ROI analysis of pharmafMRI data: an adaptive approach for global testing

Pharmacological fMRI (pharmafMRI) is a new highly innovative technique utilizing the power of functional Magnetic Resonance Imaging (fMRI) to study drug induced modulations of brain activity. FMRI recordings are very informative surrogate measures for brain activity but still very expensive and therefore pharmafMRI studies have typically small sample sizes. The high dimensionality of fMRI data and the arising high complexity requires sensitive statistical analysis in which often dimensionality reductions are crucial. We consider Region of Interest (ROI) analysis and propose an adaptive two-stage testing procedure for respectively formulating and testing the fundamental hypothesis as to whether the drug modulates the control brain activity in selected ROI. The proposed tests are proved to control the type I error rate and they are optimal in terms of the predicted chance of a true positive result at the end of the trial. Power analysis is performed by re-expressing the high dimensional domain of power function into a lower dimensional easily interpretable space which still gives a complete description of the power. Based on these results, we show under which circumstances our procedure outperforms standard single-stage and sequential two-stage procedures focusing on the small sample sizes typical in pharmafMRI. We also apply our methods to ROI data of a pharmafMRI study.