AI Pipeline Quantified Alzheimer's Mouse Behaviors

Researchers developed an automated tool to track how mice build nests as a way to measure cognitive and motor function decline.

Updated on Sept. 22, 2026 in Alzheimer’s

Isometric editorial illustration of a simplified modular laboratory habitat featuring geometric chambers and fiber bedding materials for behavioral research.
Researchers at Highwise Health have developed CozyScores, an AI-powered automated pipeline that quantifies nest-building behaviors to track cognitive and motor decline in mouse models of Alzheimer’s disease. AI Illustration. Upload story photo >

Live Poll

Do you believe using AI to analyze animal behavior improves the quality of medical research?

A new AI-powered pipeline called CozyScores has enabled researchers to objectively quantify nest-building behaviors in mouse models of Alzheimer's disease. This automated system offers a more precise method for monitoring the progression of motor and cognitive impairments in rodent studies.

Why it matters

Nest-building is a natural behavior that reflects underlying health, providing researchers with a non-invasive way to track deteriorating cognitive and motor functions. This tool enhances the depth and frequency of data collection compared to manual assessments.

In a preclinical study using C57BL6/J and 3xTg-AD mice, researchers utilized a ResNet18-based AI pipeline to score nest organization on a 5-point scale. The analysis of 92 mice revealed significant performance differences between ages and genotypes (P < 0.05).

The details

The CozyScores pipeline automates the observation of nest organization by analyzing photographs captured between 0.5 and 24 hours. By applying this deep-learning architecture, the system translates visual nest structure into quantitative scores, allowing researchers to track the gradual decline of these specific motor and cognitive skills. The study notably identified performance variations linked to both sex and age during the construction process.

Timeline

  1. 0.5 to 24 hours for photo acquisition timepoints.

  2. 4 hours for the benchmark performance score assessment.

  3. 2 to 8 hours when males performed better than females.

  4. 24 hours for the final maximum performance assessment.

Health Landscape

Behavioral testing in rodent models remains a cornerstone for understanding the progression of neurodegenerative diseases. This AI-driven approach modernizes those assessments by replacing manual scoring with continuous, automated data analysis.

While this tool is currently limited to laboratory research, it represents a shift toward more objective, automated monitoring of cognitive decline. For those managing Alzheimer's or related conditions, these advances emphasize the importance of tracking subtle changes in daily functional abilities with a physician.

The takeaway

Automated monitoring of behaviors can provide sensitive, early indicators of cognitive decline in research settings. Individuals noticing new changes in motor function or complex task performance should always discuss these symptoms with their primary care physician.

Further reading

Learn more about the latest research in our Alzheimer’s section.

Source note: This article includes information reported by Nature.

Live Poll

Do you believe using AI to analyze animal behavior improves the quality of medical research?

AI Pipeline Quantified Alzheimer's Mouse Behaviors