Related Experiment Video
Updated: May 4, 2026
Using Informational Connectivity to Measure the Synchronous Emergence of fMRI Multi-voxel Information Across Time
Published on: July 1, 2014
A novel meta-analytic approach: mining frequent co-activation patterns in neuroimaging databases
Julian Caspers1, Karl Zilles2, Christoph Beierle3
1Institute of Neuroscience and Medicine (INM-1), Research Centre Jülich, 52425 Jülich, Germany; Department of Diagnostic and Interventional Radiology, University Dusseldorf, Medical Faculty, D-40225 Dusseldorf, Germany.
We introduce PaMiNI, a novel method for neuroimaging meta-analysis. PaMiNI identifies frequent co-activation patterns, overcoming limitations of existing techniques like Activation Likelihood Estimation (ALE) and Multilevel Kernel Density Analysis (MKDA).
Area of Science:
- Neuroimaging
- Cognitive Neuroscience
- Data Mining
Background:
- Coordinate-based meta-analyses are crucial for studying brain co-activity using large databases like BrainMap.
- Existing methods (ALE, MKDA) identify convergent activity but miss within-experiment co-occurrence patterns.
- This limits the accurate evaluation of co-activation patterns in neuroimaging research.
Purpose of the Study:
- To introduce PaMiNI, a novel meta-analytic approach for data-driven detection of frequent co-activation patterns in neuroimaging datasets.
- To overcome the limitations of previous methods that do not account for within-experiment co-occurrence.
- To provide a robust and publicly available tool for enhanced neuroimaging meta-analysis.
Main Methods:
- PaMiNI combines Gaussian mixture modeling and the Apriori algorithm, established data-mining techniques.
- The approach enables the detection of frequent co-activation patterns within neuroimaging datasets.
- Feasibility was demonstrated using simulated data and a real-world working memory experiment dataset.
Main Results:
- Simulated data analysis confirmed PaMiNI's ability to identify activation foci and separate co-activation patterns, even with close or non-Gaussian foci.
- Analysis of a working memory dataset revealed a fronto-parietal core network with left-lateralization.
- PaMiNI successfully cross-validated findings with a previous Activation Likelihood Estimation (ALE) meta-analysis.
Conclusions:
- PaMiNI offers a significant advancement in coordinate-based meta-analysis by detecting frequent co-activation patterns.
- The method demonstrates robustness and accuracy in identifying brain networks, as shown in working memory studies.
- PaMiNI is implemented in a publicly available software system to promote widespread adoption and further research.
