How well do we know our own faces?

Although mirrors, photographs and proprioceptive information are widely used for feedback about our own faces, surprisingly little is known about self-face recognition thresholds (SFT) for facial features or their spatial relationship (Felisberti, 2009 Perception 38 ECVP Supplement, 80). Here SFT was investigated using digital photographs of participants‘ faces. Eyes, nose, mouth or chin areas were manipulated either individually or simultaneously. In study 1 (N=16), participants had to choose which of two images (original vs. chimeric) was the original one. In study 2 (N=30), videoclips were created showing sequences of features increasing and decreasing in area. Participants had to stop the videoclips when their original self appeared. Results revealed that SFT was not as good as many individuals expected. SFT was easier when features were made larger (F(1,15)= 44.74, p= 0.001, n2= 0.75) and for simultaneous changes in eyes, mouth and nose (F(4,60)= 6.10, p= 0.001, n2= 0.29). SFT for eyes (high contrast region) was usually lower than for mouth, nose and chin. The same pattern of results was found with videoclips, although SFTs in the latter case were overall lower. Most females were more accurate in self-face recognition than males. Further research is needed to uncover the reasons underpinning the strong individual variations in SFTs.