Meta-Analysis

Autonima can run coordinate-based meta-analysis on the NiMADS outputs produced by the pipeline.

Install Prerequisites

pip install -e .[meta]

Input Folder

autonima meta expects the folder that directly contains:

• nimads_studyset.json
• nimads_annotation.json

For standard pipeline runs, that is usually:

<output-folder>/outputs

Basic Usage

autonima meta projects/cue_reactivity/default/outputs

Optional Parameters

• --estimator Choices: ale, mkdadensity, kda
• --estimator-args JSON string of estimator arguments
• --corrector Choices: fdr, montecarlo, bonferroni
• --corrector-args JSON string of corrector arguments
• --include-ids Newline-delimited study IDs or PMIDs to include

Example:

autonima meta projects/cue_reactivity/default/outputs \
  --estimator ale \
  --corrector montecarlo

Output Layout

Meta-analysis results are written under:

<nimads-input-folder>/meta_analysis_results/

This directory contains:

• one folder per boolean annotation column
• NiMARE reports and statistical images
• summary artifacts for each analysis set

Common Failure Modes

• NiMARE is not installed
• the input folder does not contain both NiMADS JSON files
• JSON argument strings are invalid
• include-ID filtering removes all analyses for a target annotation