A new era in quantification of animal social behaviors.
A new era in quantification of animal social behaviors.
复制标题
动物社会行为量化的新时代。
DOI：
10.1016/j.neubiorev.2023.105528
复制
发表时间：
2024
影响因子：
8.2
通讯作者：
Kumar,Vivek
中科院分区：
文献类型：
--
作者：
Choi,JessicaD;Kumar,Vivek
Bordes等人利用先进的计算工具对传统社会行为研究进行了令人兴奋的回顾，强调了计算行为学的力量，以进一步加深我们对复杂精神疾病的理解。在他们的论文《通过机器学习整合行为学和比较心理学方法推进社会行为神经科学》中，Bordes等人阐述了该领域在啮齿动物社会行为研究方面的历史和现状（Bordes 2023a）。Bordes等人指出，从历史上看，行为研究主要是在自然环境中进行的观察性研究，这需要具有广泛行为知识的行为学家。最近的还原论方法允许在高度控制的环境中研究啮齿动物的特定神经领域，实现自动化和可扩展性。然而，这种控制是有代价的，通常是通过在社会隔离中进行实验，这是啮齿动物的已知压力源（Razzoli 2018）。在过去的几十年里，神经科学领域已经扩大了测量Bordes等人所描述的复杂社会行为方面的分析方法的数量。还原论分析包括试管、温点、三室、社会恐惧传递、两种选择的社会决策和社会工具任务。虽然这些方法是有用的，并测量社会行为的组成部分，但这些分析的人为背景限制了可翻译性。目前在神经科学领域使用的行为学分析包括自由移动的多动物社会互动、长期的家庭笼观察、社会回避和食物偏好分析。除了典型的实验室测试条件之外，还有一些“重新野生化”实验动物的例子（Zipple 2023）。目前的临床前实验设计通常在控制遗传机制方面表现出色，然而，它们在代表可翻译为人类治疗的自然环境方面存在不足。这些实验的重点是保持高水平的实验控制，以及提高生态现实主义的水平。野外实验可以在户外的大空间里进行。这更好地模拟了自然的社会互动，例如，在另一只老鼠取得统治地位后，为老鼠提供逃跑的空间。这些实验很难用现有的方法进行分析，因为啮齿动物不容易被摄像机捕捉到，而且它们的挖洞行为会给追踪老鼠带来更多的挑战和困难。目前可以通过wifi对人类进行非视频姿势估计（Li 2019）。研究人员还没有在啮齿动物身上测试这种方法的可行性。
Bordes et al. present an exciting review of traditional social behavioral studies utilizing advanced computational tools, emphasizing the power of computational ethology to further our understanding of complex psychiatric conditions. In their paper,“Advancing social behavioral neuroscience by integrating ethology and comparative psychology methods through machine learning,” Bordes et al. address the history and current state of the field with regard to social behavioral studies in rodents (Bordes 2023a).Bordes et al. point out that historically, behavioral research has been performed through largely observational studies in naturalistic environments that require a behaviorist with extensive knowledge on behavior. A more recent reductionist approach allows for a highly controlled environment to study specific neurological domains in rodents, enabling automation and scalability. However, this control comes at a cost, often by conducting experiments in social isolation, which is a known stressor for rodents (Razzoli 2018). In the past couple of decades, the neuroscience field has expanded the number of assays to measure facets of complex social behavior that Bordes et al. describe. Reductionist assays include tube, warm spot, three-chamber, social fear transmission, two-choice social decisionmaking, and social instrumental tasks. While these methods are useful and measure components of social behaviors, the artificial context of these assays limits translatability. Ethological assays currently used in the neuroscience field include freely-moving multi-animal social interaction, long-term home cage observation, social avoidance, and food preference assays. Beyond the scope of typical lab testing conditions, there have been examples of “rewilding” lab animals (Zipple 2023). Current preclinical experimental designs typically excel at controlling genetic mechanisms, however, they fall short at representing naturalistic environments for translatability to human therapeutics. The focus of these experiments is to retain high levels of experimental control, as well as improve the levels of ecological realism. Re-wilding experiments can occur outdoors in a large space. This better models natural social interaction by, for example, providing mice the space to flee after another mouse asserts dominance. These experiments are difficult to analyze with current methods as the rodents will not be easily captured by cameras, as well as burrowing behavior which would introduce more challenges and difficulties with tracking mice. Non-video pose estimation is currently possible in humans through wifi (Li 2019). Researchers have not yet tested the feasibility of this method in rodents.