PET can provide early diagnosis of diseases like Alzheimer's Disease (AD). FDG PET, that shows metabolic activity and FDDNP PET, that binds to neurofibrillary tangles (NFT) and beta amyloid, are two of these important PET tracers to examine AD. FDG PET and FDDNP PET both show characteristic image patterns related to a subject's cognitive decline. It is easier to determine the tracer activity in a particular region with the use of MRI. Hemispheric cortical surface maps determined from MR images have been successfully used to look at structural measures such as cortical thickness. The cortical surface maps have an extra warping step that takes sulci identified on the cortical surface and warps them to a common set which puts each cortical surface map in close alignment. In this work, I developed and evaluated a method to integrate PET image information onto the cortical surface maps. I looked at the effect of using image warping techniques that are based on level sets to warp sulci to a set of sulci in the common space on the cortical surface. Partial volume effect, kernel size as well as head movement are some issues that need to be addressed with studies that integrate dynamic PET information into MR images The resulting cortical surface maps of PET data allow us to determine the pattern changes of PET intensity as a subject's cognitive ability declines. A statistical model relating FDDNP PET signal to the MMSE score (a clinical cognitive measure) was established that allowed a subject's MMSE score to be estimated from FDDNP PET images with good accuracy.