Skip to content

Reading list: Meta analysis

Taylor Salo edited this page Apr 22, 2018 · 4 revisions

A list of papers about meta-analysis and meta-analysis-related analyses in neuroimaging

Introduction to meta-analysis

Neuroimaging meta-analysis

Image-based meta-analysis

Coordinate-based meta-analysis

The most readily available results from fMRI studies are coordinates- locations of peak activations reported in a standard brain space. While coordinates are less informative than unthresholded statistical images, the logic of coordinate-based meta-analyses is that, with enough studies, peak coordinates in regions related to an experimental manipulation will appear at a detectably higher rate than ones in unrelated regions (i.e., false positives).

Functional characterization analysis

Functional characterization analysis, also known as open-ended functional decoding or reverse inference, refers to a range of methods which attempt to predict mental states, phenotypes, or diagnoses from neural states (e.g., fMRI results). A separate collection of methods, often referred to simply as "decoding," attempt to perform a similar task, but generally use classifiers on within-subject data to predict from a restricted number of mental states or phenotypes. Conversely, functional characterization analysis uses meta-analytic databases, which ostensibly contain a wider range of mental and neural states than can be acquired within a given experiment, in order to predict across the entire range of potential mental states. As such, what functional characterization analysis loses in accuracy (as compared to decoding), it makes up for in generalizability.