Towards Robust Automated Mouse Scratching Behavior Detection for Itch Research
- 1University of Pennsylvania
Abstract
Automated detection of mouse scratching behavior could reduce manual annotation and is essential for preclinical itch research. However, robust performance across laboratories remains a challenge for current methods. We present Scratch-AID+, a framework dedicated to automated mouse scratching detection, including a diverse and progressive set of approaches: classic ResNet-based detectors, task-adaptive fine-tuning of video foundation models, and in-context learning with advanced multimodal large language models. To simulate real-world use cases, we benchmark on a public dataset and utilize additional cross-laboratory video recordings for independent testing. We find that classic baselines suffered severe performance degradation on external data under the realistic setting with sparse scratching events. In contrast, fine-tuned video foundation models successfully bridged this generalizability gap, achieving significantly better detection performance. Furthermore, frozen multimodal large language models demonstrated promising out-of-the-box ability using only few-shot demonstrations. We also examined the effects of different video preprocessing strategies. Our findings highlight the improved performance and generalizability of the adapted foundation models for automated scratching behavior analysis and shed light on the real-world deployment of such detectors. Our code is publicly available at https://anonymous.4open.science/r/scratch aid-plus/.
Related
- Semi-Automated CellProfiler Pipelines for Robust Quantification of Microglial Density, Distribution and Morphology
- Evidence for a cohesion-related function of alpaca hums: Linking automated call detection with GPS data
- Robust and accessible in situ sequencing with an open-source automated sequencer
- Robust evidence for theta-band rhythmicity in behavior across two dense-sampling datasets
- Hybrid local–global vision transformer framework for robust plant disease detection in precision agriculture
BibTeX
@misc{yang2026towards,
title = {{Towards Robust Automated Mouse Scratching Behavior Detection for Itch Research}},
author = {S. Yang and A. Pan and H. Shin and H. Chen and Z. Zhang and A. Shad and W. Huang and Y. Liu and W. Liu and H. Luo and C. Xu and J. Chen and M. Xu and W. Luo and H. Yu and L. Shen},
howpublished = {bioRxiv},
year = {2026},
month = oct,
doi = {10.64898/2026.10.12.758682},
url = {https://www.biorxiv.org/content/10.64898/2026.10.12.758682v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - Towards Robust Automated Mouse Scratching Behavior Detection for Itch Research
AU - S. Yang
AU - A. Pan
AU - H. Shin
AU - H. Chen
AU - Z. Zhang
AU - A. Shad
AU - W. Huang
AU - Y. Liu
AU - W. Liu
AU - H. Luo
AU - C. Xu
AU - J. Chen
AU - M. Xu
AU - W. Luo
AU - H. Yu
AU - L. Shen
PY - 2026
DA - 2026/10/12
PB - bioRxiv
DO - 10.64898/2026.10.12.758682
UR - https://www.biorxiv.org/content/10.64898/2026.10.12.758682v1
AB - Automated detection of mouse scratching behavior could reduce manual annotation and is essential for preclinical itch research. However, robust performance across laboratories remains a challenge for current methods. We present Scratch-AID+, a framework dedicated to automated mouse scratching detection, including a diverse and progressive set of approaches: classic ResNet-based detectors, task-adaptive fine-tuning of video foundation models, and in-context learning with advanced multimodal large language models. To simulate real-world use cases, we benchmark on a public dataset and utilize additional cross-laboratory video recordings for independent testing. We find that classic baselines suffered severe performance degradation on external data under the realistic setting with sparse scratching events. In contrast, fine-tuned video foundation models successfully bridged this generalizability gap, achieving significantly better detection performance. Furthermore, frozen multimodal large language models demonstrated promising out-of-the-box ability using only few-shot demonstrations. We also examined the effects of different video preprocessing strategies. Our findings highlight the improved performance and generalizability of the adapted foundation models for automated scratching behavior analysis and shed light on the real-world deployment of such detectors. Our code is publicly available at https://anonymous.4open.science/r/scratch aid-plus/.
ER -