2026-07-21 東京大学
撮影動画からAIが肝炎モデルマウスの症状を数値化
＜関連情報＞
- https://www.a.u-tokyo.ac.jp/topics/topics_20260721-1.html
- https://www.nature.com/articles/s41598-026-60820-3
詳細な行動表現型解析により、代謝機能障害関連脂肪性肝炎のマウスモデルにおける新たな特徴が明らかになった Deep behavioral phenotyping reveals novel features in a mouse model of metabolic dysfunction-associated steatohepatitis
Naoaki Sakamoto,Takamasa Numano,Yui Kobayashi,Masahiro Fukuda,Maria Osaki,Keisuke Omori,Taichi Yamamoto & Takahisa Murata
Scientific Reports Published:21 July 2026
DOI:https://doi.org/10.1038/s41598-026-60820-3
Abstract
Patients with metabolic dysfunction-associated steatohepatitis (MASH) often suffer from a broad range of extrahepatic symptoms including fatigue and pruritus. However, evaluating behavioral abnormalities in preclinical mouse models remains challenging due to the limitations of conventional short-duration behavioral tests. This study aimed to comprehensively profile 24-hour behaviors of a MASH mouse model using a high-resolution, artificial intelligence (AI)-based approach. C57BL/6J mice were fed a choline-deficient, L-amino acid-defined high-fat diet from 6 weeks of age. We recorded videos across the dark/light phases at 8, 10, 12, and 14 weeks of age and analyzed behaviors using a novel AI-based behavioral analysis system. Twenty-four-hour behavioral analysis revealed that this model exhibited decreased locomotor activity alongside increased grooming and scratching behaviors as steatohepatitis progressed. These mice also exhibited altered eating-drinking rhythms and rearing patterns. Notably, these behavioral changes may reflect symptoms observed in patients with MASH; for instance, the reduction in locomotor activity is indicative of a fatigue-related phenotype. This comprehensive profiling highlights that AI-based behavioral analysis can identify novel behavioral phenotypes, effectively bridging the gap between mouse models and humans.
