Gender differences in attentional bias after owning a virtual avatar with increased weight

Abstract

Introduction: Eating Disorder (ED) patients selectively attend to appearance cues in preference to other information, in a phenomenon known as Attentional Bias (AB). The latest VR Head Mounted Displays (HMD) offer the chance to include Eye-Tracking (ET) devices, and thus provide more objective measures of body-related attention. This study aims to combine VR and ET technologies and use a VR-based embodiment technique while measuring real-time attention patterns. Specifically, we assess gender differences in eye-gaze behaviors towards specific weight-related or non-weight-related body parts when participants own a virtual avatar with different body sizes. Method: Thirty-five college students (25 women and 10 men) were exposed to an immersive virtual environment in which they were embodied in three avatars with different body sizes: first, one with the same body size as the participant; second, one larger than the participant; and finally, repetition of the avatar with the same body size as the participant. To analyze the gaze data Weight-related Areas of Interest (WAOIs) and Non-Weight-related Areas of Interest (NW-AOIs) were defined. Fixation points and complete fixation time on each AOI were recorded at the three different assessment times. Raw data from the Pupil Labs eye tracking add-on for the HTC-Vive headset were subsequently transformed into percentages for further analysis. Results: Mixed between (Gender)-within (Time) analyses of variance showednon-statistically significant interaction between gender and time (p>.05) and a non-statistically significant difference in fixation points and complete fixation times (p>.05), over the three assessment times. However, a statistically significant gender difference was found in fixation points (F (1.33) =10,030, p= .003, η2 = 0.233) and complete fixation time (F (1.33) =13,017, p= .001, η2 = 0.28. Overall, women reported significantly higher levels of fixation points and complete fixation times in W-AOIs than men. Women showed an increasing gaze pattern towards W-AOIs at the three different assessment times, while men showed an opposite gaze pattern towards NW-AOIs at the three different times. Interestingly, the greatest differences between men and women were found at the third assessment, when they once again owned an avatar with the same body size as themselves. Conclusion: This study provides useful information about gender differences in gaze pattern behaviors while participants owned a virtual avatar with different body sizes. To our knowledge, this is the first study to compare gaze pattern behaviors between women and men using VR-based embodiment techniques and ET attentional bias assessment. The use of these two technologies opens a promising new area in the assessment or treatment of Eating Disorders and body image disturbances.