Spatial Correlates Integrating Image and Behavioral Biomarkers of Cognitive Dysfunction in Mouse Models of AD

Abstract

Alzheimer’s disease (AD) is a disorder that is the 6th leading cause of death in the United States. Although physicians and researchers are aware of the pathologies that come with the disease such as the presence of amyloid plaques, neurofibrillary tangles, and neuronal degeneration - the actual onset and progression of the disease is unknown. Currently, there is no known cure for Alzheimer’s. Therefore, it is critical to identify early biomarkers, to help better understand the disease in its early stages and monitor possible treatments given to patients. In this study, mouse models of AD and age matched controls are utilized in a multivariate approach using several magnetic resonance imaging protocols to identify potential image based biomarkers attributed to AD. We used predictive-based methods of imaging and behavior metrics to identify relevant anatomical networks by means of sparse regression. Another method to determine generalizable regions of interest based solely on imaging data was obtained from eigenanatomy. This knowledge of associative regions by either method was used for predictive modeling of behavioral characteristics in AD and control mice. Several regions, cerebellum, hippocampus, and motor cortex, showed significant changes in each of the three imaging contrasts and were also identified as linked to memory impairment. Successful predictions were produced for behavior assessment using the Morris Water Maze. All things considered, predictive analysis via imaging and behavior could be used to determine regions associated with AD that also match up with imaging markers of the disease.