Introducing alternative-based hypothesis testing for defining functional regions of interest in fMRI

In fMRI research, one often aims to examine activation in specific functional regions of interest (fROIs). Current statistical methods tend to localize fROIs inconsistently, only avoiding detection of false activation. Avoiding to miss true activation is however equally important in this context. We discuss the alternative-based hypothesis testing (ABHT) procedure, where evidence against the null hypothesis of no effect in a p0-value and evidence against a pre-specified alternative hypothesis in a p1-value is measured. A threshold is then imposed on each p-value to control both false positives and false negatives directly. The procedure was validated and compared to classic null hypothesis testing (NHT) on simulated brain images and was applied to a real localizer data set of 13 subjects. Finally, we examined the consistency of the results between different runs in the same subject with data from the Human Connectome Project. The ABHT method outperformed classic NHT by disregarding voxels that showed evidence against both the null and alternative and by including voxels for which the alternative hypothesis cannot be rejected despite lack of evidence against the null.