Towards multimodal atlases of the human brain

Publisher Website
Google Scholar
Abstract
Brain atlases are reference systems that associate neuroanatomical labels (nomenclature) with canonical representations of anatomy in a three-dimensional coordinate system. These systems commonly integrate multisubject data from many different sources (for example, histology, functional MRI and positron emission tomography), and could provide statistical representations of anatomy and function in whole populations. Initially, brain atlases were purely neuroanatomical, based on a single, often sparsely sampled, representative example. Now, brain atlases include population statistics on structure, gene expression, receptor patterns or connectivity over time. Modern cytoarchitectonic studies are using computational methods to pool information across subjects on such features as receptor distributions, myelination characteristics and cellular content. Correlations between functional activation and the underlying cyto- or myeloarchitecture can then be tested by using these architectural maps to define regions of interest in functional imaging studies. Integration of cytoarchitectural maps from many subjects has allowed classical maps to be re-evaluated and corrected. It has also led to a quantitative description of the intersubject variability of cytoarchitectonic areas and to the discovery of hitherto unknown cytoarchitectonic areas in the intraparietal, secondary somatosensory and extrastriate cortex. Diffusion tensor imaging (DTI) provides a new source of image contrast to map white matter integrity and connectivity, and has opened up new opportunities for brain mapping and atlasing. Population-based atlases can average anatomical features across individuals, revealing generic features that are not identifiable in individual representations owing to their considerable variability. They have identified group-specific patterns of brain structure in Alzheimer's disease, HIV/AIDS, schizophrenia, in methamphetamine users, and in developmental disorders such as fetal alcohol syndrome and Williams syndrome. Brain atlases are beginning to be used in clinical studies, including drug trials of antipsychotics or mood stabilizers, to investigate factors that influence disease expression and therapeutic response. In the next decade, population-based atlases will probably gain widespread applicability in genetic studies. Data from twins and those at genetic risk for specific diseases have been incorporated into brain atlasing studies to discover previously unknown effects on the brain of variations at specific genetic loci.