Elements of functional neuroimaging
Citation: Wager, T. D., Hernandez, L., Jonides, J., & Lindquist, M. (2007). Elements of functional neuroimaging. In J. T. Cacioppo, L. G. Tassinary & G. G. Berntson (Eds.), Handbook of Psychophysiology (4th ed., pp. 19-55). Cambridge: Cambridge University Press.
There has been explosive interest in the use of brain imaging to study cognitive and
affective processes in recent years. Examine Figure 1, for example, to see the dramatic rise in
numbers of publications using positron emission tomography (PET) and functional Magnetic
Resonance Imaging (fMRI) from 1985 to 2004. A recent surge in integrative empirical work that
combines data from human performance, neuroimaging, neuropsychology, and psychophysiology
provides a more comprehensive, but more complex, view of the human brain-mind than ever
before. Because the palette of evidence from which researchers draw is larger, there is an
increasing need to for cross-disciplinary integration and education. Our goal in this chapter is to
provide an introduction to the growing field of neuroimaging research, including a brief survey of
important issues and new directions.
The many aspects of PET and fMRI methodology are organized here into three sections
that describe the physical, social, and inferential contexts in which imaging studies are conducted.
The first section covers the physical basis of PET and fMRI imaging. This section describes the
physics of each technique, what each measures, aspects of data processing, current limits of
resolution, and a comparison of the relative advantages and disadvantages of these two
techniques. The second section concerns aspects of neuroimaging related to social issues. In this
section, we explore the kinds of questions that might be fruitfully addressed using imaging,
human factors considerations when designing imaging studies, and a ?road map? of an imaging
experiment. The third section deals with inference in neuroimaging. It contains a review of types
of experimental designs, analysis strategies and statistics, and localization of neuroimaging
results in the brain. The statistical part of the section reviews the General Linear Model (GLM;
the most commonly used analysis framework), hierarchical and robust extensions to the GLM
that are increasingly applied to neuroimaging data, and the most commonly used multivariate
analyses. In this section we also address group analyses and multiple comparisons, and pitfalls in
the use of the various analysis techniques.
Although we review the physics underlying PET and fMRI here, we would like to
emphasize that much of the material in the remainder of the chapter can stand on its own; the
reader need not have a thorough grasp of the physics before proceeding to other sections. An
outline of the major topics covered is as follows: