DeepLabCut-Based Mouse Exploratory Behavior Metrics for Aging and Periodontitis Models

A DeepLabCut-based system automatically quantifies 36 behavioral indicators across sniffing frequency, exploration duration, and novelty preference in mice. The dataset, authored by Cenfei Zhou and last updated in April 2026, captures fine-grained exploration using dynamic distance thresholds of 1 cm, 1.5 cm, and 2 cm. It was applied to detect cognitive differences in natural aging and elderly periodontitis mouse models that traditional visual inspection missed.

Use Cases

• Training machine learning models for rodent behavior analysis based on the 36 quantified indicators.
• Comparing cognitive performance between healthy and diseased animal models based on novelty preference indices.
• Analyzing fine-grained exploratory behavior based on calibrated nose tip and body center tracking with dynamic distance thresholds.
• Validating sensitive behavioral biomarkers for cognitive decline based on metrics like sniffing frequency and exploration duration.

Strengths

• Quantifies 36 specific behavioral indicators across three categories, providing multidimensional analysis.
• Uses dynamic distance thresholds (1 cm, 1.5 cm, 2 cm) to capture fine-grained exploration behavior.
• Successfully identified cognitive differences in aged and periodontitis model mice that traditional methods failed to detect.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• The dataset is very small (44.7 KB), indicating limited scope, and row count is unknown.

Provenance

Source

Cenfei Zhou via figshare

Collection Method

Data generated from a custom hardware arena and high-definition industrial camera, with tracking performed by a trained DeepLabCut deep learning algorithm.

Freshness

Last updated 2026-04-21 05:35:20; freshness should be verified.