MRI uses radio waves in the presence of a powerful magnetic field to align the nuclear magnetization of (usually) hydrogen atoms in water in the body. Atomic nuclei absorb and reemit electromagnetic waves at a resonant frequency, which falls in the radiofrequency range. Originally, MRI was used for producing anatomical images (6) but developed to a technique with a wealth of information contained in the signal regarding the physicochemical state of tissues, flow, diffusion, motion, and more recently, molecular targets. MR has much greater soft tissue contrast than CT, making it especially useful in neurological imaging.

For most medical brain examinations in general, both T₁– and T₂-weighted images are acquired for observation of white-matter and gray-matter brain structures and dementia-related morphological changes, such as changes in brain volume and size of the ventricle and specific substructures. Other specific MRI parameters correlate with the diffusion of water molecules (i.e., Brownian motion) in biological tissues (7) and are quite valuable in studying neuropathology-related changes in intra/extracellular water balance and subcellular structural changes (8). The recent development of diffusion tensor imaging (DTI) enables diffusion measurement simultaneously in different directions from which tissue anisotropy can be calculated per voxel (9). From these images, brain maps of fiber directions can be reconstructed to study brain connectivity, a feature affected in several neuropathologies (10). fMRI is capable of mapping functional regions of the cortex in real time during specific task activation. This is achieved using magnetic susceptibility image contrast, which detects the local increase in CBF due to activation. Deoxyhemoglobin is paramagnetic and interacts with the magnetic field, whereas oxyhemoglobin is not. As a consequence, a higher influx of oxygenated blood during activation increases local T₂ and T₂* values. Such signal intensity change is termed blood oxygenation level-dependent (BOLD) contrast (11). MRI is currently still a new modality in the fields of in vivo molecular and cellular imaging compared with the established methods such as PET and optical imaging. This delayed development is probably inherent to the low sensitivity of MRI. However, MRI is gaining headway, since its main contribution to molecular imaging is its high spatial resolution compared to other methods. MRI can achieve a microscopic resolution of 100 µm in all dimensions on a routine basis in living animals, and 10 µm is feasible in fixed specimens (12). This almost represents conventional histology, while allowing the object to be viewed interactively in any plane. Merely prolonging scan time to increase spatial resolution is not a viable option in fragile, anesthetized transgenic animals. However, with the development of higher field strength magnets and the use of small cryogenically operated coils (13), spatial resolution is still improving because of the increase in signal-to-noise ratio. Some in vivo MRI studies of mouse brain reported so far cover a wide range of voxel sizes and field strengths, e.g., 98 × 156 × 1,500 μm³ at 11.7 T (14), 140 × 140 × 1,200 μm³ at 4.7 T (15), 156 × 156 × 1,000 μm³ at 14 T (16), and 78 × 156 × 343 μm³ at 7 T (17).

Advances in MRI have expanded its use in neuroimaging by applying exogenous contrast agents. The range of applications of these compounds is determined largely by their biodistribution and pharmacokinetics. Contrast agents are typically injected intravenously and, depending on their chemical structure, may remain in the vasculature (iron oxide particles), enter the interstitial space (Gadolinium chelates), or be taken up by cells (manganese). Macromolecular contrast agents are intravascular agents and are well suited to quantitative imaging of perfusion and vascular volume in brain activity studies (18). Freely diffusible contrast agents such as Gadolinium chelates are used in tissue permeability studies such as blood–brain barrier (BBB) leakage, as none of the contrast agents can pass the BBB. For molecular imaging applications, “targeted” and “smart” contrast agents were developed (19,20). Targeted agents incorporate ligands, such as antibodies that bind to specific molecular markers in the tissue, whereas smart agents are activated by the presence of specific ions or enzymes. These agents allow in vivo visualization of gene expression, an emerging field within molecular imaging (21). Moreover, MRI is also one of the completely safe imaging tools since this technique does not make use of ionizing or nuclear compound injections.

MRS is a magnetic resonance technique that can measure brain biochemistry. This technique together with other MRI applications provides metabolic information such as anatomical, physiological, functional, or molecular imaging. There is increasing evidence that impairment of energy metabolism, mitochondrial dysfunction, and altered glutamatergic neurotransmission are critical factors in the development and progression of neurodegenerative diseases. Therefore, in vivo MRS provides a useful tool in the evaluation of the underlying processes, as many of the observable metabolites prove to be relevant in the assessment of neurochemical changes associated with neurodegeneration and dementia.

In localized ¹H-MRS, the target is the proton with a natural abundance of almost 100%, ultimately allowing the quantification of more than 18 ¹H containing metabolites in vivo (22). Localized MRS techniques allow assessment of metabolites resorting from a well-defined volume, e.g., an affected brain region, and can return single- or multivolume information with microliter resolution. Multivolume methods are known as chemical shift imaging (CSI) or spectroscopical imaging (SI), which give a voxelwise metabolite distribution, i.e., a metabolite image (23,24) (Fig. 5).

Metabolites relevant for neurodegeneration can be used as fingerprints for the diseased brain. The ¹H-MRS spectrum of the normal brain is dominated by the large N-acetyl aspartate (NAA) resonance at 2.01 ppm. The exact role of NAA remains unknown, but it has been used extensively as a marker for neuronal density, thereby giving a measure for neuron survival and death (25). Myo-inositol (mI) is a cyclic sugar alcohol present in high concentrations. The precise function of mI is not known, but it has been dedicated as being an osmolyte, a glial marker, or a breakdown product of myelin. Typical changes of mI over time are seen in AD, cognitive impairment, and brain injury (26,27). Taurine (Tau) is another putative osmolyte of the brain. It has also been implicated as a modulator of neurotransmitter action. Tau is present in all cell types of the CNS, being a prominent metabolite in rodents rather than in humans (28). Glutamate (Glu), the major excitatory neurotransmitter and main precursor of γ-aminobutyric acid (GABA) and gluthation (GSH) (29), can be detected with MRS together with glutamine (Gln). Lactate (Lac) is normally present at low concentrations in the brain and is elevated in cases such as stroke, ischemia, or tumors (30–32). One of the most apparent resonances in a ¹H-MRS spectrum of the brain is the combined resonance of both creatine (Cr) and phosphocreatine (PCr) at 3.03 ppm, also referred to as total creatine (tCr). The tCr signal is relatively constant with only negligible changes reported with age or disease. This is the reason why it is often used as an internal reference for quantification (33). Choline (Cho)-containing compounds form a prominent resonance in ¹H-MRS spectra of the brain at 3.2 ppm. Since there are contributions of glycerophosphocholine (GPC), Cho, and phosphocholine (PCho), it is often referred to as total choline (tCho). Together with mI, tCho is often considered as a glial marker, and thus a marker for gliosis, since concentrations of these compounds are typically much higher in glia than in neurons (34,35).

The techniques of CT and MRI are commonly used as instrumental aids for structural imaging in the diagnosis of dementia (21,36,37). However, new imaging tools are emerging and might significantly change the way CT and MRI are used nowadays in clinical settings. CT is a medical imaging method employing tomography, which means that digital geometry processing is used to generate a 3D image of the inside of an object from a large series of 2D X-ray images taken around a single axis of rotation. CT visualizes differences in the physical densities of biological tissues through X-ray absorption. The main issue within radiology today is how to reduce the radiation dose during CT examinations without compromising image quality.

For animal research, µCT systems are available. They are fivefold cheaper than µMRI systems and are straightforward to site in a laboratory. 3D µCT has proven to be a powerful technique for imaging and analysis of bone structure and density in small animals (38). Although initial µCT imaging was performed ex vivo, recent advances allow for similar imaging protocols to be performed in vivo (39). Unlike ex vivo µCT, which provides very fine spatial resolution (typically less than 10 μm), the spatial resolution of in vivo µCT studies suffers from physiologic (e.g., cardiac and respiratory) motion. High-resolution in vivo µCT has become important for examining morphology in small animal models for tumor disease (40,41).

Experimental CT imaging in animal models for dementia is very limited since soft tissue imaging using CT requires the use of X-ray absorbing contrast agents due to the lack of native CT contrast in the brain. Especially in the brain, distinction between gray and white matter (in quantitative terms, gray matter 20–35 HU (Hounsfield units), white matter 30–40 HU) needs addition of a radio-opaque component (iodine anion). There is, however, a CT study in a mouse model for dementia wherein postmortem material amyloid plaques could be discerned (42).

Upon application of intravenous injection of a CT contrast agent, investigation of cerebral vascular system impairments is made possible (43). The occurrence of cerebral vascular abnormalities is not only linked to vascular dementia but is also an issue often discussed concerning all types of dementia.

Together with MRI, PET and SPECT are the most commonly used imaging techniques in animal models. PET uses biologically active molecules in µmolar or ηmolar concentrations that have been labeled with short-lived positron-emitting isotopes. Opposed to PET, the imaging agent used in SPECT emits gamma rays.

Both PET and SPECT can be used to measure regional CBF, which is physiologically coupled to brain metabolism making its use relevant in dementia studies. The radiotracer is assumed to accumulate in areas of the brain, in proportion to the rate of delivery of nutrients to that volume of brain tissue.

PET also provides measures of regional cerebral glucose metabolism (rCMRglc) by using a positron-emitting glucose analog, fluorodeoxyglucose (FDG). FDG-PET has become a standard technique in oncology and is gaining headway in dementia research in animal models.

Limitations to the widespread use of PET arise from the high costs of cyclotrons needed to produce the short-lived radionuclides for PET scanning and the need for specially adapted on-site chemical synthesis apparatus to produce the radiopharmaceuticals. SPECT scans using ^99mTc-HMPAO (hexamethylpropylene amine oxime) competes with FDG-PET scanning of the brain to provide very similar information about local brain damage from many processes. SPECT is more widely available because the radioisotope generation technology is longer lasting and far less expensive. The same is true for the gamma scanning equipment.

A combined or multimodality scanner, such as SPECT/CT or PET/CT, can acquire coregistered structure and function in a single study. The data are complementary, allowing CT to accurately localize functional abnormalities and SPECT or PET to highlight areas of abnormal metabolism. Although technically more challenging, the simultaneous acquisition of MR and PET has also been demonstrated, initially for preclinical imaging of small animals, but more recently for the human brain (44).

PET is an extremely sensitive technique allowing for the identification of subtle changes in brain neurochemistry by labeling the molecules with high specificity. Selective tracers for many neurotransmitter receptors and other cellular proteins or tracers that act as an enzyme substrate, which allow the detection of a specific neuronal function, are available. Novel disease-specific probes are being developed very rapidly. New perspectives are opened by tracers for imaging amyloid plaques, tracers for measuring local acetylcholinesterase (AChE) activity and the binding capacity of nicotinic and serotonergic receptors, which address neurotransmitter deficits in dementia.

Nuclear medicine is a part of molecular imaging because it produces images that reflect biological processes taking place at the cellular and subcellular level. Nuclear imaging techniques, however, suffer from poor spatial resolution, with the highest resolution scanners today giving at best 4-mm resolution and for animal systems up to 1.5 mm. Therefore, PET and SPECT scans are increasingly read alongside CT or MRI scans, taking advantage of the coregistration of both anatomic and metabolic information to estimate brain regions of altered radioligand uptake. For the same reason, and very recently, small animal SPECT/MRI dual hybrid device combining pinhole SPECT imager adjacent to a small low-field MR imager has been developed (45).

BLI and FLI enable visualization of genetic expression and physiological processes at the molecular level in living tissues. Reporter genes that encode either fluorescent or bioluminescent proteins have been used to create internal biological light in the study of animal models for human biology and are commonly used in tumor research (46). Longitudinal recordings of living animal models with these modalities have shown to provide clues for the development of clinical applications. Because of the presence of the skull, optical methods for brain applications are more restricted, but nevertheless abundant.

Near-infrared fluorescent (NIRF) probes are typically small fluorescent molecule dyes designed to absorb and emit light in the near-infrared (650–950 nm), where tissue scattering is lowest. The simple synthesis, the low cost, and the long life of NIRF probes present an attractive alternative for MRI and PET techniques in the field of molecular imaging.

BLI does not rely on an external light as for FLI, instead uses luminescence sourcing (e.g., luciferase) from, e.g., the firefly. Sensitive charged-coupled device (CCD) cameras are used to capture the bioluminescent data, which are superimposed on the photographic image for interpretation. Because there is no competing background signal, those techniques can detect low levels of light emission. The DNA encoding the luminescent protein is incorporated into the laboratory animal either via a viral vector or by creating a transgenic animal. Within these experimental animals, luciferine is injected and its oxidation by luciferase provides bioluminescence. Because luciferase imaging requires genetic transfection, it is very unlikely to ever find an application for this technique in human studies. Nevertheless, it can be very useful in predicting human response to therapy in the early stages of drug development.

In dementia research, progression is made in designing smart optical probes that emit a characteristic fluorescence signal only when bound to the studied target, i.e., a biomarker for the disease. The first NIRF probes for use in dementia were designed in 2001 and targeted amyloid in the brain of mouse models for AD (47). Until now, targeting amyloid is the most common application in dementia research.

The potential of BLI in dementia research finds its application in stem-cell-based therapies using bioluminescent stem cells and BLI as a superior viability marker. Stem-cell-based therapies may offer a potentially attractive and unlimited source of cells for the repair of damaged tissue in CNS disease. Experimental CNS transplantation studies have largely evaluated the in vivo survival, proliferation, and migrational capacity of neural grafts within the host tissue with histology at several different time points. Suitable noninvasive imaging technologies may serve as tools for probing neural graft behavior in the living brain. While other imaging approaches, such as MRI or PET, might have been chosen for this purpose, the limitations associated with these modalities outweigh their benefits. Particularly in MRI, which can monitor stem-cell populations that are iron-labeled, the in vivo distinction between viable and nonviable cells is not possible. Moreover, iron uptake can be confounded by other resident cells such as macrophages.

The etiologies of neurodegenerative diseases may be diverse; however, a common pathological denominator is the formation of aberrant or misfolded proteins. There is an ongoing debate about the nature of the harmful proteins and how toxic conformations selectively can damage neuronal populations. Soluble oligomers are associated with early pathological alterations, and strikingly, oligomeric assemblies of different disease-associated proteins may share common structural features. A major step toward the understanding of mechanisms implicated in neuronal degeneration is the identification of related genes. Studies based on these disease-associated genes illuminated the two faces of protein misfolding in neurodegeneration: a gain of toxic function and a loss of physiological function, which can even occur in combination (60).

Here, we review the occurrence of specific protein accumulations in several dementia disorders and summarize how these early-stage accumulations can be detected through neuroimaging methods in animal models.

MEMRI can also be used to study brain connectivity and potential changes upon neurodegeneration. In that case, specific brain nuclei of the affected circuit are injected (targeted) with manganese, and subsequently, the signal changes in the projection areas reveal the integrity and the activity of the connections. Pelled et al. (94) performed such a study in healthy and in unilateral 6-hydroxydopamine (6-OHDA) Parkinson disease rat model to test the basal ganglia interhemispheric connectivity after injection of MnCl₂ into the entopeduncular nucleus, substantia nigra (SN), and the habenula. They discovered modulations in the effective intra- and interhemispheric basal ganglia connectivity in the unilateral 6-OHDA rat model, which suggests a linkage between the dopaminergic and serotonergic systems in PD, in line with clinical symptoms. The use of manganese has been, however, limited due to concerns about toxicity, specifically acute cardiovascular depression, as it acts as a transient competitive antagonist to calcium activity, reducing contractility and systemic blood pressure. To date, it has been approved for clinical imaging only in a chelated form, i.e., manganese dipyridoxyl diphosphate (Mn-DPDP). While chelation markedly improves the acute safety profile of manganese, it does so at the expense of many of the properties that make it desirable for neuroimaging.

In humans, changes in functional connectivity in specific regions of the brain affected by dementia are investigated with DTI, resting-state, and activation fMRI studies. A multitude of studies exist in patients with AD but also in patients with HD (95,96). However, no such studies were performed on animal models, but some recent papers have exploited the technique in anesthetized control rats (97).

On the other hand, DTI studies to detect white matter changes associated with dementia were done in both patients (98) and in animal models of dementia. DTI studies in animals offer the enormous benefit of providing insight into the underlying mechanism of the observed changes in the DTI parameters being the radial and the axial component. This allows discriminating connectivity loss due to demyelinization or axonal loss (99,100). DTI studies in AD (86) and PD models (101) have been helpful in discerning the specific degenerating fibers in each disease.

The mammalian subventricular zone (SVZ) is the largest germinative zone of the adult brain, which contains a well-characterized stem cell niche with neuronal progenitor cells (NPCs). While most studies highlight the neurogenic potential of SVZ progenitors, recent data indicate that SVZ cells become reactivated in response to different pathological cues, such as trauma, ischemia, inflammation, demyelination, and also neurodegeneration like in PD (102) and HD (103). Similar multiple studies exist and what they have in common is that they assess cell proliferation and differentiation with the gold standard being 5-bromo-2’-deoxyuridine (BrdU) labeling, and subsequent, histology and immunohistochemistry for cell type-specific markers.

However, recently, several in vivo visualization methods were developed, which would ultimately permit to follow neuronal recruitment at the rostral migratory stream (RMS) and olfactory bulb (OB) as a read-out for neurogenesis and/or cell survival in animal models for different brain pathologies. NPCs were labeled, in vivo, in situ in adult rats with fluorescent, micron-sized iron oxide particles (MPIOs), and using MRI, the migrating neural precursors carrying MPIOs were discerned along the RMS to the OB. Immunohistochemistry and electron microscopy indicated that particles were located inside NPCs positive for glial fibrillary acid protein (GFAP+) in the SVZ, inside migrating NPCs positive for polysialylated neural cell adhesion molecule (PSA-NCAM+), in doublecortin positive NPCs within the RMS and OB, and in mature neurons in the OB staining positive for neural nucleus (Neu-N+) (104). This work demonstrates that in vivo cell labeling of progenitor cells for MRI is possible and enables serial, noninvasive visualization of endogenous progenitor/precursor cell migration. Undoubtedly, this method will be used in the near future to phenotype neurodegenerative models and to study the efficacy of endogenous progenitor cells as an attractive alternative to the existing therapeutic options for treating neurodegenerative diseases.

The group of Aigner L. et al. (105) developed a new tool based on the doublecortin promoter, driving the expression of the luciferase reporter gene (DCX-promo-luciferase) in transgenic mice to perform in vivo imaging of neurogenesis. Indeed, the DCX-promo-luciferase mice allowed optical in vivo imaging of the onset of and increase in neurogenesis in developing fetal brains, as well as imaging of neurogenesis in the intact adult mouse central nervous system. This imaging approach allows longitudinal study of neurogenesis in intact animals without the requirement of cellular prelabeling. Moreover, it guarantees that detection is specific for neuronal precursors and restricted to viable cells (106). The group of Baekelandt finally extended this to a truly quantitative method (107,108), as illustrated in mice in Fig. 3.

The dentate gyrus is the second germinative zone in the mammalian brain that continuously provides the adult hippocampus with new neurons. Because the germinative layer is in close vicinity to the target region, the above-mentioned methods are not applicable. However, in dementia such as in AD, the neuroproliferative (subgranular) zone of the dentate gyrus seems to be affected, while the physiological destination of these neurons (granule cell layer), and the CA1 region of Ammon’s horn, is the principal site of hippocampal pathology. Jin et al. (109) reported that neurogenesis is enhanced in hippocampus of patients with AD, and they reproduced their findings in a transgenic mouse model, PDGF-APP(Sw, Ind) mice, which express the Swedish and Indiana APP mutations (109). Others have shown impairment in hippocampal neurogenesis in mice carrying targeted mutations in APP, PS1, or both APP and PS1 (110).

A recent study (111) used in vivo MRS to determine the load of NPCs in both human and rat dentate gyrus. Also, this might become an attractive tool for further investigating alterations in neurogenesis capacity in neurodegenerative animal models.