78 CATALYZING INQUIRY
terminology, particularly neuroanatomical nomenclature, is vast, nonstandard, and confusing. Ana-
tomical entities may have multiple names (e.g., caudate nucleus, nucleus caudates), the same term may
have multiple meanings (e.g., spine [spinal cord] versus spine [dendritic spine]), and worst of all, the
same term may be defined differently by different scientists (e.g., basal ganglia). To minimize semantic
confusion and to situate cellular and subcellular data from the CCDB in a larger context, the CCDB is
mapped to several shared knowledge sources in the form of ontologies.
Concepts in the CCDB are being mapped to the Unified Medical Language System (UMLS), a large
metathesaurus and knowledge source for the biomedical sciences.^51 The UMLS assigns each concept in
the ontology a unique identifier (ID); thus, all synonymous terms can then be assigned the same ID. For
example, the UMLS ID number for the synonymous terms Purkinje cell, cerebellar Purkinje cell, and
Purkinje’s corpuscle is C0034143. Thus, regardless of which term is preferred by a given individual, if
they share the same ID, they are asserted to be the same. Conversely, even if two terms share the same
name, they are distinguishable by their unique IDs. In the example given above, spine (spinal cord) =
C0037949, whereas spine (dendritic spine) = C0872341.
In addition, an ontology can support the linkage of concepts by a set of relationships. These rela-
tionships may be simple “is a” and “has a” relationships (e.g., Purkinje cell is a neuron, neuron has a
nucleus), or they may be more complex.^52 From the above statements, a search algorithm could infer
that “Purkinje cell has a nucleus” if the ontology is encoded in a form that would allow such reasoning
to be performed. Because the knowledge required to link concepts is contained outside of the source
database, the CCDB is relieved of the burden of storing exhaustive taxonomies for individual datasets,
which may become obsolete as new knowledge is discovered.
The UMLS has recently incorporated the NeuroNames ontology^53 as a source vocabulary.
NeuroNames is a comprehensive resource for gross brain anatomy in the primate. However, for the
type of cellular and subcellular data contained in the CCDB, the UMLS does not contain sufficient
detail. Ontologies for areas such as neurocytology and neurological disease are being built on top of the
UMLS, utilizing existing concepts wherever possible and constructing new semantic networks and
concepts as needed.^54
In addition, imaging data in the CCDB is mapped to a higher level of brain organization by register-
ing their location in the coordinate system of a standard brain atlas. Placing data into an atlas-based
coordinate systems provides one method by which data taken across scales and distributed across
multiple resources can reliably be compared.^55
Through the use of computer-based atlases and associated tools for warping and registration, it is
possible to express the location of anatomical features or signals in terms of a standardized coordinate
system. While there may be disagreement among neuroscientists about the identity of a brain area
giving rise to a signal, its location in terms of spatial coordinates is at least quantifiable. The expression
of brain data in terms of atlas coordinates also allows them to be transformed spatially to offer alterna-
tive views that may provide additional information (such as flat maps or additional parcellation
(^51) B.L. Humphreys, D.A. Lindberg, H.M. Schoolman, and G.O. Barnett, “The Unified Medical Language System: An Informatics
Research Collaboration,” Journal of the American Medical Informatics Association 5(1):1-11, 1998.
(^52) A. Gupta, B. Ludascher, J.S. Grethe, and M.E. Martone, “Towards a Formalization of a Disease Specific Ontology for
Neuroinformatics,” Neural Networks 16(9):1277-1292, 2003.
(^53) D.M. Bowden and M.F. Dubach, “NeuroNames 2002,” Neuroinformatics 1:43-59, 2002.
(^54) A. Gupta, B. Ludascher, J.S. Grethe, and M.E. Martone, “Towards a Formalization of a Disease Specific Ontology for
Neuroinformatics,” Neural Networks 6(9):1277-1292, 2003.
(^55) A. Brevik, T.B. Leergaard M. Svanevik, J.G. Bjaalie, “Three-dimensional Computerised Atlas of the Rat Brain Stem
Precerebellar System: Approaches for Mapping, Visualization, and Comparison of Spatial Distribution Data,” Anatomy and
Embryology 204(4):319-332, 2001; J.G. Bjaalie, “Opinion: Localization in the Brain: New Solutions Emerging,” Nature Reviews:
Neuroscience 3(4):322-325, 2003; D.C. Van Essen, H.A. Drury, J. Dickson, J. Harwell, D. Hanlon, and C.H. Anderson, “An Inte-
grated Software Suite for Surface-based Analyses of Cerebral Cortex,” Journal of the American Medical Informatics Association
8(5):443-459, 2001; D.C. Van Essen, “Windows on the Brain: The Emerging Role of Atlases and Databases in Neuroscience,”
Current Opinion in Neurobiology 12(5):574-579, 2002.