Statistical testing and power analysis for brain-wide association study

Abstract

The identification of connexel-wise associations, which involves examining functional connectivities between pairwise voxels across the whole brain, is both statistically and computationally challenging. Although such a connexel-wise methodology has recently been adopted by brain-wide association studies (BWAS) to identify connectivity changes in several mental disorders, such as schizophrenia, autism and depression [Cheng et al., 2015a[Cheng et al., ,b, 2016, the multiple correction and power analysis methods designed specifically for connexel-wise analysis are still lacking. Therefore, we herein report the development of a rigorous statistical framework for connexel-wise significance testing based on the Gaussian random field theory. It includes controlling the family-wise error rate (FWER) of multiple hypothesis testings using topological inference methods, and calculating power and sample size for a connexel-wise study. Our theoretical framework can control the false-positive rate accurately, as validated empirically using two resting-state fMRI datasets. Compared with Bonferroni correction and false discovery rate (FDR), it can reduce false-positive rate and increase statistical power by appropriately utilizing the spatial information of fMRI data. Importantly, our method considerably reduces the computational complexity of a permutation-or simulation-based approach, thus, it can efficiently tackle large datasets with ultra-high resolution images. The utility of our method is shown in a case-control study. Our approach can identify altered functional connectivities in a major depression disorder dataset, whereas existing methods failed. A software package is available at https://github.com/weikanggong/BWAS.

Assessing Overall Reproducibility for Large-scale High-throughput MRI-based Association Studies

“…M otivated by the Dice similarity coefficient that divides z-scores into three parts, we consider three categories for the underlying/unobserved association analysis related true status: positive change/correlation, no-change/correlation and negative change/correlation. Instead of thresholding, we consider a three-component normal-mixture model for the joint distribution of paired z-scores (2) ] (see above for z-score calculation).…”

“…As an application of M 2 For each list of paired z-scores, we applied M 2 RI to assess the related reproducibility (see Table 3 for results). For PPM I and UK Biobank data sets, the median reproducibility was 0.99993 with the range of interquartile (Q1-Q3) 0.99964-0.99998.…”

“…This index is termed the mixture model reproducibility index (M 2 RI). To evaluate M 2 RI, we present the analysis results from simulations based on sM RI data in the UK Biobank (17,18). Then, we present a comprehensive reproducibility assessment analysis of gray-matter-related human behaviors, brain task state activation and connectivity -behavior analysis in which the reproducibility rates are known.…”

“…. M2 RI was compared to Dice similarity coefficient (DSC) in three simulation analysis scenarios and four M RI association analysis scenarios. For each analysis scenario, the true reproducibility index is shown in the table.…”

“…For each M RI association analysis scenario, the median and 95% confidence intervals (CIs) were obtained based on bootstrapping the paired z-scores. The asymptotic theoretical 95% CIs for M2 RI was also presented. DSC: Dice similarity coefficient.…”