ASSESSING THE RELATIONSHIPS BETWEEN FUNCTIONAL CONNECTIVITY, STRUCTURAL MORPHOMETRY, AND COGNITIVE MEASURES IN ALZHEIMER’S DISEASE

Abstract

As Alzheimer’s disease (AD) progresses, various changes occur such as decline in cognitive and brain functions, as well as increased brain atrophy. Measures of functional connectivity, structural morphometry, and the CERAD neuropsychological battery (CERAD) have been used to evaluate the progression of AD. These measures are sensitive because they distinguish healthy control (HC) from both mild cognitive impairment (MCI) and AD participants. In this study we investigate how some of these measures relate to stages of AD. Structural MRI, resting state functional MRI, and CERAD were collected in 16 HC, 16 MCI, and 16 Early AD (EAD) participants. Relationships between functional networks (default mode, salience, right executive control, and left executive control networks), structural morphometry (cortical thickness and volume, and subcortical volume), and CERAD were analyzed using Multiple Factor Analysis (MFA). MFA is a principal component analysis based technique used to analyze multiple tables that measure sets of variables collected on the same participants. We found a negative relationship between brain function and brain structure. Cortical thickness, left cortical volume, and subcortical volume differentiated HC and MCI from EAD. CERAD differentiated MCI from EAD. Salience and right executive control networks, left cortical volume, left cortical thickness, and CERAD had the most reliable between group effect (p< .05). The brain regions that mainly contributed to the between group effect belonged to temporal brain regions and hippocampus. For CERAD, the scores which contributed to the between group effect were verbal fluency, Boston naming task, word list recall, word list learning, and constructional praxis recall. We found two important results. First, structural measures were best at separating HC and MCI from EAD participants, whereas CERAD was best at separating MCI from EAD. Second, the brain regions and cognitive measures which drove the separation between cohorts were the regions known to be affected early in the course of disease. Together, these two results show that atrophy is most sensitive to detect the early stages of AD whereas functional connectivity and cognition are less sensitive but can still significantly separate HC from EAD.