Feature-specific neural reactivation during episodic memory.

Area of Science:

• Neuroscience
• Cognitive Psychology
• Machine Learning

Background:

• Episodic visual recall involves complex neural processes.
• Understanding how features are represented and reactivated in the brain is crucial for memory research.

Purpose of the Study:

• To introduce a novel analytical approach, feature-specific informational connectivity (FSIC).
• To decode neural reactivation patterns associated with different feature types during visual recall.
• To investigate the relationship between feature reactivation, memory vividness, and recognition accuracy.

Main Methods:

• Utilized functional magnetic resonance imaging (fMRI) data from participants performing an episodic visual recall task.
• Applied a neural network to extract hierarchical representations for feature decoding.
• Developed and implemented the feature-specific informational connectivity (FSIC) method.

Main Results:

• Neural reactivation of low-level (edges), high-level (facial features), and semantic (object categories) features was observed across visual and frontal cortices.
• Reactivation of both low- and high-level features correlated positively with memory vividness.
• Reactivation of low-level features, but not high-level features, correlated with recognition accuracy for semantically similar images.

Conclusions:

• FSIC is a valuable tool for mapping feature-specific neural reactivation.
• Both low- and high-level visual features contribute to the vividness of episodic memories.
• Findings challenge a strict posterior-to-anterior progression of visual information processing in memory recall.