The relation between statistical power and inference in fMRI.
The relation between statistical power and inference in fMRI.
复制标题
DOI：
10.1371/journal.pone.0184923
复制
发表时间：
2017
期刊：
影响因子：
3.7
通讯作者：
Yarkoni T
中科院分区：
文献类型：
--
作者：
Cremers HR;Wager TD;Yarkoni T
即使所有其他的实验考虑都无可挑剔地解决了，统计上的不足也会导致实验失败。在功能磁共振成像中，大量的因变量，相对较少的观察（受试者），以及需要对多重比较进行校正的组合，可能会显著降低统计能力。这个问题已经得到了明确的解决，但仍然存在争议——特别是在fMRI的预期效应大小方面，特别是在受试者之间的效应，如群体比较和大脑行为相关性方面。我们的目的是通过考虑和对比两种可能的大脑行为相关性的模拟情景：弱扩散效应和强局部效应，来澄清功率问题。从这些场景中抽样表明，特别是在弱扩散场景中，常见样本量（n = 20-30）显示出极低的统计能力，难以代表整个样本中的实际效果，并且在后续重复中显示出很大的变化。来自人类连接组计划的经验数据更类似于弱扩散情景，而不是局部强情景，这强调了许多研究中功率问题的程度。功率问题的可能解决方案包括增加样本量，使用不那么严格的阈值，或者关注感兴趣的区域。然而，这些方法并不总是可行的，有些还存在重大缺陷。可能有助于解决权力问题的最突出的解决方案包括基于模型的（多变量）预测方法和具有相关面向综合方法的元分析。
Statistically underpowered studies can result in experimental failure even when all other experimental considerations have been addressed impeccably. In fMRI the combination of a large number of dependent variables, a relatively small number of observations (subjects), and a need to correct for multiple comparisons can decrease statistical power dramatically. This problem has been clearly addressed yet remains controversial—especially in regards to the expected effect sizes in fMRI, and especially for between-subjects effects such as group comparisons and brain-behavior correlations. We aimed to clarify the power problem by considering and contrasting two simulated scenarios of such possible brain-behavior correlations: weak diffuse effects and strong localized effects. Sampling from these scenarios shows that, particularly in the weak diffuse scenario, common sample sizes (n = 20–30) display extremely low statistical power, poorly represent the actual effects in the full sample, and show large variation on subsequent replications. Empirical data from the Human Connectome Project resembles the weak diffuse scenario much more than the localized strong scenario, which underscores the extent of the power problem for many studies. Possible solutions to the power problem include increasing the sample size, using less stringent thresholds, or focusing on a region-of-interest. However, these approaches are not always feasible and some have major drawbacks. The most prominent solutions that may help address the power problem include model-based (multivariate) prediction methods and meta-analyses with related synthesis-oriented approaches.
登录
查看更多内容
影响因子：
5.7
作者：
Bartra, Oscar;McGuire, Joseph T.;Kable, Joseph W.
通讯作者：
Kable, Joseph W.
影响因子：
6.9
作者：
Cremers HR;Veer IM;Spinhoven P;Rombouts SA;Yarkoni T;Wager TD;Roelofs K
通讯作者：
Roelofs K
影响因子：
4.2
作者：
Bennett, Craig M.;Wolford, George L.;Miller, Michael B.
通讯作者：
Miller, Michael B.
影响因子：
5.7
作者：
Dudoit, S;Shaffer, JP;Boldrick, JC
通讯作者：
Boldrick, JC
影响因子：
9.8
作者：
Chang LJ;Gianaros PJ;Manuck SB;Krishnan A;Wager TD
通讯作者：
Wager TD