Analysis of scalp voltage asymmetries using Hotelling's T2 methodology.

Topographic mapping procedures are becoming increasingly popular, and scalp voltage asymmetries are a measure of specific interest. Standard statistical techniques in the field to detect voltage asymmetries, "significance probability mapping (SPM)" and repeated-measures analysis of variance (ANOVA) have statistical assumptions which are not in accord with electrophysiological data, and which therefore may lead to serious statistical errors. These limitations and a superior method using multivariate analysis are discussed. This paper provides a sample data set, for illustration of specific computational steps, and to provide an empirical example and reference for the theoretical discussion.