Purchase this article with an account.
Nicolas Davidenko, Hema Kopalle, Bruce Bridgeman; A strong bias to fixate the upper eye in tilted faces. Journal of Vision 2018;18(10):164. doi: https://doi.org/10.1167/18.10.164.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
There is a well-known left-gaze bias when looking at faces: the left side of the face (from the point of view of the observer) receives more first fixations and has a greater influence on perceptual decisions than the right side of the face (Campbell, 1978; Guo et al., 2009). Here we investigated whether the in-plane orientation of a face image influences the left-gaze bias. In Experiment 1, we developed an expression classification task that elicited a strong left-gaze bias in upright faces, regardless of whether face images had been mirror-reversed (ruling out stimulus effects). In Experiment 2, a new group of participants completed the same task on faces that were presented either upright, rotated clockwise by 45°, or counterclockwise by 45°, while their eye movements were being tracked. We predicted that in addition to a left-gaze bias to upright faces, participants might show an upper eye bias to tilted faces, manifesting as more first fixations to the left eye of clockwise-rotated faces, but more first fixations to the right eye of counterclockwise-rotated faces. Our findings confirmed this prediction; in fact, the tendency to fixate the upper eye in rotated faces completely dominated any lateral bias (Figure 1A). In Experiment 2, we tested additional orientations (0°, ±11.25°, ±22.5°, ±33.75°, ±45°, ±90°, and 180°) to determine at what angle this upper-eye bias overrides the left-gaze bias. To our surprise, even a small counterclockwise tilt of 11.25° was sufficient to eliminate the left-gaze bias, with the upper eye bias peaking for ±45° faces (Figure 1B). We consider a potential mechanism for this upper eye bias in light of recent findings (de Haas et al., 2016) showing that cortical responses to facial features are enhanced when features appear in canonical positions relative to fixation.
Meeting abstract presented at VSS 2018
This PDF is available to Subscribers Only