Studies associated with pain
Term-based meta-analyses: Frequently Asked Questions
This page displays information for an automated Neurosynth meta-analysis of the term "pain". The meta-analysis was performed by automatically identifying all studies in the Neurosynth database that loaded highly on the term, and then performing meta-analyses to identify brain regions that were consistently or preferentially reported in the tables of those studies.
What do the "forward inference" and "reverse inference" maps mean?
For a detailed explanation, please see our Yarkoni et al (2011). In brief, the forward inference map displays brain regions that are consistently active in studies that load highly on the term "pain". Regions with large z-scores are reported more often than one would expect them to be if activation anywhere in the brain was equally likely. Note that this is typically not so interesting, because it turns out that some brain regions are consistently reported in a lot of different kinds of studies (again, see our paper). So as a general rule of thumb, we don't recommend paying much attention to forward inference maps.
Reverse inference maps are, roughly, maps displaying brain regions that are preferentially related to the term pain. The reverse inference map for pain displays regions that are reported more often in articles that include the term pain in their abstracts than articles that do not. Most of the time this a much more useful way of thinking about things, since reverse inference maps tell you, in some sense, which brain regions are more diagnostic of the term in question, and not just which regions are consistently activated in studies associated with that term.
How do you determine which studies to include in an analysis?
For all term-based meta-analyses visible on this website, we consider a study to load on a given term if the term is used at least once anywhere in the article abstract. We have applied various other modeling approaches in the past (e.g., increasing the cut-off, using continuous-valued weights, and using the full-text of articles rather than just the abstract), but there is generally surprisingly little effect on results within a fairly broad range of parameter variation.
Are these maps corrected for multiple comparisons?
Yes, they're corrected using a false discovery rate (FDR) approach, with an expected FDR of 0.01.
I need more details! How exactly were these maps and data generated?
If you want to know exactly how things work, we encourage you to clone the Neurosynth python tools from the github repository and work through some of the examples and code provided in the package. Everything you see on this page was generated using the default processing stream, so you should be able to easily generate the exact same images (unless the underlying database has grown or changed) for yourself.