Decoding the details of visual working memory representations from ERP and alpha oscillation activities

Area of Science:

• Cognitive Neuroscience
• Neuroscience
• Psychology

Background:

• Working memory (WM) temporarily stores and processes information for action guidance.
• Attribute amnesia in WM leads to forgetting even with sufficient attention.
• Understanding feature priorities in WM is crucial for explaining memory limitations.

Purpose of the Study:

• To investigate if different object features (color, shape) have varying priorities for WM reselection.
• To examine underlying electroencephalography (EEG) signals for detailed feature information storage in WM.
• To explore the neural basis of attribute amnesia.

Main Methods:

• Participants performed a change detection task while EEG data was collected.
• Univariate analysis and multivariate pattern analysis (MVPA) were used to analyze EEG signals.
• Behavioral accuracy and reaction times were measured.

Main Results:

• Behaviorally, color was recalled faster and more accurately than shape.
• Event-related potential (ERP) analysis showed a larger P300 amplitude for color compared to shape.
• MVPA successfully decoded specific color and shape values from EEG data.

Conclusions:

• Color features are prioritized over shape features during WM reselection.
• Detailed feature information, including specific values, is retained in WM.
• EEG signals contain discriminative information about stored object features.