Unconscious familiarity and local context effects on low-level face processing: a reconstruction hypothesis

Area of Science:

• Cognitive Psychology
• Neuroscience
• Computer Vision

Background:

• Traditional face recognition models propose distinct, static representations for each known face.
• It is commonly believed that semantic or memory-based information does not influence low-level face processing.

Purpose of the Study:

• To investigate the nature of face representations in memory.
• To examine the flow of information processing during face recognition.
• To test whether semantic information influences low-level face perception.

Main Methods:

• Two experiments involved participants searching for a primed target face among distractors.
• Experiment 1 manipulated the fame of distractor faces (same vs. different fame).
• Experiment 2 manipulated the priming of the target face (semantic vs. no information).

Main Results:

• Increasing the number of faces (display set size) slowed down face detection time.
• This increase in detection time was dependent on the fame of the target face.
• The search context and semantic priming significantly affected famous face detection speed.

Conclusions:

• Face representations are not static but are actively reconstructed.
• High-level semantic information and low-level perceptual information are integrated during face processing.
• This integration process occurs automatically and outside of conscious awareness.