Contextual prediction errors reorganize naturalistic episodic memories in time

Area of Science:

• Cognitive Neuroscience
• Neuroscience
• Psychology

Background:

• Episodic memories are contextual and temporally ordered, relying on associative binding.
• Prediction errors are crucial for declarative memory but their role in episodic memory's temporal dimension is unexplored.

Purpose of the Study:

• To investigate the impact of prediction errors on the temporal ordering of complex episodic memories.
• To understand how prediction errors influence the learning and reorganization of event sequences over time.

Main Methods:

• Participants experienced naturalistic contexts with event sequences, some containing prediction errors.
• Drift-diffusion modeling was used to analyze recall speed and decision thresholds.
• Correlation analysis examined the relationship between decision threshold adjustments and recall speed.

Main Results:

• Prediction errors altered the temporal ordering of event structures, weakening wrongly predicted older sequences.
• Newly encoded sequences with prediction errors showed high accuracy, even after single exposure.
• Lower decision thresholds and faster recall were observed for newer sequences with prediction errors.

Conclusions:

• Prediction errors play a temporally distinct and adaptive role in learning and reorganizing episodic temporal sequences.
• The findings highlight the dynamic nature of episodic memory and the influence of prediction errors on temporal structure.