Identifying Age Related Atrophy Using Levelset Registration of Embedded Maps

Structural brain changes that occur during the process of normal aging are of great interest to neuroscientists investigating age-related changes in cognitive function. A variety of techniques currently exist to identify group differences in brain anatomy with the greatest focus having been placed on cortical anatomy. Although cortical anatomy is of considerable significance, the underlying white matter tracts are at great risk in the aged brain. Patchy white matter lesions referred to as leukoaraiosis (LA) have long been identified radiologically as T2-weighted hyperintensities in older adults. Techniques such as voxel-based morphometry (VBM), deformation-based morphometry (DBM), or tissue thickness measurements allow for quantitative comparisons of white matter in a regionally specific manner. However, VBM is dependent upon segmentation algorithms to measure tissue density, a task difficult in images displaying the diffuse contrast changes associated with LA. Similarly, DBM is dependent on registration algorithms to establish a spatial correspondence between images.

Unfortunately, current image registration methods do not perform effectively in regions where tissue contrast has changed, such as in leukoaraiosis. The goal of this Driving Biologic Project (DBP) is to evaluate the relationship between changes in cognitive function and cerebral white matter in older adults. Since current VBM and DBM methods cannot be applied to subjects with leukoaraiosis, the project will expand current image registration techniques developed by the Center for Computational Biology (CCB) at UCLA to accommodate topological changes in the images.

Specific Aims

  • Aim 1: Develop a deformable registration method for images represented as embedded maps, employing the level-set framework. We will develop a deformable registration approach representing each image as an embedded map. This registration framework will simultaneously account for changes in intensity and morphology between image pairs.
  • Aim 2: Identify the relationship between executive cognitive functions and age-related changes in cerebral white matter using the level-set based registration of embedded maps approach. To evaluate white matter integrity, quantitative T1 and T2 maps will be registered using the embedded maps algorithm developed in Aim 1. We will identify white matter intensity and volume changes (i.e. leukoaraiosis) in a regionally specific manner and correlate these changes with cognitive function.