Neuroimaging and Cognition in Parkinson's Disease Dementia

Abstract

The prevalence of cognitive impairment and dementia in Parkinson's disease (PD) is high and can potentially occur as the result of multiple differing pathologies. Neuroimaging has provided evidence of decreased cortical volume, increased white matter diffusion changes, and decreased resting metabolic activity that appears to begin prior to the onset of dementia in PD patients. Cognitive impairment has been found to be associated with multiple neurotransmitter transmission deficiencies, including dopamine and acetylcholine, indicating a widespread neurotransmitter dysfunction in PD‐related dementia. Findings of increased Pittsburgh Compound B (PiB) binding in subjects with Lewy Body Disease (LBD) compared with Parkinson's disease and dementia (PDD) may explain phenotype differences in the spectrum of Dementia with Lewy Bodies (DLB), and show promise in guiding future therapeutic trials aimed at this disease. Advances in neuroimaging now allow for the detection of volumetric, pharmacologic, and pathological changes that may assist in the diagnosis and prediction of cognitive impairment in Parkinson's patients so that better evaluation of disease progression and treatment can be obtained.

INTRODUCTION

The prevalence of dementia in Parkinson's disease (PD) is approximately 30%, with a sixfold increased risk for developing dementia in PD compared with non‐PD elderly (1). It has been demonstrated that at the time of disease onset, 25%–30% of PD patients have some degree of cognitive impairment (55), with one recent longitudinal study reporting the eventual development of dementia in up to 80% of patients with a PD diagnosis (29). Early investigations into PD‐related cognitive impairment focused on commonly observed frontal executive dysfunction. However, the involvement of multiple cognitive domains, including visuospatial functions, as well as memory and language performance (9) have subsequently been described. Theories regarding the etiology of cognitive impairment in PD include striatal dysfunction that results in secondary effects on the frontal lobe, primary frontal lobe dysfunction and more widespread cortical dysfunction secondary to global neurotransmitter system deficits (ie, cholinergic, serotonergic, norepinephrine, etc.). The underlying pathophysiology responsible for cognitive impairment in PD is variable, and likely includes the presence of Lewy bodies (LB), Lewy neurites (LN) and Alzheimer's disease (AD) pathology. While certain risk factors for cognitive impairment in PD have been discovered, the means for indentifying dementia susceptibility on an individual basis are currently unavailable. Imaging biomarkers may help aid in the diagnosis and prediction of cognitive impairment in Parkinson's patients, reveal underlying pathophysiology and neurotransmitter dysfunction, and assist in the evaluation of disease treatment and progression.

The phenotype of dementia and parkinsonism can result from multiple potential underlying pathologic correlates, including advanced AD, cerebrovascular disease, some forms of frontotemporal dementia, dementia with Lewy bodies (DLB) and mixed pathologies. For the purposes of this review, imaging differences and associations with cognitive outcomes will focus on comparisons between subjects with a clinical diagnosis of Parkinson's disease and dementia (PDD), Parkinson's disease subjects without dementia (PD‐ND), Lewy body dementia (LBD) and controls. Although LBD and PDD are thought by some to represent different points on the same spectrum of LB, or alpha‐synuclean‐associated dementias, a distinction based on the temporal course of cognitive impairment may be helpful in determining potentially divergent pathologies. Difficulties with comparisons between studies include the use of different operational definitions for dementia subgroups. Accordingly, studies included in this review have used inclusion criteria that are in accordance with the current recommendations contained within International Consensus criteria which suggest that DLB be diagnosed when dementia occurs before or concurrently with parkinsonism, and PDD be used to describe dementia that occurs in the context of well‐established PD (51).

VOLUMETRICS

Hippocampus and amygdala volumes

Numerous cross‐sectional magnetic resonance imaging (MRI) studies have demonstrated the presence of hippocampal atrophy in PDD subjects when compared with appropriate control subjects, utilizing both manual tracing 10, 18, 41, 45 and voxel‐based morphometry (VBM) 17, 37, 66 methodologies. Furthermore, smaller hippocampal volumes have been shown to be associated with poorer memory performance in PDD subjects 18, 37, 41. A similar degree of medial temporal lobe/hippocampal atrophy between PDD and DLB subjects has been reported, with a general pattern of atrophy consisting of: controls <PDD ≈ DLB < AD (68). One common difficulty in this type of study, however, is in controlling for the severity of dementia or disease state in the three dementia conditions. Hippocampal atrophy has also been observed, albeit to a lesser degree, in PD‐ND 10, 14, 18, 37, 45, 66, 68, and PD‐ND subjects with mild cognitive impairment (MCI) 53, 60, with a greater degree of atrophy being associated with worse performance on tests of memory function 10, 14, 18, 56. The more recent of these studies have shown that the hippocampal head may be affected to a greater degree than the posterior hippocampal region 10, 37, and that atrophy of the hippocamal head is associated with poorer performance on delayed recall testing in Parkinson's subjects without dementia (10) and in PD‐ND subjects with visual hallucinations (37). Atrophy of the amygdala that is associated with delayed recall performance has also has been reported in PPD (41) and PD‐ND subjects (10) vs. controls.

Neuroimaging evidence of hippocampal and amygdala atrophy can be seen in PDD and PD‐ND subjects when compared with controls, and is related to memory performance in PD subjects with and without dementia (see Table 1). The delineation of PD‐ND subjects into those who meet criteria for MCI may further help identify a subset of this cohort at greater risk for dementia conversion. Imaging findings support previously observed pathological changes of higher LB and LN density in the CA‐2‐3 field of the hippocampus of PDD patients (19), and may explain memory deficits observed in this disorder. Pathological evidence of significant amygdalar degeneration and LB formation has also been reported 22, 28 in PD patients, and is thought to be related to common clinical features of anosmia and visual hallucinations. Although imaging evidence of hippocampal and amygdalar atrophy is consistent with the formation of PD and PDD‐specific pathology (ie, LB‐containing neurons), a similar pattern of atrophy and regional pathology is also observed in early AD. Therefore, co‐existing AD pathology, which is thought to occur in 10%–60% of PD patients (48), may be contributing to these imaging findings in some PD and PDD subjects.

Table 1.

Volumetric MRI summary table. Abbreviations: PD‐ND = Parkinson's disease subjects without dementia; PD‐MCI = Parkinson's disease diagnosed with mild cognitive impairment; PDD = Parkinson's disease and dementia; AD = Alzheimer's disease.

Regional brain volumes	PD‐ND < controls	PD‐MCI < controls or PD‐no MCI	PDD < controls	PDD < PD‐ND	AD vs. PDD
Hippocampus/parahippocampus	(56)‡ (66) (68) (10)‡ (18)‡ (45) (37) (14)‡	(53) (60)‡	(10) (37) (6) (66) (17) (41) (18)‡ (45)	(6) (56) (66)	(17) (AD < PDD) (45) (PDD < AD)
Amygdala	(10)‡ (41)†		(6) (41)	(6)
Temporal lobe	(66) (6)†	(53) (6)†	(6) (68) (17)	(6) (56) (66)	(17) (AD < PDD)
Frontal lobe	(56)‡ (66) (17) (14)‡	(53) (6)†	(6) (66) (17)	(6) (56)
Parietal lobe			(17)	(6)
Occipital lobe			(6) (17)	(17)
Thalamus			(66)	(6) (56)
Caudate				(56)
Putamen			(66)
Brain stem			(6) (red nucleus)

^† Reductions did not reach statistical significance.

^‡ Atrophy was associated with performance on cognitive testing.

Lobar volumes

In cross‐sectional studies, greater atrophy of the frontal lobes has been observed in PDD subjects compared with controls 6, 17, 56, 66, with some reporting regionally specific atrophy of the anterior cingulate gyrus 56, 66. In addition to frontal lobe involvement, many have reported atrophy of the temporal lobes 6, 17, 56, 68, and less commonly, the occipital lobes (17). Similar findings of greater lobar atrophy (specifically, in the frontal, parietal and temporal lobes) have also been reported in PD patients who developed dementia early in the disease course compared with those who had a later onset of dementia (5). In one study, no volumetric differences between PDD and DLB subjects were observed (17). Subjects with Parkinson's or Lewy Body disease and mild cognitive impairment (PDLB‐MCI) have been shown to have greater frontal and temporal lobe atrophy than control subjects (53). In this study, there were no differences in MRI volumes between those with PDLB‐MCI and a mixed cohort of PDD and DLB subjects. In PD‐ND subjects, greater atrophy in the frontal lobe has been reported 14, 17, 56, 66, with an association observed between smaller prefrontal cortex and poorer performance on tests of executive function 14, 56. Greater temporal lobe atrophy has also been seen in PD‐ND subjects when compared with controls 6, 66.

One of the few longitudinal studies tracking lobar atrophy over time has reported significantly greater rates of cerebral atrophy in PPD subjects (1.2% per year) when compared with PD‐ND (0.3% per year) and control subjects (0.4% per year) with no difference in cortical atrophy rates between PD‐ND and control subjects (16). In this study, no associations between cognitive function test scores and atrophy rates were observed. This is in contrast to an earlier study in which a greater annual rate of cortical atrophy was seen in PD‐ND (0.8% per year) subjects compared with controls (0.04% per year), and rates of brain volume loss correlated with a decline on global measures of cognitive function (33). Differences in MRI methodology, subject age, duration of follow up and baseline cognitive function in the PD‐ND subjects groups likely explain, in part, the conflicting results between these two studies. In a VBM longitudinal study, PDD subjects were shown to have significant atrophy in the fusiform gyrus, parahippocampal gyrus and hippocampus, temporo‐occipital lobe and medial anterior temporal gyrus compared with baseline scans after approximately 1 year. Over that same period of time, atrophy of anterior and posterior cingulate gyrus, temporo‐occipital region, insula, hypothalamus, nucleus accumbens and hippocampus were observed in PD‐ND subjects (59).

Evidence from numerous cross‐sectional and a few longitudinal studies suggests that atrophy of medial temporal and limbic structures, with some involvement of the frontal lobe, occurs prior to the onset of overt cognitive impairment in PD subjects (see Table 1). More widespread involvement of the neocortex and subcortical regions likely occurs in relation to subsequent functional cognitive decline. The relationship between regional volumetric brain changes and performance on cognitive testing appears to be most robust in non‐demented PD subjects, perhaps prior to the onset of more global atrophy and cognitive decline. Whether or not MRI volumetric measures can be used as sensitive predictors of those at risk for dementia has yet to be determined.

RESTING BRAIN METABOLISM AND PERFUSION

Early imaging studies employing positron emission tomography (PET) to evaluate dementia in PD reported a decrease in resting glucose utilization in the temporal and parietal cortex similar to that seen in AD 44, 58, 64. Later studies confirmed the similarities between AD and PDD in glucose consumption; however, a direct comparison between the two revealed regional differences, with PDD subjects showing greater metabolic reduction in the visual cortex and relatively preserved metabolism in the medial temporal cortex when compared with AD subjects (70). Furthermore, several additional studies using both PET and single photon emission computed tomography (SPECT) have shown significantly decreased cortical metabolism in the frontal lobe in PDD subjects when compared with both controls 3, 43, 58, 62, 70 and PD‐ND subjects 58, 72, 74. Observations of decreased cerebellar and occipital glucose metabolism in PDD subjects compared with controls 2, 58 lend further support to the notion that cortical metabolic changes contributing to cognitive decline in PD are distinct from that occurring in AD. A similar pattern of diminished regional cerebral blood flow (rCBF) in PDD and LBD subjects has been observed in the parietal, frontal, occipital and temporal lobes 26, 43, with significantly decreased rCBF and metabolism occurring in the frontal lobe of DLB subjects when compared with PDD subjects 43, 74.

Although some have not observed differences in resting cerebral blood flow in PD‐ND subjects when compared with controls 62, 72, many others have reported a decrease in metabolic activity within the frontal, temporal, occipital and parietal lobes, as well as the basal ganglia and thalamus 2, 3, 32, 39, 43, 46, 58 in these subjects, indicating cortical metabolic abnormalities prior to the onset of cognition‐related functional decline. Accordingly, in subjects with PD diagnosed with mild cognitive impairment (PD‐MCI), frontal, temporal and parietal lobe hypoperfusion has been observed when compared with PD‐ND subjects (72). Links between parietal hypoperfusion and overall cognitive decline have been reported (62), including one longitudinal study which used SPECT to evaluate changes in resting cerebral blood flow in 30 PD‐ND subjects over a 1‐year period (67). At baseline, there was relative decrease in tracer activity in the right parietal cortex in PD‐ND patients compared with controls. At follow up, only parietal cortex tracer activity was decreased relative to controls and baseline rCBF measurement. In this study, six subjects converted to dementia, and the decline in Mini‐Mental State Examination (MMSE) scores was correlated with decreased mean relative tracer activity of the parietal cortex during follow up. Others have demonstrated a correlation between diminished rCBF activity in the right visual association cortex and visuoperceptual function on cognitive testing (2).

Using fluorodeoxyglucose (FDG)‐PET and neuropsychological tests, discrete patterns of glucose metabolism activity associated with visuospatial and memory function (52), and executive function 35, 46 in PD subjects have been reported. The topographic patterns associated with cognition have been found to be independent and distinct from those associated with dysphoria (52) and motor dysfunction 36, 46, and include metabolic reductions in the frontal and parietal association areas, and relative increases in the cerebellar vermis and dentate nuclei (53). This type of network analysis has been used to demonstrate a longitudinal regional metabolic decline in the medial frontal and inferior parietal lobe in early stage PD‐ND subjects followed for 48 months (36). The cognitive‐related network pattern has been shown to be significantly increased in PD subjects with multi‐domain MCI relative to PD controls subjects, with single‐domain MCI patients subjects having a cognitive expression pattern somewhere in between the two groups (34).

Resting brain metabolism and perfusion, as measured by PET and SPECT studies, have shown decreased tracer uptake in the temporal and parietal lobes of PDD patients, similar to that seen in AD. However, the additional involvement of the basal ganglia, frontal and occipital lobes distinguishes the metabolic derangements in PDD from that seen in AD, and indicates that a more global cerebral process is involved in PD‐associated cognitive decline than that explained just by frontal de‐afferentation secondary to striatal dopamine dysfunction (see Table 2). Whether or not changes in resting brain metabolism can predict those PD‐ND or PD‐MCI subjects who will eventually convert to dementia is still unknown.

Table 2.

Resting brain metabolism and perfusion summary table. Abbreviations: PD‐ND = Parkinson's disease subjects without dementia; PD‐MCI = Parkinson's disease diagnosed with mild cognitive impairment; PDD = Parkinson's disease and dementia; AD = Alzheimer's disease; PET = positron emission tomography; SPECT = single photon emission computed tomography.

Regional resting brain metabolism and perfusion	PD‐ND < controls	PD‐MCI < controls or PD‐no MCI	PDD < controls	PDD < PD‐ND	AD vs. PDD
Frontal lobe	(46) (PET) (58) (PET) (3) (SPECT) (43) (SPECT)	(34) (PET) (72) (SPECT)	(58) (PET) (70) (PET) (3) (SPECT) (62) (SPECT) (43) (SPECT)	(58) (PET) (74) (PET)	(43) (SPECT) (AD < PDD)
Temporal lobe	(32) (PET) (58) (PET) (39) (SPECT)	(72) (SPECT)	(58) (PET) (70) (PET) (3) (SPECT) (62) (SPECT) (43) (SPECT)	(58) (PET) (74) (PET)	(70) (PET) (AD < PDD) (43) (SPECT) (AD < PDD)
Parietal lobe	(32) (PET) (58) (PET) (2) (SPECT) (43) (SPECT)	(72) (SPECT)	(58) (PET) (70) (PET) (3) (SPECT) (62) (SPECT) (26) (SPECT) (43) (SPECT) (67) (SPECT)	(58) (PET) (43) (SPECT)	(43) (SPECT) (AD < PDD)
Basal Ganglia	(46) (PET) (58) (PET) (43) (SPECT)		(58) (PET)
Cerebellum	(58) (PET)	(34) (PET) (PD‐MCI > controls)	(58) (PET)
Occipital lobe	(58) (PET) (2) (SPECT)		(58) (PET) (43) (SPECT)	(43) (SPECT)	(70) (PET) (AD > PDD)
Brainstem		(34) (PET) (PD‐MCI > controls)
Hippocampus	(46) (PET)

MOLECULAR IMAGING

Dopaminergic function

Both PET and SPECT have been used to evaluate the relationship between cognitive function and regional dopaminergic tracer uptake in patients with PD. Frequently employed methods include the evaluation of dopamine decarboxylase activity, assessed with PET studies using 6‐[¹⁸F]fluoro‐L‐dopa (Fdopa) as a radiotracer. This technique allows the evaluation of the uptake and conversion of Fdopa to fluorodopamine inside the neuron and is considered a reliable method for the study of dopaminergic system function (24). Using SPECT, dopamine transporter (DAT) tracers, such as ¹²³I‐2β‐carbomethoxy‐3β‐(4‐iodophenyl)‐N‐(3‐fluoropropyl)‐N‐nortropane ([¹²³I]‐FP‐CIT) and ¹²³I‐2β‐carbomethoxy‐3β‐(4‐iodophenyl)‐N‐(3‐fluoropropyl) nortropane ([I¹²³]‐beta‐CIT) provide a measurement of presynaptic dopaminergic terminal function (24), and PET ¹¹C‐S‐Nomifensine (¹¹C‐S‐NMF) provides a presynaptic marker of dopamine denervation (63). As expected, several studies have shown diminished tracer uptake in the striatum in PDD 30, 38, 57 and PD‐ND subjects 12, 13, 30, 31, 38, 57 when compared with controls. In those with PDD, reduced caudate dopamine uptake has been shown to be associated with global cognitive impairment (57). While some have reported no association between cognitive performance in PD‐ND subjects and the degree of striatal dopaminergic depletion (12), relationships between reduced dopamine tracer uptake in the caudate and worse performance on memory 31, 40, 69 and executive, or frontal lobe function 15, 49, 54, 61, 69 have been reported. In general, decreased putaminal dopamine tracer uptake has been related to the degree of motor disability and not cognitive performance 12, 31, 61. A few, however, have reported an association between decreased putamen Fdopa uptake and worse performance on tests of memory (49) and executive function and fluency 54, 69, which may in part be related to motor actions required after some test of frontal lobe function (69). Interestingly, an increase in dopamine tracer uptake in the frontal lobe of early stage, drug naive PD‐ND subjects that was associated with worse performance on a tests of sustained attention, but improved performance on tests of suppressed attention has been reported (13). Such findings could represent an early compensatory mechanism brought on by striatal dopaminergic dysfunction, or may reflect changes in other neurotransmitter systems (13).

In one of the few longitudinal studies investigating dopamine metabolism on cognitive function in PD, investigators followed DLB, PD‐ND, PDD and control subjects over a 1‐year period with baseline and follow up ¹²³I‐FP‐CIT SPECT imaging (21). Significant decreases in striatal binding between baseline and follow‐up imaging were evident in DLB and PDD, but not in PD‐ND subjects. Diminished tracer uptake between baseline and follow‐up scans were observed in the posterior and anterior putamen in both PDD and DLB subjects, while significant reductions in caudate uptake between the two scans were seen just in those with PDD. In PDD subjects, higher annual percentage change of caudate tracer uptake was associated with greater age and degree of cognitive impairment at baseline. Rates of decreased striatal dopamine binding were greater for all patients (DLB, PD, PDD) than for controls in the caudate and posterior putamen. However, in the anterior putamen, only PDD subjects showed significant decline. Rates of cognitive decline correlated with greater loss of dopamine tracer uptake in the posterior putamen in DLB, while no relationship between changing cognition and rates of decreased striatal binding was seen in either PD or PDD.

Molecular imaging using SPECT and PET radioligands for evaluating the dopaminergic system has consistently shown diminished striatal tracer uptake in PDD and PD‐ND subjects. Decreased cognitive performance, including both memory and frontal executive function, has been shown to be related to caudate and perhaps putamen dopaminergic depletion. The role of extrastriatal dopamine uptake in PD‐related cognitive decline is unclear. Nevertheless, serial dopaminergic molecular imaging may be useful in monitoring disease progression in PDD.

Cholinergic transmission

Methods for assessing brain cholinergic function include the measurement of acetylcholinesterase (AChE) activity with [¹¹C]‐methyl‐4‐piperidyl acetate ([¹¹C]‐MP4A) or [11C]methyl‐piperidin‐4‐yl propionate ([11C]PMP) and PET, and the measurement of muscarinic acetylcholine receptor (mAChRs) with ¹²³I‐iodo‐quinuclidinyl‐benzilate ([¹²³I]‐QNB) and SPECT. Previously reported post‐mortem evidence of cholinergic dysfunction in PD 4, 73 has been subsequently observed in multiple in vivo cholinergic molecular imaging studies of PD‐ND patients 7, 8, 30, 65. Global cortical cholinergic tracer uptake has been shown to be significantly decreased in those with PDD and DLB (−20.9% to −29.7%), compared with PD subjects without dementia (−10.7% to −12.9%) relative to controls 7, 8, 30. One recent study showed no difference in brain AChE activity levels between patients with early PD‐ND and advanced PD‐ND, possibly indicating that cholinergic deficits do not progress during the course of PD without dementia (65). While a similar degree of cholinergic dysfunction has been shown in both LBD and PDD subjects 8, 21, mean cortical AChE hydrolysis rates have been reported to be significantly lower in PDD/LBD (−20%) than in AD (−9.1%) subjects (8). Regional differences in cholinergic transmission have been reported, with decreased tracer uptake observed in the posterior cingulate gyrus 30, 65, frontal 8, 30, parietal 8, 30 and temporal lobes 8, 65 in those with PDD or PDD/DLB compared with PD‐ND subjects. While some have shown decreased AChE activity (65) and cortical cholinergic transmission (8) in the occipital cortex of PD‐ND subjects, others have reported an increase in muscarinic acetylcholine receptors in this same area in those with PDD or DLB (20). In a combined cohort of PDD, DLB and PD‐ND subjects, cortical cholinergic denervation was associated with worse performance on tests of attention and executive functioning, but not memory (7), and diminished posterior cingulate cortex AChE activity correlated with poorer global cognition (65).

Molecular imaging studies have demonstrated widespread cholinergic dysfunction in PD subjects that is significantly increased in magnitude in those who also have dementia. Although present in PD‐ND subjects, cholinergic deficiency does not appear to progress with advanced motor disease in those without dementia. This finding, along with the reported relationships between AChE activity and neuropsychological testing, supports the role of cholinergic deficiency as a significant contributor to cognitive dysfunction in both PDD and DLB.

DIFFUSION TENSOR IMAGING

Diffusion tensor imagine (DTI) is capable of measuring brain tissue characteristics of local microstructural water diffusion. One DTI measurement is fractional anisotropy (FA), which describes the degree of water molecule anisotropy, or directional preference of the water diffusion process. Increased FA values indicate a greater degree of water molecule directionality, and presumably, greater microstructural integrity. In a recent study, reduced FA of the substantia nigra (SN) was found to be 100% sensitive and specific for distinguishing PD‐ND patients from healthy controls (71). Using statistical parametric mapping (SPM), others have shown regional FA decreases in the frontal lobe, including the medial frontal, cingulate (anterior and posterior) and dorsolateral prefrontal cortex (42). One of the few studies that have looked specifically at PD‐related cognitive changes in relation to DTI parameters has shown a significant FA reduction in the frontal, temporal and occipital white matter in both PDD and PD‐ND subjects compared with controls (50). In this study, PDD subjects showed a significant FA reduction in the bilateral posterior cingulate bundles compared with PD‐ND subjects that were correlated with global cognition and memory function. In patients with DLB, global white matter FA reductions have been reported by some, including involvement of the corpus callosum, pericallosal areas, frontal, parietal, occipital, and to a lesser degree, temporal white matter (11). Others, however, report abnormalities more focal to the parietal lobe (precuneus) in DLB patients when compared with controls (25).

DTI has been shown to be exquisitely sensitive and specific to microstructural changes within the SN, the site of dopaminergic depletion, in PD patients. More widespread white matter diffusion changes have also been observed in PD‐ND, PDD and DLB subjects. DTI‐detected disruption of diffusivity within the white matter may be an indicator of early microvascular ischemic disease which is associated with brain white matter hyperintensity lesions and thought to contribute to both motor and nonmotor symptoms in PD‐related dementia.

AMYLOID IMAGING

Several recent studies have employed Pittsburgh Compound B (PiB) PET imaging to elucidate the role of amyloid pathology in the etiology of PD‐related dementia 23, 27, 47. PDD subjects have been shown to have significantly decreased PiB binding compared with subjects with either AD 27, 47 or DLB 23, 27, 47 with similar dementia severity. Amyloid burden in DLB appears to be of a similar magnitude as that seen in AD (27), while the amyloid load in PDD is comparable with that observed in PD‐ND and control subjects 23, 27. While DLB subjects show a relatively widespread increase in amyloid burden compared with PDD subjects, the occipital and medial temporal lobe amyloid load appears to be similar between the two groups (27). Compared with DLB, PDD and PD‐ND subjects, there is a relative sparing of the occipital cortex to PiB‐binding derived measures of amyloid pathology in AD subjects (27). In this study, cognitive measures of visuospatial skill function was not associated with occipital amyloid load, but was correlated with greater regional amyloid burden in the parietal/posterior cingulate cortex. Greater global amyloid burden was associated with diminished global cognition in those with DLB, PDD and PD‐ND (27).

Results from PiB PET imaging reveal that cortical amyloid deposition is significantly increased in DLB, while amyloid burden in PDD subjects remains similar to PD‐ND and healthy controls. Findings suggest that amyloid pathology is a likely contributor to cognitive impairment in PD‐related dementias, with the combination of amyloid and Lewy body pathology leading to the development of clinical features of DLB. Whether or not the presence of PiB‐detected cortical amyloid in PD‐ND patients is predictive of subsequent cognitive decline has yet to be determined.

DISCUSSION

Neuroimaging has provided evidence of decreased cortical volume, increased white matter diffusion changes and decreased resting metabolic activity that appears to be global in PDD and appears to begin prior to the onset of dementia in PD patients. Cognitive impairment has been found to be associated with multiple neurotransmitter transmission deficiencies, including dopamine and acetylcholine, indicating a widespread neurotransmitter dysfunction in PD‐related dementia. In general, the imaging abnormalities seen in PDD appear to be similar in magnitude and pattern to that observed in LBD, but with some distinctions from that seen in AD. Findings of increased PiB binding in LBD subjects compared with PDD may explain phenotype differences in the spectrum of DLB, and show promise in guiding future therapeutic trials aimed at this disease. While there is some evidence that the severity of neuroimaging abnormalities may distinguish those with greater cognitive impairment, longitudinal studies are needed to determine such abnormalities are predictive of future dementia. Imaging biomarkers have thus far shown great promise in the evaluation of cognitive impairment in PD. Future neuroimaging studies including longitudinal outcomes with more comprehensive cognitive testing and pathological follow‐up will be most helpful in advancing our understanding of PD‐related dementia and better evaluate disease progression and treatment.