Skip to Main Content
Large-scale neuroimaging databases provide a rich fundus of functional neuroimaging experiments exhibiting maximum activation coordinates for specific task conditions. Aiming to explore major neuronal networks of the human brain, we developed a meta-analytic pattern-mining approach which combines Gaussian mixture modeling with the Apriori algorithm to identify frequent activation patterns within these databases. The approach has been implemented in the PaMiNI (Pattern Mining in NeuroImaging) system, providing manifold facilities for the finding, inspection, and analysis of relevant patterns. After briefly sketching the background of PaMiNI, we give an overview of the system and describe its architecture. Using an example application, a system walkthrough illustrates how PaMiNI can be used for the discovery of networks comprising functionally connected brain regions.