Willner Paul
Dept. of Psychology, Swansea University, Swansea SA2 8PP, UK.
Neurobiol Stress. 2016 Aug 22;6:68-77. doi: 10.1016/j.ynstr.2016.08.001. eCollection 2017 Feb.
The chronic mild stress (CMS) model of depression is considered by many to be the animal model of depression that has the greatest validity and translational potential, but it has often been criticized for a perceived lack of reliability. The aims of this study were to establish the extent to which the procedure is reproducible, and to identify experimental variables relevant to its reliability. Because failures to replicate frequently remain unpublished, a survey methodology was used. A questionnaire was circulated to 170 labs identified from a PubMed search as having published a CMS study in the years 2010 or 2015 (with no selection in respect of the results reported). Responses were returned by 71 (42%) of the recipients, followed by further correspondence with some of them. Most of the respondents (n = 53: 75%) reported that the CMS procedure worked reliably in their hands. Of the others, 15 (21%) reported that the procedure was usually reliable, but not always (n = 9: 13%) or not for all measures (n = 6: 8%). Only three respondents (4%) reported being unable to reproduce the characteristic effects, two of whom may be using an insufficient duration of CMS exposure. A series of analyses compared the 75% of 'reliable' labs with the 25% of 'less reliable' labs on a range of experimenter, subject, stress and outcome variables. Few if any significant differences between these two samples were identified, possibly because of the small size and diversity of the 'less reliable' sample. Two other limitations of the study include the (unavoidable) omission of labs that may have worked with the model but not published their data, and the use of ad hoc measures to compare the severity of different stress regimes. The results are discussed in relation to relevant published observations. It is concluded that CMS is in fact a rather robust model, but the factors that result in a less effective implementation in a minority of laboratories remain to be firmly established.
许多人认为，慢性轻度应激（CMS）抑郁症模型是最具效度和转化潜力的抑郁症动物模型，但它常因被认为缺乏可靠性而受到批评。本研究的目的是确定该程序的可重复性程度，并识别与其可靠性相关的实验变量。由于未能重复实验的情况通常未发表，因此采用了调查方法。向通过PubMed搜索确定的170个实验室发放了问卷，这些实验室在2010年或2015年发表过CMS研究（对所报告的结果未作筛选）。71名（42%）收件人回复了问卷，随后与其中一些人进行了进一步通信。大多数受访者（n = 53：75%）报告称，CMS程序在他们手中运行可靠。在其他受访者中，15名（21%）报告称该程序通常可靠，但并非总是如此（n = 9：13%）或并非对所有测量指标都可靠（n = 6：8%）。只有三名受访者（4%）报告无法重现特征性效应，其中两名可能是因为CMS暴露持续时间不足。一系列分析在一系列实验者、受试者、应激和结果变量方面，将75%的“可靠”实验室与25%的“较不可靠”实验室进行了比较。这两个样本之间几乎没有发现显著差异，可能是因为“较不可靠”样本的规模较小且具有多样性。该研究的另外两个局限性包括（不可避免地）遗漏了可能使用过该模型但未发表数据的实验室，以及使用临时措施来比较不同应激方案的严重程度。结合相关已发表的观察结果对结果进行了讨论。得出的结论是，CMS实际上是一个相当稳健的模型，但导致少数实验室实施效果较差的因素仍有待明确。
