Talk by Thomas R. Knösche, MPI for Human Cognitive and Brain Sciences entitled "Connectivity based cortex parcellation".
Local contact: Tim Dyrby, timd @ drcmr.dk

Abstract: There is ample evidence for the existence of areas on the human cortex that are distinct with respect to certain structural and functional criteria. For example, cytoarchitectonic areas have been defined according to the type, density and arrangement of cells by early [1] as well as modern [2-3] researchers and often appear to bear functional meaning.

However, a closer look reveals that any particular parcellation of the cortical sheet provides an incomplete view on its functional-anatomical organization. First, the parcellation very much depends on the criteria used to define similarity and segregation. These can be, for example, BOLD activation patterns, cytoarchitecture, myeloarchitecture, receptor architecture and connectivity. Second, even within a particular set of criteria, the question remains, how similar two pieces of cortex have to be to belong to the same area. In other words, the parcellation depends on the desired granularity with which we want to describe the structural and/or functional similarity structure of the cortex.

As in modern neuroscience brain function is considered to be supported mainly by networks on different scale levels – from the local circuits within a microcolumn to brain-wide networks mediated by white matter fibers – the connectivity of a particular grey matter unit to the rest of the brain has to be a very important criterion for the organization of the cortex. It has been demonstrated that cortex parcellations based on long-range connectivity computed from diffusion MRI is meaningful and reproducible [4-5], as well as functionally meaningful [5-6]. However, the techniques used in these studies always yield one particular parcellation at a predefined level of granularity (e.g., by stating the number of clusters). In order to completely capture the information content of the similarity structure of long-range connectivity in an orderly and intuitive way, one can use hierarchical clustering techniques [7] and represent their results in form of dendrograms (trees).

In this presentation I will first discuss some fundamental question concerning the relationship between the functional and structural organization of the cortex, the particular role of connectivity in this respect, as well as some principal connectivity based approaches along with their potential role in a general multimodal framework. Next, I will review some studies to illustrate the chances and risks of connectivity based parcellation. Finally, I will present a group of novel hierarchical whole-brain parcellation techniques [8]. In particular I will discuss the issues of tree simplification, identification of relevant parcellations from trees and comparison between trees, and demonstrate the usefulness of hierarchical parcellation with some examples. In conclusion I will argue that hierarchical parcellation is adequate and useful. At the same time I will also provide a critical discussion of some potential problems and limitations. Some of these limitations may be overcome by better imaging techniques [9], others are more fundamental and need to be resolved within multimodal frameworks.

References

1.              Brodmann, K., Vergleichende Lokalisationslehre der Großhirnrinde in ihren Prinzipien dargestellt auf Grund des Zellaufbaues. 1909, Leipzig: Barth.

2.              Amunts, K., et al., Broca’s region revisited: cytoarchitecture and intersubject variability. J Comp Neurol 1999. 412: p. 319-341.

3.              Amunts, K., A. Schleicher, and K. Zilles, Cytoarchitecture of the cerebral cortex--more than localization. NeuroImage, 2007. 37(4): p. 1061-5; discussion 1066-8.

4.              Anwander, A., et al., Connectivity-based parcellation of Broca's area. Cereb Cortex, 2007. 17(4): p. 816-825.

5.              Johansen-Berg, H., et al., Changes in connectivity profiles define functionally distinct regions in human medial frontal cortex. Proc Natl Acad Sci U S A, 2004. 101(36): p. 13335-40.

6.              Schubotz, R.I., et al., Anatomical and functional parcellation of the human lateral premotor cortex. NeuroImage, 2010. 50(2): p. 396-408.

7.              Gorbach, N.S., et al., Hierarchical Information-based Clustering for Connectivity-based Cortex Parcellation. Frontiers in Neuroinformatics, 2011. 5.

8.              Moreno, D., et al., Connectivity-based Whole-Brain Hierarchical Parcellation of the Human Brain, in 17. Annual Meeting of the Organization for Human Brain Mapping. 2011: Quebec, Canada.

9.              Heidemann, R.M., et al., : k-space and q-space: Combining ultrahigh spatial and angular resolution in diffusion imaging using ZOOPPA at 7T. NeuroImage, accepted.