Perceptual Adaptation to Facial Asymmetries
Rhodes, Gillian, Louw, Kim, Evangelista, Emma, Psychonomic Bulletin & Review
Humans and other animals are highly sensitive to deviations from bilateral symmetry and prefer symmetric mates. Fluctuating asymmetries (FAs) are random deviations from perfect symmetry that can result from developmental instability. Human perceptions of facial asymmetry are driven by FAs and insensitive to directional asymmetries (DAs), which have a consistent direction of bias (e.g., left side always larger) across the population and are unrelated to developmental stability. We hypothesized that perceptual adaptation may filter out DAs and provide a proximate mechanism for this perceptual focus on FAs. We created a small population of faces with DAs by applying a unilateral distortion to the same side of each face. After 5 min of adaptation, (new) faces with these DAs looked less asymmetric and the most symmetric-looking distortion shifted toward the adapting asymmetry level. Parallel changes occurred for attractiveness. We suggest that perceptual adaptation may provide the proximate mechanism for an evolutionarily adaptive focus on FAs.
Humans and many other animals are highly sensitive to deviations from bilateral symmetry, and they prefer symmetric mates (Møller & Swaddle, 1997; Rhodes, 2006; Rhodes & Simmons, 2007). This preference may be an evolutionary adaptation for finding high-quality mates, given the link between developmental instability and fluctuating asymmetry (FA; Livshits & Kobyliansky, 1991; Mather, 1953; Møller & Thornhill, 1998; Polak, 2003; Van Valen, 1962; Zakharov, 1981). FA is characterized by random deviations from perfect symmetry in bilaterally paired traits (e.g., left eye equally likely to be larger or smaller than right eye) that arise when environmental stress and/or genetic factors disrupt development (Møller & Swaddle, 1997; Polak, 2003). In contrast, directional asymmetry (DA) has a consistent bias across a population (e.g., left eye always larger than right eye) and does not reflect developmental instability (Polak, 2003).
Interestingly, human perceptions of facial symmetry are driven largely by FA and are unrelated to DA (Simmons, Rhodes, Peters, & Koehler, 2004).1 Given that humans find symmetric faces attractive (for reviews, see Rhodes, 2006; Rhodes & Simmons, 2007), this focus on facial FA could be evolutionarily adaptive if facial FA covaries negatively with mate quality. Evidence that facial asymmetry is elevated in some chromosomal disorders (Thornhill & Møller, 1997) suggests that it might. In addition, facial asymmetry has been linked with poor health in nonclinical populations, although this finding remains controversial (for a recent review, see Rhodes & Simmons, 2007).
What proximate mechanism could filter out DA and focus perception on asymmetries that may signal developmental instability? A good candidate is perceptual adaptation, which occurs when response properties change in response to changes in the inputs (Clifford & Rhodes, 2005; Clifford et al., 2007; Schwartz, Hsu, & Dayan, 2007). Perceptual adaptation may serve to calibrate responses to the distribution of inputs experienced and would ensure sensitivity to distributional properties of the inputs, such as whether an asymmetric trait in a face is FA or DA.
Numerous studies have reported perceptual adaptation in face perception (MacLin & Webster, 2001; Rhodes, Jeffery, Watson, Clifford, & Nakayama, 2003; Rhodes et al., 2004; Robbins, McKone, & Edwards, 2007; Watson & Clifford, 2003; Webster & MacLin, 1999). In these studies, participants typically view a series of faces to which a consistent distortion has been applied. For example, the internal features may have been "expanded" spherically from a central point, making the faces look as if they were viewed through a fish-eye lens. After the viewer has adapted to these faces, the distortion becomes less noticeable and faces with slightly expanded features look more normal than undistorted faces. …