The speed of human social interaction perception

Area of Science:

• Cognitive Neuroscience
• Social Cognition
• Neuroimaging

Background:

• Perceiving social interactions is a core human ability, developing early and shared with primates.
• The neural mechanisms and timing of social interaction recognition remain largely unknown.
• Key questions involve whether recognition is a rapid feedforward process or a slower inference.

Purpose of the Study:

• To investigate the temporal dynamics of neural computations underlying social interaction recognition.
• To determine if social interaction information is processed spontaneously and abstractly in the brain.
• To differentiate between rapid feedforward and slower post-perceptual processing of social cues.

Main Methods:

• Magnetoencephalography (MEG) decoding was employed to analyze brain activity.
• Participants viewed static images of people interacting or acting independently.
• An orthogonal task was used to assess spontaneous processing, followed by an explicit detection task.

Main Results:

• Social interaction presence was decodable from MEG data around 300ms post-image onset, indicating abstract representations.
• This late latency persisted even during an explicit social interaction detection task.
• Different types of social interactions (e.g., mutual gaze vs. joint attention) were distinguished even later, around 500ms.

Conclusions:

• The human brain spontaneously extracts social interaction information, but this process is relatively slow.
• Social interaction recognition likely involves iterative top-down computations rather than purely feedforward processing.
• These findings shed light on the neural basis of social perception and its temporal characteristics.