Waveform-based classification of dentate spikes.

Area of Science:

• Neuroscience
• Computational Neuroscience
• Memory Research

Background:

• Dentate spikes (DSs) are prominent local field potential (LFP) patterns in the hippocampus, linked to memory consolidation during quiet states.
• Current classification of DSs into types 1 and 2 relies on current source density (CSD) analysis, requiring multi-electrode recordings.
• The functional roles of different DS types in mnemonic processes remain largely unknown.

Purpose of the Study:

• To develop an unsupervised method for classifying DS types using waveform analysis from single-electrode recordings.
• To enable the investigation of DS type-specific functions in memory consolidation.
• To apply the method to study DS alterations in a mouse model of Alzheimer's disease.

Main Methods:

• Developed an unsupervised classification method using Gaussian mixture models based on DS waveforms.
• Validated the method against CSD-based classification in mice with DG laminar profiles.
• Applied the classification technique to analyze LFPs from apolipoprotein (apo) E3 and apoE4 knock-in mice.

Main Results:

• The Gaussian mixture model approach achieved high classification accuracy (>80%).
• Classified DS types exhibited CSDs, waveforms, rates, and widths consistent with CSD-based classification.
• ApoE4 knock-in mice, modeling Alzheimer's disease, showed wider DSs (both types) from a young age, particularly type 2.

Conclusions:

• The developed waveform-based classification method effectively distinguishes DS types from single-electrode recordings.
• DS waveforms contain information about their origin, suggesting distinct network dynamics and roles in memory.
• Early alterations in DS morphology in apoE4 mice may indicate pathophysiological changes in the entorhinal cortex-dentate gyrus network.