Magnetic resonance imaging and spectroscopy in at-risk populations and preclinical Alzheimer's disease

Abstract

Alzheimer’s disease (AD) is the most common form of dementia. One of the earliest and most distinct features of AD is memory loss, followed by difficulties in learning and a decline in cognitive abilities. People afflicted with AD usually develop symptoms in their late seventies, but we know today that the very first signs of pathology can be detected decades before symptom onset. A considerable part of AD research today is focused on the detailed characterization of this asymptomatic “silent” phase of AD.

The main pathological hallmarks of AD are amyloid plaques – abnormal extracellular deposits of the amyloid-β (Aβ) protein and intracellular neurofibrillary tangles (NFTs) – aggregates of the phosphorylated tau protein. AD is also characterized by progressive neurodegeneration – a deterioration of the structure and function of neurons, which leads to loss of brain tissue. Atrophy first takes place the medial temporal lobe (the entorhinal cortex and the hippocampus) and subsequently propagates to other areas of the brain. Magnetic resonance imaging (MRI) is a powerful method used to assess the extent of atrophy of the whole brain or specific structures. Another useful tool for studying AD pathology is magnetic resonance spectroscopy (MRS). This method allows quantification of certain brain metabolites in vivo. The most relevant MRS metabolites in the context of AD are myo-inositol – an organic osmolyte and N-acetyl-aspartate – a marker of neuronal integrity. The overall aim of this thesis is to further characterize structural and metabolic changes associated with incipient AD pathology.

Study I assesses a common methodological issue of volumetric MRI studies related to inter-individual differences in intracranial volume (ICV). In a study setting where regional brain volumes are analyzed, it is often of interest to compare groups, e.g. control vs patient, in order to quantify atrophy due to pathology. This type of group comparison is confounded by the fact that people with larger ICV usually have larger brain structures, making it difficult to isolate disease-related atrophy. This work examines multiple procedures that can be used to compensate for ICV in volumetric studies, highlighting that the choice of ICV normalization approach may have profound effects on the interpretation of study results. Study II examines brain morphology from a network perspective. Here, the brain is represented as a graph – a set of nodes interconnected by edges – where the nodes are based on anatomical regions and the edges are measures of the “connection” (i.e. structural co-variance) between these regions. We examine global and local network properties in cognitively healthy elderly with evidence of amyloid pathology. Study II reveals that the changes in cerebral network topology in asymptomatic individuals at risk for AD occur before to any detectable cortical thinning.

Studies III and IV explored whether brain metabolites measured with MRS may be useful biomarkers of ongoing amyloid-related pathological processes. Previous MRS studies have found that in a typical AD spectrum, mI is elevated and NAA is decreased. However, relatively little is known about the time course of these changes as well as the interplay between MRS and established biomarkers/risk-factors for the disease. In Study III, MRS spectra of non-demented individuals at varying degrees of risk for AD was examined in conjunction with information about Aβ, tau and APOE ε4 carriership – the main genetic risk factor for AD. Our findings highlight the very early involvement of brain mI in AD. We show that this metabolite is changed already at presymptomatic disease stages, and that elevated mI is linked to a higher Aβ plaque load. Study IV is an extensive follow-up of Study III, and the first longitudinal MRS study, taking into account individual amyloid status. We demonstrate that during a four-year follow-up, non-demented individuals with pathological baseline Aβ accumulate mI at a higher rate, suggesting that mI may have the ability not only to detect but also to track ongoing Aβ pathology. Ultimately, we hope that non-invasive cost-efficient MRS markers may be useful for early patient screening and evaluation of disease-modifying strategies.