Perceptual learning modifies inversion effects for faces and textures.

Area of Science:

• Cognitive Psychology
• Neuroscience
• Visual Perception

Background:

• The face-inversion effect, a phenomenon where inverted faces are harder to recognize than upright faces, is a key aspect of human face perception.
• Current models debate whether this effect relies on holistic face encoding or featural analysis.

Purpose of the Study:

• To investigate the role of training and orientation specificity in modulating the face-inversion effect.
• To explore how learning influences the encoding of facial features and configural information.

Main Methods:

• Participants were trained to identify either upright or inverted faces using a 10-alternative forced-choice (10AFC) task.
• Performance was subsequently assessed using same- and different-orientation face pairs (upright and inverted).
• The study also included experiments with phase-scrambled faces to examine configural processing.

Main Results:

• Both upright- and inverted-face training improved recognition performance.
• Performance gains were specific to the trained face exemplars and orientations.
• Upright training increased the face-inversion effect, while inverted training decreased it for trained faces.
• Similar orientation-specific learning effects emerged for phase-scrambled faces, indicating learning impacts configural processing.

Conclusions:

• Face perception learning is highly specific to trained orientations and exemplars.
• Training can dynamically alter the face-inversion effect, suggesting flexible face-encoding mechanisms.
• These findings challenge purely holistic models and support views incorporating adaptable feature-based learning in face recognition.