Stimuli

Thirty-five face images from an in-house database³⁴ were adopted as candidate stimuli and rated in terms of distinctiveness¹⁵ and beauty on a 7-point scale (1 = least distinctive, 7 = most distinctive; 1 = least beautiful, 7 = most beautiful) by an independent group of 20 undergraduates. All images were grayscale Chinese male faces with the external contours removed and were randomly presented at the center of the screen. Two faces that were rated as least distinctive (Face A: 3.35, Face B: 3.15) and almost equally beautiful (Face A: 3.65, Face B: 3.45) were chosen as the stimuli for the formal experiments.

The two images were morphed (using FantaMorph version 5.4.1) to create a continuum. A classification task¹⁵ was carried out to assess the classification of the morphed faces and identify the category boundary of the continuum. An independent group of 30 undergraduates completed this task. Nine morphed images with a 10% interval between the endpoint images (Face A and Face B) were used as the stimuli in this task (e.g., 10% A/90% B, 20% A/80% B… 90% A/10% B). The two endpoints were presented side-by-side at the top of the screen, and a morphed image was presented at the center of the screen below the morphed image. The participants were asked to indicate which endpoint (left or right) most closely resembled the morph by pressing a key. The positions of the endpoint faces were counterbalanced. Each of the nine morphed images was presented five times, yielding a total of 45 trials. Figure 1 shows the mean classification of each morph from Face A to Face B. The categorical boundary was near the midpoint of the continua (50% A/B).

Classification of each morph from Face A to Face B. Nine morphed images with a 10% interval between the endpoint images (Face A and Face B) were presented in the classification task. Each point represents percentage that each morph was classified as Face A.