Improving behavioral assessment in genetically modified mice.

改善转基因小鼠的行为评估。

• 批准号: 20K14267
• 负责人: Giovanni Sala
• 金额: $ 2.66万
• 依托单位: National Center for Geriatrics and Gerontology (2021-2022)Fujita Health University (2020)
• 依托单位国家: 日本
• 项目类别: Grant-in-Aid for Early-Career Scientists
• 财政年份: 2020
• 资助国家: 日本
• 起止时间: 2020-04-01 至 2023-03-31
• 项目状态: 已结题
• 来源: https://kaken.nii.ac.jp/en/grant/KAKENHI-PROJECT-20K14267/
• 关键词: mouse behaviour; statistics; machine learning; psychometrics; mouse; mouse behavior; data science; factor analysis; cluster analysis; Data reduction; Animal research; Genetics

I have established, via dimensionality reduction statistical analyses, that locomotor activity and anxiety-like behaviour in mice are, to date, two indistinguishable constructs to a large extent. This finding is of extreme interest. In a nutshell, this outcome leads to the ascertainment that pretty much all conclusions relative to anxiety-like behaviour in mice models of psychiatric disease are, at least to a large extent, moot. Consequently, future research will be necessary in order to improve the trustworthiness of mouse behaviour research. Specifically, some non-locomotor-activity-related measurements of mouse anxiety should be developed and introduced in mouse model research in order to reliably dissociate between the two above constructs.The second research achievement refers to how correctly model mouse locomotor activity in time series settings. Most researchers in the field employ simple ANOVA models. Regrettably, such models suffer from serious shortcomings including unrealistic statistical assumptions (e.g., linearity and Gaussian distribution). I have designed a Generalized Additive Model Location Shape and Scale (GAMLSS) model that significantly improve the accuracy of the results.These results have been disseminated in several international conferences.

通过降维统计分析，我已经确定，到目前为止，老鼠的运动活动和类似焦虑的行为在很大程度上是两个难以区分的概念。这一发现极具吸引力。简而言之，这一结果导致确定，几乎所有与精神疾病小鼠模型中的焦虑行为有关的结论，至少在很大程度上是没有意义的。因此，未来的研究将是必要的，以提高老鼠行为研究的可信度。具体地说，在小鼠模型研究中，应该开发和引入一些与运动活动无关的小鼠焦虑测量方法，以便可靠地分离上述两种结构。第二个研究成果是如何在时间序列环境下正确地模拟小鼠的运动活动。该领域的大多数研究人员使用简单的方差分析模型。遗憾的是，这些模型存在严重缺陷，包括不切实际的统计假设(例如，线性和高斯分布)。我设计了一个广义加性模型位置、形状和尺度(GAMLSS)模型，显著提高了结果的准确性。这些结果已经在几个国际会议上传播。

项目成果

Assessment of Mouse Behaviour

小鼠行为评估

• 发表时间: 2020
• 作者: 松島佳苗;渥美剛史;Mrinmoy Chakrabarty;井手正和;Giovanni Sala
• 通讯作者: Giovanni Sala

Improving Reliability and Phenotyping in Mouse Behavior with Machine Learning.

通过机器学习提高小鼠行为的可靠性和表型。

• 发表时间: 2021
• 作者: 吉本 早苗;吉田 昌弘;原田 泰彦;富田 望;請園正敏;Giovanni Sala
• 通讯作者: Giovanni Sala

An integrated data management system for mouse behavior data

小鼠行为数据综合数据管理系统

• 发表时间: 2021
• 作者: 吉本 早苗;早坂 智幸;Tsuyoshi Miyakawa
• 通讯作者: Tsuyoshi Miyakawa