Dynamic human brain mapping and analysis: From statistical atlases to patient-specific diagnosis and analysis

Christos Davatzikos

Research output: Chapter in Book/Report/Conference proceedingChapter


This chapter describes methodologies for measuring spatiotemporal patterns of brain structure in brain images, a problem that arises very often in monitoring disease progression and treatment responses, from a series of scans. A4-dimensional shape transformation is used to map images to a stereotactic coordinate space, in order to standardize the coordinates of anatomical structures across different individuals and remove interindividual variability. A statistical atlas is then constructed from a set of data that has been mapped to the same stereotactic space, and reflects the variation of brain structure across individuals of the population used to construct the statistical atlas; the transformation that was used to map the images to the stereotactic space is also measured, as it often constitutes a key morphometric measurement reflecting morphological characteristics of the respective individual relative to a standardized template brain that resides in the stereotactic space. Individual patient scans are then compared against one or more statistical atlases, in order to diagnose disease or predict likelihood of disease progression. This statistical comparison is typically performed via pattern classification systems, which are trained to recognize spatiotemporal patterns of brain structure that are highly characteristic of a disease of interest.

Original languageEnglish (US)
Title of host publicationPrinciples and Advanced Methods in Medical Imaging and Image Analysis
PublisherWorld Scientific Publishing Co.
Number of pages26
ISBN (Electronic)9789812814807
ISBN (Print)9789812705341
StatePublished - Jan 1 2008
Externally publishedYes

ASJC Scopus subject areas

  • Medicine(all)


Dive into the research topics of 'Dynamic human brain mapping and analysis: From statistical atlases to patient-specific diagnosis and analysis'. Together they form a unique fingerprint.

Cite this