Functional volumes modeling: Scaling for group size in averaged images

Publisher Website
Google Scholar
Abstract
Functional volumes modeling (FVM) is a statistical construct for metanalytic modeling of the locations of brain functional areas as spatial probability distributions. FV models have a variety of applications, in particular, to serve as spatially explicit predictions of the Talairach‐space locations of functional activations, thereby allowing voxel‐based analyses to be hypothesis testing rather than hypothesis generating. As image averaging is often applied in the analysis of functional images, an important feature of FVM is that a model can be scaled to accommodate any degree of intersubject image averaging in the data set to which the model is applied. In this report, the group‐size scaling properties of FVM were tested. This was done by: (1) scaling a previously constructed FV model of the mouth representation of primary motor cortex (M1‐mouth) to accommodate various degrees of averaging (number of subjects per image = n = 1, 2, 5, 10), and (2) comparing FVM‐predicted spatial probability contours to location‐distributions observed in averaged images of varying n composed from randomly sampling a 30‐subject validation data set. Hum. Brain Mapping 8:143–150, 1999.