Hierarchical information-based clustering for connectivity-based cortex parcellation

Abstract

One of the most promising avenues for compiling connectivity data originates from the notion that individual brain regions maintain individual connectivity profiles; the functional repertoire of a cortical area (“the functional fingerprint”) is closely related to its anatomical connections (“the connectional fingerprint”) and, hence, a segregated cortical area may be characterized by a highly coherent connectivity pattern. Diffusion tractography can be used to identify borders between such cortical areas. Each cortical area is defined based upon a unique probabilistic tractogram and such a tractogram is representative of a group of tractograms, thereby forming the cortical area. The underlying methodology is called connectivity-based cortex parcellation and requires clustering or grouping of similar diffusion tractograms. Despite the relative success of this technique in producing anatomically sensible results, existing clustering techniques in the context of connectivity-based parcellation typically depend on several non-trivial assumptions. In this paper, we embody an unsupervised hierarchical information-based framework to clustering probabilistic tractograms that avoids many drawbacks offered by previous methods. Cortex parcellation of the inferior frontal gyrus together with the precentral gyrus demonstrates a proof of concept of the proposed method: The automatic parcellation reveals cortical subunits consistent with cytoarchitectonic maps and previous studies including connectivity-based parcellation. Further insight into the hierarchically modular architecture of cortical subunits is given by revealing coarser cortical structures that differentiate between primary as well as premotoric areas and those associated with pre-frontal areas.

Brain parcellation based on information theory

“…The main issues with these methods are the definition of the number of clusters a priori and the reliance on initial random sampling, as it has been shown that iterative repetitions of the same method may lead to different results [9]. To overcome these limitations, and assuming that brain networks have hierarchical properties [10,11], several hierarchical 25 clustering methods that compute a parcellation at each level in the hierarchy have been proposed [12,13,14,15]. These methods obtain brain parcellations at multiple granularities without the need to define the number of clusters.…”

“…Gorbach et al [13] proposed a hierarchical method 30 that clusters voxels using the mutual information between tractograms, and obtained promising results for specific regions of the brain. The use of mutual information as a similarity measure is, therefore, an effective solution to group voxels.…”

“…The use of mutual information as a similarity measure is, therefore, an effective solution to group voxels. However, Gorbach et al's method [13] assumes that the whole cluster can be represented by only one tractogram. …”

Mapping the connectome: Multi-level analysis of brain connectivity

“…This Research Topic of Frontiers in Neuroinformatics, dedicated to the memory of Rolf Kötter (1961Kötter ( -2010 and his pioneering work in the field of brain connectomics, comprises contributions that elucidate different levels of connectivity analysis (from MRIbased methods, through axonal tracing techniques, to mapping of functional connectivity in relation to detailed 3-D reconstructions Gorbach et al, 2011;Yendiki et al, 2011] to ex vivo mapping of detailed fiber architectures [(C,e) Axer et al, 2011b; (e,F) Annese, 2012]. (g-H) Novel experimental methods in animal models include combined optogenetic and functional MRI mapping of specific connections (g) (Lee, 2011) (Borisyuk et al, 2011).…”

“…Knowledge about the specific hodological organization of different brain regions may thus predict the various functional properties of such regions. Gorbach et al (2011) explore relationships between the functional and connectional "fingerprints" of cerebrocortical areas in the human brain, by using hierarchical information-based clustering of MRI-based connectivity measures. They propose an automated hierarchical parcelation approach to identify cortical subunits that are consistent with cytoarchitectonic maps and previous connectivity-based parcelation schemes ( Figure 1K).…”