Oral and Maxillofacial Radiology, Applied Oral Sciences and Community Dental Care, Faculty of Dentistry, The University of Hong Kong, Hong Kong, China.
Oral and Maxillofacial Surgery, Faculty of Dentistry, The University of Hong Kong, Hong Kong, China.
Neuroimage. 2019 Nov 15;202:116111. doi: 10.1016/j.neuroimage.2019.116111. Epub 2019 Aug 22.
Multiple neuroimaging meta-analyses have been published concerning gustation, food and taste. A meta-evaluation of these meta-analyses was conducted to qualitatively evaluate the presented evidence. A systematic search was done using multiple databases, in which no restriction was placed on participants and nature of interventions (stimuli vs control). Twenty-three meta-analyses were identified and analyzed. All of them have met 4-9 criteria, out of 11, from the modified checklist constructed by Müller et al. (2018), which implied moderate to high quality of evidence. One of the concerns we found was that no meta-analysis surveyed had been explicitly pre-registered. Also, only three meta-analyses (13.0%) provided clear explanation of how they accounted for sample overlap. Only six meta-analyses (26.1%) explicitly described how they double checked the data. Only two of the 20 meta-analyses (10.0%) using GingerALE software used both the debugged version (v2.3.6) as well as the recommended cluster-level inference with familywise error rate correction. Overall, meta-analyses are increasingly adopting more stringent statistical thresholds, but unfortunately not larger number of studies contained in the analyses.
已经发表了多篇关于味觉、食物和味道的神经影像学荟萃分析。对这些荟萃分析进行了元评估，以定性评估所呈现的证据。使用多个数据库进行了系统搜索，对参与者和干预措施的性质（刺激与对照）没有限制。确定并分析了 23 项荟萃分析。根据 Müller 等人（2018 年）构建的修改后的清单，所有这些都符合 11 个标准中的 4-9 个标准，这意味着证据质量中等偏高。我们发现的一个问题是，没有一项荟萃分析被明确预先注册。此外，只有三项荟萃分析（13.0%）清楚地解释了他们如何解释样本重叠。只有六项荟萃分析（26.1%）明确说明了他们如何仔细检查数据。在使用 GingerALE 软件的 20 项荟萃分析中，只有两项（10.0%）同时使用了经过调试的版本（v2.3.6）和推荐的具有家族错误率校正的簇级推断。总体而言，荟萃分析越来越多地采用更严格的统计阈值，但不幸的是，分析中包含的研究数量并没有增加。
