Perfusion and metabolism
The techniques of single-photon emission computed tomography (SPECT) and positron emission tomography (PET) have traditionally been used to study cerebral perfusion and metabolism. Recently there has been interest in the application of advanced MRI techniques such as arterial spin labelling (ASL-MRI) and magnetic resonance spectroscopy. Unlike PET and SPECT these techniques are non-invasive and do not require intravenously injected radioactive ligands to form brain images, thus making them attractive for longitudinal studies.
Glucose metabolism has been measured using 18F-fluorodeoxyglucose (FDG)-PET. Perfusion imaging using SPECT is achieved with a variety of ligands: 99mTc-hexamethylpropyleneamine oxime (HMPAO), N-isopropyl-_p-123I-iodoamphetamine (IMP), and 99mTc-ethylcysteinate dimer (ECD). Using blood samples the FDG uptake signal can be quantified in absolute units, but is often analysed relative to a reference region, with either a whole brain average or cerebellar uptake being chosen. Perfusion, likewise, is nearly always semi-quantitative. The choice of reference region is also important; a potential disadvantage of using average whole brain (or grey matter) values as a reference is that if there is a large area of hypoperfusion unaffected areas will appear relatively hyperperfused (and the affected area will seem relatively less hypoperfused). The choice of a reference region is not completely standard, the cerebellum has been shown to be a good choice in AD since it is relatively unaffected . However, cerebellar activity may be increased in PD, and Borghammer et al.  have suggested using the white matter region as a reference, but this has not yet been widely adopted.
FDG-PET has shown widespread reductions in glucose metabolism, including in the midline frontal and parietal regions, lateral frontal, lateral parieto-temporal, and occipital cortex, in those with PD-D [77-79]. Using FDG and principal components analysis, Huang et al.  identified specific patterns of relative regional metabolism which accounted for the variability in FDG scans across PD subjects. They then identified one of these component patterns where FDG uptake correlated with cognitive function in PD. This consisted of relatively reduced metabolism in the midline frontal, precuneus, inferior parietal, and prefrontal regions, with increased metabolism in the cerebellum and pons. The same authors also demonstrated that this pattern of altered regional metabolism was associated with PD-MCI  and that longitudinal changes in it were associated with decreasing cognitive ability . Bohnen et al.  have further demonstrated reduced metabolism in the occipital cortex and posterior cingulate cortex in those with PD without dementia at baseline who then went on to develop PD-D.
SPECT studies of perfusion in PD-D report similar results to those using FDG-PET, with reduced perfusion in the midline and lateral parietal regions [84-88]; moreover, posterior parietal perfusion has been found to correlate with cognitive ability  (see Fig. 10.2) These changes are very similar to those seen in dementia with Lewy bodies (DLB) [77, 88, 90].
ASL-MRI provides a surrogate measure of cerebral blood flow and perfusion. The magnetic state of the protons in the cerebral circulation provides contrast; as the blood flows towards the brain the magnetization of the protons is inverted and when the blood reaches the brain tissue the change in magnetization is detected. In studies of AD, ASL-MRI has demonstrated patterns
Fig. 10.2 SPECT perfusion in AD and PD-D. Top row (adapted from data in Firbank et al 2003): overlaid on an axial MRI are averaged perfusion deficits—pD-D in black; Ad in white; both Ad and pDD in horizontal shading. Second row: perfusion image from a typical subject with PDD. Third row: normal perfusion image.
reprinted from Neuroimage, 20, firbank MJ, colloby sJ, burn DJ, McKeith iG, O'Brien JT., regional cerebral blood flow in Parkinson's disease with and without dementia, 1309-19, ©(2003), with permission from Elsevier.
of temporo-parietal hypoperfusion comparable with those observed with FDG-PET . In subjects with PD-D, hypoperfusion has been observed in the precuneus, cuneus, and middle frontal gyri when compared with healthy controls [92, 93], and there is preservation of perfusion in the globus pallidus, putamen, anterior cingulate, and pre- and post-central gyri .
Magnetic resonance spectroscopy provides an in vivo surrogate measure of specific aspects of cerebral metabolism. Proton magnetic resonance spectroscopy (1H-MRS) detects the small signal changes of protons from molecules of particular metabolites found within the brain on the basis that the resonance frequency is altered by the surrounding chemical milieu. These key ‘biomolecules’ include: N-acetyl aspartate (NAA), considered to be a marker of neuronal density and integrity reflecting cell metabolism and mitochondrial activity; creatine (Cr), a molecule found in all types of neuronal cell and used as a concentration reference because of its steady state; myo-inositol, principally found in glial cells acting as a major osmosregulator within these cells; and choline, a precursor of acetylcholine which is increased in membrane breakdown and found in glycerophosphorylcholine as a breakdown product of membrane phosphatidylcholine. Either multislice or single-voxel spectroscopy is performed and a large size is often chosen for the single voxel owing to the small concentrations of the measured molecules of interest . A few small studies have demonstrated that 1H-MRS may be able to identify metabolic changes in the cerebral structures implicated in PD-related cognitive decline.
An observation that supports the concept of significant ‘posterior’ cortical dysfunction driving the evolution of PD-D  is the lower NAA/Cr ratios in the hippocampus  and occipital cortices of patients with PD-D relative to subjects with PD-NC and healthy controls . In comparison with those with AD, subjects with PD-D had reduced glutamate and NAA/Cr was reduced relative to controls . A further study observed reduced NAA concentrations in the right dorsolateral prefrontal cortex compared with PD-NC  which correlated with performance on frontal subcortical tasks. In subjects with seemingly normal cognition, reduced NAA/Cr has also been observed in the posterior cingulate gyrus .
In summary, PD-D patients have markedly reduced perfusion and metabolism throughout the brain, particularly in the parietal and frontal regions. The degree of reduction in perfusion appears to be linked to cognitive decline, and patients with specific cognitive impairments also demonstrate metabolic changes. Currently the data from 1H-MRS and ASL-MRI are too limited to inform any role as possible biomarkers for cognitive decline in PD and further research is required to characterize the changes in cerebral blood flow and metabolism detectable with MRI and how these compare with changes observed with PET and SPECT.