Abstract
Faghel-Soubeyrand et al. (in press) trained observers to use the facial information most correlated with skilled face-sex discrimination — the eye on the right of the face stimulus from the observer’s viewpoint — and showed that these observers’ performance increased more than that of control participants. Here, using a similar implicit induction procedure, we attempted to train observers to use the information associated with skilled — mostly the two eyes— or unskilled face identification (Tardif et al., 2018). First, participants completed 500 Bubbles trials where they were asked to identify a celebrity, to reveal their use of information pre-induction. Second, participants carried out 500 more trials of the Bubbles task, during which, unbeknownst to them, the base face stimuli were tampered with. In the best-information induction subject group, the information related to skilled face identification was made available (N=8; mean age=21.9; 2 women) and, in the worse-information induction subject group, the information related to unskilled face identification was made available (N=7; mean age=21.9; 3 women). Third, and finally, observers completed 500 more Bubbles trials to reveal their use of information post-induction. For each subject group, we computed classification images, showing the visual information used before and after the induction trials. As expected, results show that participants from the worse-information group used the mouth before and after the induction, whereas participants from the best-information group used mainly the mouth before induction and the two eyes after induction (Cluster Test: p< .05; sigma=26; tC=2.70; Sr=21901; Chauvin et al., 2005). We believe this induction procedure shows promise as a mean for individuals specifically impaired in face recognition (e.g. developmental prosopagnosics) and professionals relying on strong face processing (e.g. police officers) to improve their abilities.