Discrimination surfaces with application to region-specific brain asymmetry analysis
Gabriel Martos
FCEyN-UBA
Discrimination surfaces are here introduced as a diagnostic tool for localizing brain regions where discrimination between diseased and non-diseased subjects is higher. To estimate discrimination surfaces, we introduce a Mann--Whitney type of statistic for stochastic processes, and present large-sample results characterizing its asymptotic behavior. Simulation results demonstrate that our estimator accurately recovers the true surface and corresponding interval of maximal discrimination. The empirical analysis suggests that in the anterior region of the brain, schizophrenia patients tend to present lower local asymmetry scores in comparison to patients in the control group.