Parcellation of the cingulate cortex at rest and during tasks: a meta-analytic clustering and experimental study

Front Hum Neurosci. 2013 Jun 14;7:275. doi: 10.3389/fnhum.2013.00275. eCollection 2013.


Anatomical, morphological, and histological data have consistently shown that the cingulate cortex can be divided into four main regions. However, less is known about parcellations of the cingulate cortex when involved in active tasks. Here, we aimed at comparing how the pattern of clusterization of the cingulate cortex changes across different levels of task complexity. We parcellated the cingulate cortex using the results of a meta-analytic study and of three experimental studies. The experimental studies, which included two active tasks and a resting state protocol, were used to control the results obtained with the meta-analytic parcellation. We explored the meta-analytic parcellation by applying a meta-analytic clustering (MaC) to papers retrieved from the BrainMap database. The MaC is a meta-analytic connectivity driven parcellation technique recently developed by our group which allowed us to parcellate the cingulate cortex on the basis of its pattern of co-activations during active tasks. The MaC results indicated that the cingulate cortex can be parcellated into three clusters. These clusters covered different percentages of the cingulate parenchyma and had a different density of foci, with the first cluster being more densely connected. The control experiments showed different clusterization results, suggesting that the co-activations of the cingulate cortex are highly dependent on the task that is tested. Our results highlight the importance of the cingulate cortex as a hub, which modifies its pattern of co-activations depending on the task requests and on the level of task complexity. The neurobiological meaning of these results is discussed.

Keywords: activation likelihood estimation; data driven parcellation; hierarchical clustering; k-means clustering; meta-analytic clustering; meta-analytic connectivity modeling; voronoi parcellation; voxel-based meta-analysis.