Second-order facial features are processed analytically in composite faces

Area of Science:

• Cognitive Psychology
• Visual Perception
• Human Face Recognition

Background:

• Previous research suggested composite faces undergo holistic processing.
• Contradictory findings indicated composite face morphs are processed similarly to inverted or misaligned faces.
• Schematic faces offer a controlled method to investigate face processing mechanisms.

Purpose of the Study:

• To replicate and extend findings on composite face processing using schematic stimuli.
• To investigate whether schematic faces exhibit holistic processing characteristics.
• To determine the processing mechanisms (serial vs. parallel) for aligned and misaligned schematic faces.

Main Methods:

• Replication of Cheng et al. (2018) using schematic faces with varying second-order features.
• Utilizing a complete composite face task to assess processing.
• Multidimensional scaling (MDS) analysis to determine the best fitting metric (integrality vs. separability).

Main Results:

• Schematic faces showed hallmarks of holistic processing in the composite face task.
• Processing metrics differed from morph faces, aligning more with Euclidean distance (integrality).
• Evidence suggests a mixture of serial and parallel processing for both aligned and misaligned faces, with no strong holistic pooling.

Conclusions:

• Composite faces may not be uniquely processed holistically but rather like objects with separable dimensions.
• Alignment facilitates more parallel processing for composite faces compared to misaligned ones.
• Findings challenge strong holistic processing claims for composite faces, emphasizing separable feature processing.