The Online Brain Atlas Reconciliation Tool

The Online Brain Atlas Reconciliation Tool is an NIMH-sponsored project that aims to provide a quantitative solution to the so-called neuroanatomical nomenclature problem by comparing overlap relations between regions defined as spatial entities in different brain atlases.

At present we have released a prototype web tool that presents initial project results.  Specifically, we have developed an analytical framework that quantifies the similarities and differences between different human MRI-based partitioning schemes, which will enable meaningful multi-atlas meta-analysis of clinical and scientific questions.  These methods offer a quantitative answer to the “nomenclature problem” in neuroscience by comparing brain parts on the basis of their geometrical definitions rather than on the basis of name alone.   Thus far our tools have been used to quantitatively compare eight distinct parcellations of the International Consortium for Brain Mapping (ICBM) single-subject template brain, each created using existing atlasing methods.  We provide measures of global and regional similarity, and offer visualization techniques that allow users to quickly identify the correspondences (or lack of correspondences) between regions defined by different atlases.  By making these methods interactive and putting them online we hope to facilitate a better understanding, particularly within the brain imaging community, of the relationships between different atlases.

One important use for such a tool will be to provide a single place where researchers can compare existing atlases, and we will encourage users to add parcellations created by other atlasing methods, automated labeling tools, and particularly by expert manual delineation of regions.  This will both expand the power of the analysis itself and provide a valuable resource simply by supplying a set of labeled brains.  Further, the tool will help to improve meta-analyses of results in online databases and will help researchers guide their everyday reviews of existing literature by finding alternative labels for the same or nearby brain regions.  This effort will be made more powerful by our proposal to integrate the quantitative results from our spatial analysis with existing ontological efforts for neuroanatomy, and by integrating the resources within the existing NIF, BIRN and CaBIG infrastructures.

Additional information about the underlying analysis framework of the Online Brain Atlas Reconciliation Tool can be found in Bohland et al. (2009).  The brain atlas concordance problem: quantitative comparison of anatomical parcellations.  PLoS ONE.  4(9): e7200.