Figure 2.

Behavioral paradigm and results. (A) Visual stimuli. All images represent one face identity, showing five levels of facial traits. Dominance and trustworthiness axes were manipulated in three steps (−3, 0, 3) of one standard deviation straddling the neutral. Faces were generated using the Facegen Modeler program (http://facegen.com version 3.1), based upon an extensively validated model (9). (B) Experiment 1: 2AFC temporal detection. Patients fixated on a central cross, with onset of each 1,500 ms interval alerted by a low (interval 1) or high pitch (interval 2) tone. Stimuli were located inside the scotoma (see Figure 1), and could appear in either interval at random, for a period of 500 ms. At the end of the trial, participants had to decide in which interval it appeared. Stimuli consisted of a face of 5.25 or 7.25° height, exhibiting either a dominant, submissive, trustworthy, untrustworthy, or neutral expression, at random. Experiment 2: 2AFC face trait discrimination. Throughout each trial of 2,500 ms duration, participants were required to fixate on a central black cross. During this time, the stimulus appeared inside the scotoma for 500 ms with jittered onset. At the end of the trial, patients had to indicate whether the face had been “friendly” or “unfriendly.” If they saw nothing, they were instructed to guess. (C) Mean behavioral performance for all participants who completed this task (n = 11) ± standard error of the mean (SEM) for 2AFC temporal detection, as a function of face trait. Statistically significant performance above chance was shown with red symbols, below chance in blue, all others in black. (D) Mean behavioral performance for all participants who completed this task (n = 11) ± SEM for 2AFC face trait discrimination.