Content area
Abstract
Good research requires, among other virtues, (i) methods that yield stable experimental observations without arbitrary (post hoc) assumptions, (ii) logical interpretations of the sources of observations, and (iii) sound inferences to general causal mechanisms explaining experimental results by placing them in larger explanatory contexts. In The New Phrenology, William Uttal examines the research tradition of localization, and finds it deficient in all three virtues, whether based on lesion studies or on new technologies for functional brain imaging. In this paper I consider just the arguments concerning brain imaging, especially functional Magnetic Resonance Imaging. I think that Uttal is too harsh in his methodological critique, but correct in his assessment of the conceptual limitations of localist evidence. I propose instead a data-driven test for assessing relative modularity in brain images, and show its use in a secondary analysis of fMRI data from the National fMRI Data Center (www.fmridc.org). Although the analysis is a limited pilot study, it offers additional empirical challenge to localism.