Spatial extent inference (SEI) is widely used across neuroimaging modalities to study brain-behavior associations that inform our understanding of disease. Recent studies have shown that Gaussian random field (GRF) based tools can have hugely inflated family-wise error rates (FWERs). We developed a semiparametric bootstrap joint (sPBJ) testing procedure to address the limitations of available methods; our approach is the image equivalent of a robust sandwich covariance estimator and yields consistent estimates of standard errors, even under model misspecification.
Speaker: Dr. Simon Vandekar, Assistant Professor, Vanderbilt Department of Biostatistics