Skip to main content

Meta-Analysis Specification

Now that your studyset's studies have all the necessary information (metadata, coordinates, annotations), you can configure and run a meta-analysis!

A wizard will guide you through the following key steps of meta-analysis specification.

1) Select the Algorithm and Corrector​

Choose the algorithm for your meta-analysis and, optionally, select a statistical correction method (corrector).

Neurosynth-Compose supports several meta-analysis algorithms, powered by NiMARE, a Python library for neuroimaging meta-analysis. For an overview to the coordinate-based meta-analysis methods supported by NiMARE, refer to this guide.

Supported coordinate-based methods include:

  • MKDADensity (default): Multi-Kernel Density Analysis. A popular density-based meta-analysis algorithm. Creates a binary sphere around each coordinate.

  • ALE: Activation Likelihood Estimation, convolves coordinates with a 3D Gaussian distribution.

  • MKDAChi2: Allows you to compare your meta-analysis studies with a larger reference set of studies. It tests whether a higher proportion of studies in your meta-analysis activate a specific voxel compared to a larger population of studies that were not included. For more details, see the MKDAChi2 Association tutorial

    For each algorithm, a default set of arguments is used, but you can easily modify them. For a complete reference on available arguments, see the NiMARE API Docs.

Next, select a method for multiple comparisons correction:

  • FDRCorrector (default): False Discovery Rate correction.
  • FWECorrector: Family-Wise Error Rate correction. We strongly recommend using FWECorrector for publication-quality results with 10,000s iteration if possible. This is computationally intensive however, hence not the default option.

Wizard 1

2) Select Analyses​

Now you will select which studies & analyses you want to include in your meta-analysis. To do so, select the annotation column for inclusion, which determines which analyses are used in the meta-analysis. By default, the "included" column contains all studies and analyses, but you can modify this based on your needs by editing Annotations in the Extraction step.

A preview of the included studies based on the selected annotation column will be displayed.

Wizard 2

3) Meta-Analysis Details​

Enter a name and description for your meta-analysis to help you find it later. A default name will be provided, but you can customize it as needed.

4) Review​

Finally, you can review your selected options before creating your specification.

Wizard 3 Review

5) Finalize​

To complete the specification process, click the "Create Meta-Analysis Specification" button.

Running the Meta-Analysis​

After clicking "Create Meta-Analysis Specification," you will be taken to a page displaying the status of your running analyses. Refer to the next section of the documentation for details on running and monitoring your meta-analysis.