Alzheimer's disease (AD), which is the most common form of dementia, is increasingly prevalent and represents a worsening healthcare burden. Typically, the symptoms of the disease begin with mild memory difficulties and evolve towards increasing cognitive impairment and dysfunctions in complex daily activities [1]. At present, there are few disease-modifying therapies available, although some treatments are available to improve AD symptoms [2].

AD is characterized as a slowly progressive neurodegenerative disease. Post-mortem histopathology has shown the presence of neuritic plaques and neurofibrillary tangles in patients with AD, as a result of amyloid-beta (Aβ) accumulation and tau aggregation. With the advent of radiotracers with a molecular affinity to amyloid and tau proteins, the spatiotemporal evolution of AD pathology has been characterized in depth using positron emission tomography (PET). Neurodegenerative changes are thought to occur downstream of abnormal protein accumulation, and these changes can be measured during life using either PET tracers or magnetic resonance imaging (MRI). Previous studies have shown limited correlation between amyloid pathology and cognitive changes, while both tau pathology and neurodegeneration are thought to be more closely linked to cognition and disease severity [3,4] . To better understand the associated pathological changes with AD, preclinical animal models have been developed to recapitulate specific AD-associated features. One of the most widely used models of AD is the triple-transgenic model (3xTg-AD) [5], which exhibits both amyloid and tau pathology, along with cognitive changes.

Standard structural MRI [6] is a non-invasive technique that can yield information on gross anatomical changes, including hippocampal atrophy [[6], [7], [8]]. VBM is an automated technique that spatially normalizes, segments, and smooths high resolution structural MR images before performing voxel-wise statistical tests to examine differences in brain anatomy between groups. Using VBM, differences in the relative grey matter (GM) density (or concentration) can be compared across groups of subjects on a voxel-wise basis, without the need for a priori regions of interest. While such differences do not directly relate to cellular density [9], the biophysical basis of these differences has been previously attributed to the combined effects of tissue volume changes, vascular changes, and cellular changes such as cell count, distribution, and morphology [[10], [11], [12]]. This technique has been successfully used to investigate morphological changes in a wide range of neurological disorders, including AD [13,14] and Parkinson's disease [15].

More advanced MRI techniques include diffusion MRI (dMRI), which probes the microscopic movement of water throughout the brain [16]. The microscopic sensitivity of dMRI can highlight microstructural changes that may be occurring earlier in the disease. For example, previous studies have shown that changes in water diffusivity [17,18], structural connectivity [19], and white matter (WM) microstructural integrity [20,21] are associated with AD. The predominant dMRI technique to assess WM microstructure is diffusion tensor imaging (DTI), which yields metrics such as fractional anisotropy (FA). Unfortunately, single-shell DTI (the prototypical DTI technique for clinical MRI scanners, defined by a single diffusion b-value) has several limitations that may reduce its accuracy and specificity. For instance, traditional DTI cannot represent multiple and independent intra-voxel orientations within a brain voxel [22]; additionally, DTI is sensitive to partial volume effects (PVEs), wherein the resulting DTI metrics reflect a weighted average of multiple diffusion components within a voxel and are thus no longer specific to a single tissue type [23]. To overcome this latter limitation, free-water (FW) correction algorithms for DTI (FW-DTI) have been developed [24] and were recently shown to improve the accuracy and sensitivity of WM microstructural analysis in AD [25].

Because GM atrophy and abnormalities in the architecture and microstructure of WM have been well demonstrated in human AD [26,27], the purpose of this study was to assess GM density and microstructural differences in WM using VBM and FW-DTI between 3xTg-AD and wild-type (WT) mice. These methods used in combination may yield complementary insight into GM and WM changes in AD mouse models. Additionally, post-mortem histopathology was performed to verify amyloid and tau accumulation in regions across the brain. Ultimately, studies linking GM and WM changes in preclinical models to amyloid and tau deposition could substantiate the link between specific AD pathology and structural/microstructural neurodegeneration. Given the link between cognitive decline in humans and neurodegeneration, these studies may provide mechanistic insight into whether GM and WM changes are subsequent drivers of cognitive decline.