Nonparametric Analysis of Statistic Images from Functional Mapping Experiments

Abstract

The analysis of functional mapping experiments in positron emission tomography involves the formation of images displaying the values of a suitable statistic, summarising the evidence in the data for a particular effect at each voxel. These statistic images must then be scrutinised to locate regions showing statistically significant effects. The methods most commonly used are parametric, assuming a particular form of probability distribution for the voxel values in the statistic image. Scientific hypotheses, formulated in terms of parameters describing these distributions, are then tested on the basis of the assumptions. Images of statistics are usually considered as lattice representations of continuous random fields. These are more amenable to statistical analysis. There are various shortcomings associated with these methods of analysis. The many assumptions and approximations involved may not be true. The low numbers of subjects and scans, in typical experiments, lead to noisy statistic images with low degrees of freedom, which are not well approximated by continuous random fields. Thus, the methods are only approximately valid at best and are most suspect in single-subject studies. In contrast to the existing methods, we present a nonparametric approach to significance testing for statistic images from activation studies. Formal assumptions are replaced by a computationally expensive approach. In a simple rest-activation study, if there is really no activation effect, the labelling of the scans as "active" or "rest" is artificial, and a statistic image formed with some other labelling is as likely as the observed one. Thus, considering all possible relabellings, a p value can be computed for any suitable statistic describing the statistic image. Consideration of the maximal statistic leads to a simple nonparametric single-threshold test. This randomisation test relies only on minimal assumptions about the design of the experiment, is (almost) exact, with Type I error (almost) exactly that specified, and hence is always valid. The absence of distributional assumptions permits the consideration of a wide range of test statistics, for instance, "pseudo" t statistic images formed with smoothed variance images. The approach presented extends easily to other paradigms, permitting nonparametric analysis of most functional mapping experiments. When the assumptions of the parametric methods are true, these new nonparametric methods, at worst, provide for their validation. When the assumptions of the parametric methods are dubious, the nonparametric methods provide the only analysis that can be guaranteed valid and exact.

A statistical framework for neuroimaging data analysis based on mutual information estimated via a gaussian copula

“…4A). A common approach with neuroimaging studies is to apply a permutation test together with the method of maximum statistics in order to correct for multiple comparisons [Holmes et al, 1996; Nichols and Holmes, 2002]. We calculated the two‐sample Kolmogorov–Smirnov (KS) test statistic [Massey, 1951] for each time point and sensor from all the available 1000 trials and used this as the ground‐truth to evaluate tests performed with smaller numbers of trials (Fig.…”

“…4A). Multiple‐comparison correction is required over time points and sensors/sources: this can be achieved using permutation testing (repeating the calculation with shuffled stimulus values) combined with the method of maximum statistics [Holmes et al, 1996; Nichols and Holmes, 2002], or cluster sum statistics [Maris and Oostenveld, 2007] possibly with threshold‐free cluster enhancement [Pernet et al, 2015; Smith and Nichols, 2009]. An advantage of this design is that, because each time point is analyzed separately, there is no assumption that the signal is stationary.…”

Progressive brain changes in patients with chronic fatigue syndrome: A longitudinal MRI study

“…Fig. S1).DARTEL SnPM is the same as option 2 except that a nonparametric paired t‐test23 was used in step (d) (Supp. Fig.…”

A non-parametric approach for co-analysis of multi-modal brain imaging data: Application to Alzheimer's disease

“…Since the voxel value distribution of a combining function image is often unknown, we employ a permutation test for statistical inference (Holmes et al, 1996;Nichols and Holmes, 2002). Furthermore, when a combining function and a permutation test are used together, the correlation between t images is implicitly accounted for (Hayasaka and Nichols, 2004;Pesarin, 2001).…”

“…Such statistic -local maximum, cluster size, or some other statistics -is used in the subsequent test as a test statistic, rather than each voxel value W(v). Only the largest of such test statistics is recorded in each permutation, in order to control the family-wise error (FWE) rate to correct for multiple comparisons (Hayasaka and Nichols, 2003;Holmes et al, 1996). The entire process is repeated for a sufficient number of times, typically between 1000 and 3000 permutations for sufficient confidence in the permutation distribution, each with a different permutation of As and Bs.…”