Fig. 1

FDR vs. Power decision plot for single measurement simulation. The alternative hypothesis is \(H_{1,j}: \left| \varvec{L}^T_{10\%, j}\varvec{\beta }\right| =5\). The solid curve is the estimated FDR based on Eq. (11) averaged over 300 simulation replicates, and the dashed curve is the empirical true FDR calculated by averaging the proportions of false discoveries \(\frac{\varvec{V}(\phi )}{\varvec{R}(\phi )}\) over 300 simulation replicates. The black horizontal dot-dash line represents the empirical true FDR calculated by averaging the proportions of false discoveries over 300 simulation replicates when using \(\alpha =0.05\) as the significance level. The solid and dashed curves are overlapped on the top panel