-BD young women. In comparison to the controls, women with BD showed sustained maintenance biases on thin and fat body images during both early automatic and late strategic processing stages. This study highlights a clear need for research on the dynamics of attentional biases related to body image and eating disturbances.

Introduction

Body dissatisfaction (BD) is one of the prominent risk and maintenance factors for eating disorders (Stice & Shaw, 2002). In Western cultures, approximately 40%–50% of women have some level of body dissatisfaction (Bearman, Martinez, Stice, & Presnell, 2006), and the proportion in China is similar to that in Western societies (Chen, 2006).

Sociocultural factors have been found to play a prominent role in the development of body image dissatisfaction and disordered eating. According to the sociocultural model, body dissatisfaction and disordered eatings are results of internalizing the increasing pressures for women to achieve an ultra-slender figure/the thin ideal, which current culture emphasizes as an essential component of beauty (Fitzsimmons-Craft, 2011). In particular, the mass media has been identified as a potent source of unrealistic thin ideals (Thompson, Coovert, & Stormer, 1999). When watching fashion magazines or web sites, appearance social comparisons involuntary happen and are generally upward, as females tend to compare themselves to thin and attractive media images (Morrison et al., 2004, Van den Berg et al., 2002). Therefore, appearance social comparison serves as a link between mass media and body dissatisfaction, and it could lead to many negative outcomes, such as feelings of discontent (Thompson & Heinberg, 1999), weight concerns (Posavac, Posavac, & Posavac, 1998) and depressions (Lorenzen, Grieve, & Thomas, 2004).

Visual attentional bias has important functions during the upward appearance comparisons (Cho and Lee, 2013, Glauert et al., 2010). The way people attend to their own body and someone else's body may contribute to the outcome of a social comparison process (Festinger, 1954). Contrast effects in an upward comparison happen when you allocate relatively more attention to the thin ideal media images and to the unattractive body parts of your own bodies, which increases the perceived attractiveness of the comparison target and unattractiveness of your own body. This may then negatively affect the outcome of the social comparison process, and thus possibly increase body dissatisfaction (Roefs et al., 2008). Previous research suggested that individuals with body image concerns may have upward appearance comparison tendencies (Corning, Krumm, & Smitham, 2006), and the unique cognitive processes often characterize individuals with body dissatisfaction or disordered eating. For examples, Onden-Lim, Wu, and Grisham (2012) reported that body dysmorphic concern is positively associated with attention toward attractive appearance images rather than unattractive ones. Jansen, Nederkoorn, and Mulkens (2005) found that participants with eating symptoms paid most visual attention to the self-identified ugly body part of their own and beautiful part of another woman's body, while the normal controls did exactly the opposite. Roefs et al. (2008) replicated these findings among women with a lower rating of attractiveness to their own body. Also, Cho and Lee (2013) reported that in comparison to low BD women, high BD women observed the thin ideal bodies longer and more frequently. This maintenance attentional bias to thin ideal images has been also found among patients with bulimia nervosa (Blechert, Nickert, Caffier, & Tuschen-Caffier, 2009).

However, there have been several limitations within this research area. Firstly, relatively little research has investigated the impact of exposure to overweight/fat media images (i.e., downward comparison), which is not to be neglected. If exposure to overweight/fat media images leads to an assimilation effect (i.e., a comparer's self-evaluation is assimilated toward the perceived evaluation of the comparison target), it could result in a more negative self-evaluation (Harrison, 2000, Jansen and de Vries, 2002). Several studies have reported negative effects of exposure to unattractive body images. For example, Papies and Nicolaije (2012) reported that restrained eaters felt worse about themselves if they felt similar to a plus-size model than if they felt similar to a slim model, but they felt better about themselves if they felt dissimilar to a plus-size model than if they felt dissimilar to a slim model. Moreover, the effects of exposure to fat body images exhibited individual differences that after exposure to overweight/fat media images some females can feel worse than others about their body, but some females may not be affected at all. For example, one study reported that women higher in both BMI and neuroticism experienced greater BD after being exposed to overweight images (Dalley, Buunk, & Umit, 2009). Some other studies also indicated that women with higher BMIs compatible with body dissatisfaction or drive for thinness (Gao et al., 2013, Roefs et al., 2008) were more vulnerable to the negative effects of exposure to fat body images. Therefore, having a higher BMI may not be problematic unless the individual also has maladaptive cognitive structures. Upon the above findings along with the facts that BMI was found to be related to body dissatisfaction in large amounts of studies (Annis et al., 2004, Schwartz and Brownell, 2004), the pre-existed body dissatisfaction may be one of the critical factors that could promote an assimilation effect on exposure to overweight/fat body images. Women high in BD would observe the overweight/fat body images carefully and may assimilate self-evaluations of their own body to the unattractive bodies. Although the attention directed to overweight/fat media images among women with BD or disordered eating is mixed (Janelle, Hausenblas, Fallon, & Gardner, 2003), empirical evidence has emerged that women high in BD showed attention to (Gao, Wang, Chen, Wang, & Zhang, 2012) or disengagement difficulty from fat body images (Gao et al., 2013). Following this reasoning, we predict that women high in BD will allocate more attention to fat body images.

Another limitation of previous research is that most of the previous studies failed to investigate time course of attentional bias when confronted with body-related media images. The maintained attention to thin body images reported in previous studies (Blechert et al., 2009, Cho and Lee, 2013, Gao et al., 2012, Jansen et al., 2005, Roefs et al., 2008) almost reflected the slower, strategic and voluntary processing as they depended on the total gaze durations within the whole stimuli presentation. One study has reported attentional bias toward body areas typical of dissatisfaction during latter processing stages among women with BD (Janelle, Hausenblas, Ellis, Coombes, & Duley, 2009). However, rarely did the above eye movements research investigate attentional bias to body images during early processing stages, which was governed mainly by faster and automatic processes. Thereby, attentional components during early processing stages, such as attentional orientation, detection speed or initial maintenance, could not be distinguished. Rinck and Becker (2006) suggested that in order to understand the processes involved in attentional biases, it is necessary to differentiate fast, more automatic processes from slower, cognitively controlled processes, and the distinction is important both theoretically and empirically (Cisler & Koster, 2010). There has been evidence showing that women with body dissatisfaction have attentional bias toward body images during early automatic processing stages. Using the dot probe task and short presentation durations (150 ms and 500 ms) when there is not enough time to shift eye gaze, or to initiate strategic control of attention, Glauert et al. (2010) found that undergraduate females were faster to discriminate the direction of an arrow cue when it appeared in the location previously occupied by a thin than a fat body. This bias was weakest for women who had a higher BMI and elevated body dissatisfaction. Significant correlations between this attentional bias and either BMI or body dissatisfaction were eliminated if either of these two variables was controlled for. However, since the dot probe task provides only a discrete snapshot of responses after the onset of stimuli, it fails to elucidate how attention is deployed prior to behavioral responses (Hermans, Vansteenwegen, & Eelen, 1999). Thus, the components of attentional bias reported by Glauert et al. (2010) were unclear, as the attentional bias could be resulted from faster detection speed, more frequent orientation, longer initial maintenance or some of them.

Therefore, we conducted the present eye tracking study to investigate time course of attentional bias to slim and overweight/fat body shape images. In order to reduce cognitive strategies when responding to experiment tasks, free viewing task with four categories of pictures being presented simultaneously was used in the current study (Rinck & Becker, 2006). This task has good ecological validity and attentional bias will be more likely to happen when multiple stimuli compete for attention (Mathews & Mackintosh, 1998). Although 3 s (Hermans et al., 1999) or 60 s (Rinck & Becker, 2006) presentation durations were used in previous studies, 15 s presentation time was employed in the current study. Three-second presentation time is too short to study time course of attention over a longer interval (Rinck & Becker, 2006), whereas, participants in our pilot study reported that 1 min was too long making their eyes uncomfortable when staring at the screen. Most of the participants found it uninteresting to watch the stimuli 15 to 20 s after the stimuli appeared. Therefore, after being revised several times, 15 s presentation time was adopted. Early attentional bias could be measured by event-related analyses of the critical first eye movement event, with first fixation frequency assessing orienting biases, first fixation latency measuring detection speed and first fixation duration exploring initial attentional maintenance (e.g., Gao, Wang, et al., 2011, Gao et al., 2012, Garner et al., 2006). Also, the allocation of early attention was measured by a detailed analysis of gaze duration in the first 3 s, which was divided into six intervals of 500 ms each (Hermans et al., 1999, Rinck and Becker, 2006, Rohner, 2002). For the epoch-related analyses of the complete presentation time, five intervals of 3 s each were defined to study the whole and late attentional deployment.

On the above review of the literatures, we hypothesized that 1) due to upward social comparison, women high in BD would show sustained maintenance bias to thin body images, that they, in comparison to the normal controls, would allocate their attention more to thin body images (greater gaze durations bias during the 15 s presentation time) than to neutral images; 2) due to downward social comparison, women high in BD, in comparison to the normal controls, would also show attentional maintenance bias to overweight/fat body images (longer gaze durations during the 15 s presentation time) than to neutral pictures; and 3) body dissatisfaction level would increase among women high in BD after exposure to both thin and fat body images, while the controls would have no such change in body dissatisfaction. We also investigated attentional bias components in early processing stages, including attentional orientation, detection speed and initial attentional maintenance.