Random fields of multivariate test statistics, with applications to shape analysis

Abstract

Our data are random fields of multivariate Gaussian observations, and we fit a multivariate linear model with common design matrix at each point. We are interested in detecting those points where some of the coefficients are nonzero using classical multivariate statistics evaluated at each point. The problem is to find the $P$-value of the maximum of such a random field of test statistics. We approximate this by the expected Euler characteristic of the excursion set. Our main result is a very simple method for calculating this, which not only gives us the previous result of Cao and Worsley [Ann. Statist. 27 (1999) 925--942] for Hotelling's $T^2$, but also random fields of Roy's maximum root, maximum canonical correlations [Ann. Appl. Probab. 9 (1999) 1021--1057], multilinear forms [Ann. Statist. 29 (2001) 328--371], ${\bar{\chi }}^2$ [Statist. Probab. Lett 32 (1997) 367--376, Ann. Statist. 25 (1997) 2368--2387] and ${\chi }^2$ scale space [Adv. in Appl. Probab. 33 (2001) 773--793]. The trick involves approaching the problem from the point of view of Roy's union-intersection principle. The results are applied to a problem in shape analysis where we look for brain damage due to nonmissile trauma.

Published in the Annals of Statistics (http://www.imstat.org/aos/) by the Institute of Mathematical Statistics (http://www.imstat.org)

References in corpus (2)

• Validity of the expected Euler characteristic heuristic
• A Gaussian kinematic formula

Cited by in corpus (12)

• A Cramér moderate deviation theorem for Hotelling's $T^2$-statistic with applications to global tests
• Bayesian matching of unlabeled marked point sets using random fields, with an application to molecular alignment
• Testing One Hypothesis Multiple Times: The Multidimensional Case
• Confidence Tubes for Curves on SO(3) and Identification of Subject-Specific Gait Change after Kneeling
• The geometry of least squares in the 21st century
• Tilted Euler characteristic densities for Central Limit random fields, with application to "bubbles"
• Accelerating Permutation Testing in Voxel-wise Analysis through Subspace Tracking: A new plugin for SnPM
• Informative Goodness-of-Fit for Multivariate Distributions
• Testing One Hypothesis Multiple times
• The volume-of-tube method for Gaussian random fields with inhomogeneous variance
• Sparse Network Modeling
• Extremes of Spherical Fractional Brownian Motion