Retrieving noncued episodic memories by taking a cascading walk down memory lane(s).

Area of Science:

• Cognitive Psychology
• Neuroscience
• Computational Neuroscience

Background:

• Episodic memory retrieval ease varies.
• Computational models for searching past experiences are needed.

Purpose of the Study:

• Propose a cascading memory retrieval model.
• Explain variability in recall ease and searchability of episodic memories.

Main Methods:

• Modeled memory retrieval as sequential steps on a directed memory graph.
• Conducted two experiments to validate predictions.

Main Results:

• Memory retrieval exhibits predictable ordering.
• Recall is less likely if mnemonic paths have dead ends.
• Memory retrieval is guided by the associative network's topology.

Conclusions:

• Cascading walk model explains episodic memory retrieval.
• Provides a mechanism for hippocampal (p)replay and mind wandering.
• Suggests episodic and semantic knowledge retrieval may use random walk processes.