2 - Experimental Design and Statistical Issues

Abstract

This chapter focuses on the experimental design and analysis issues that arise when using functional neuroimaging. These issues comprise those that are generic to all functional neuroimaging and some considerations that are specific to modeling the effects of pathophysiology. An important issue in design and analysis is the relationship between the neurobiological hypotheses and conceptual models of pathophysiology and how this relationship is realized in terms of the statistical models used to analyze neuroimaging time series. Functional mapping studies are usually analyzed with some form of statistical parametric mapping which refers to the construction of spatially extended statistical processes to test hypotheses about regionally specific effects. Statistical parametric maps (SPMs) are image processes with voxel values that are, under the null hypothesis, distributed according to a known probability density function. The success of SPM is largely due to the simplicity of the idea. One analyzes each and every voxel using any standard statistical test. The analysis of functional neuroimaging data involves many steps that can be broadly divided into (i) spatial preprocessing, (ii) estimating the parameters of a statistical model, and (iii) making inferences about modeled effects using the parameter estimates and their associated statistics.