Serial dependence in face-gender classification revealed in low-beta frequency EEG

Area of Science:

• Cognitive Neuroscience
• Perception
• Human Brain Activity

Background:

• Perception is influenced by immediate past experiences, known as serial dependence.
• This phenomenon is particularly pronounced in face perception.
• The underlying neurophysiological mechanisms remain unclear.

Purpose of the Study:

• To investigate the neural basis of serial dependence in face-gender perception.
• To determine if past facial stimuli can be decoded from ongoing neural activity.
• To link neural signals to the strength of perceptual bias.

Main Methods:

• Participants discriminated face-gender from sequences of male, female, and androgynous images.
• Electroencephalography (EEG) was recorded during the task.
• Machine learning classification techniques were used to decode previous trial information from EEG.

Main Results:

• Strong serial dependence was observed, with androgynous faces biased by preceding gender.
• The strength of this bias oscillated at specific beta frequencies (14 Hz for female, 18 Hz for male priors).
• Classification accuracy of previous stimulus gender from EEG correlated with individual serial dependence strength.

Conclusions:

• Recent face-gender experience is selectively encoded in beta-frequency neural oscillations (14-20 Hz).
• Neural signals from previous trials are maintained and influence current perception.
• These findings implicate beta-frequency neural activity as the substrate for serial dependence in face perception.