Tracking the flow of hippocampal computation: Pattern separation, pattern completion, and attractor dynamics.

Area of Science:

• Neuroscience
• Computational Neuroscience
• Cognitive Neuroscience

Background:

• Hippocampal theories traditionally assign pattern separation to the dentate gyrus (DG) and pattern completion to the CA3 region.
• Limited neurophysiological data from DG, CA3, and entorhinal cortex historically hindered testing these theories.
• Recent advancements have increased recordings from CA3, medial entorhinal cortex (MEC), lateral entorhinal cortex (LEC), and DG.

Purpose of the Study:

• To investigate the mnemonic functions of the hippocampus, specifically pattern separation and completion.
• To test computational theories using neurophysiological recordings in a novel experimental paradigm.
• To compare hippocampal and entorhinal cortex representations during a spatial memory task.

Main Methods:

• Multi-region electrophysiological recordings in behaving animals.
• Analysis of neural representations in anterior thalamus, MEC, LEC, DG, CA3, and CA1.
• Utilized a local-global cue mismatch (double-rotation) experiment to probe spatial memory functions.

Main Results:

• DG representations showed greater change than entorhinal inputs in the cue-mismatch environment, supporting DG's pattern separation role.
• CA3 representations showed less change than inputs, supporting CA3's pattern completion/error correction function.
• Findings align with continuous attractor network models of hippocampal function.

Conclusions:

• The dentate gyrus (DG) appears to perform an automatic pattern separation function.
• The CA3 region's attractor dynamics enable flexible pattern separation or completion based on input.
• Results provide crucial neurophysiological evidence for distinct computational roles within the hippocampus.