This page starts off with documentation on background resources and setup prior to conducting a meta-analysis. These steps primarily include downloading necessary software and preregistration.
Here you will find useful resources for guiding through meta-analyses.
- Relevant papers
- Handbook of research synthesis and meta-analysis (available via UPenn library)
- GitHub Repo following Healthy Pain Meta-analysis
- GitHub Repo following Chronic Pain fMRI Meta-analysis
Before starting the systematic review, do a background search on previous reviews in this topic. You don’t have to try to find every single review that’s been out there (you’ll get a better sense of it as you screen through abstracts), but the most cited reviews are a helpful start.
The most important thing in a meta-analysis is a rigorously defined question and outcome measure of interest. These pieces of information are most helpful to note when defining your question:
- How a condition is defined and examples of it (e.g., chronic pain)
- MRI measure of interest (e.g., BOLD response to a task, ALFF, ReHo, VBM, etc.)
- Experimental paradigm of interest (e.g., acute pain stimulation in a scanner)
- Another great place to look at is any professional society sites that have rigorously defined certain conditions (e.g., pain as defined by IASP).
- MATLAB (for meta-analyses)
- Anatomy Toolbox
- Microsoft Excel
- Recommended: R, FSL (or other analyses software you’re comfortable with)
Steps to set up:
- Create a Covidence review for systematic review.
- Set up a Google Drive project folder (e.g., project folder for structural imaging pain meta-analysis).
- Set up a folder on your local drive with these subfolders (these will be helpful for organization):
After setup, your first task should be to pre-register your systematic review on PROSPERO. PROSPERO isn’t the only site to pre-register your study, but it is the most well-known site for pre-registering a systematic review.
PROSPERO is a site to preregister systematic reviews. However, before preregistration, first create a 1-page meta-analysis protocol using this example to guide how the meta-analysis will be conducted. This will be a reference for the team to use. Part of this involves developing a search strategy. Search strategies from previous meta-analyses will be helpful as a guideline.
After, conduct preliminary literature search in databases to find an optimal # of studies for screening. This number should be in the thousands range (less than 8,500 prior to duplication of records removed would be a good number to aim for). This provides a sense of the scope of the literature. You can also check that you aren’t getting copious records of irrelevant literature (e.g., canine pain when you’re looking for human pain).
Once these details have been sorted, here are the steps to preregistration on PROSPERO:
- In the Google Drive folder, create a draft for PROSPERO pre-registration using this example.
- Author affiliations and e-mail addresses would be needed during this stage.
- After, go to https://www.crd.york.ac.uk/prospero/ to fill in the information and submit. This should be done under Ted’s account.