Brain and eye movement dynamics track the transition from learning to memory-guided action

Area of Science:

• Cognitive Neuroscience
• Neuroscience
• Psychology

Background:

• Humans continuously learn and update knowledge to optimize memory-guided actions.
• Mastering an environment involves a gradual shift from acquiring new information to relying on established memories.

Purpose of the Study:

• To investigate the neural and behavioral dynamics underlying the transition from learning to memory-guided action.
• To examine how electroencephalogram (EEG) patterns and eye movements reflect this learning process.

Main Methods:

• A spatiotemporal image learning task ('Memory Arena') was designed for participants to learn 50 items.
• Multivariate EEG patterns and anticipatory eye movements were recorded during task-free periods between image presentations.
• Decoding analyses were performed on EEG data to identify representations of previous and upcoming images.

Main Results:

• Decoding of previous and upcoming image identities was successful during task-free periods.
• The strength of decoding showed opposing trends: decreasing for previous images and increasing for upcoming images over time.
• Behavioral learning rates correlated with the ratio of upcoming to previous image evidence in EEG.
• Eye movement precision increased, anticipating upcoming image locations, mirroring EEG decoding and behavioral performance.

Conclusions:

• The study demonstrates a dynamic neural and behavioral shift from active learning to memory-based action.
• Opposing trends in EEG decoding reflect the brain's transition in representing past versus future information.
• Anticipatory eye movements and neural representations align, providing a comprehensive view of learning dynamics.