Parkinson's disease with dementia (PDD) and dementia with Lewy bodies (DLB) share many clinical and neurobiological similarities, including similar parkinsonian motor symptoms, neuropsychological profiles, and neurochemical and neuropathological characteristics of α-synuclein.1–4 The clinical distinction between PDD and DLB depends on the onset of cognitive symptoms relative to parkinsonian motor symptoms. However, several lines of evidence suggest that PDD and DLB may represent two distinct subtypes, although they have a similar pathological spectrum. The density of Lewy bodies and the burden of Alzheimer's disease-like pathology are more severe in DLB than PDD.1, 5–7 In addition, cortical atrophy and metabolic patterns are more severe, and cognitive performance is poorer in DLB than in PDD.8–10
Voxel-based morphometry (VBM) is an unbiased whole-brain MR technique that is capable of detecting regionally specific differences in brain tissue composition, including gray matter (GM) and white matter (WM). Previous studies comparing brain structure in patients with PDD and DLB using VBM have focused on GM, and the results have been controversial.8, 11 In this study we compared GM and WM densities in patients with PDD and DLB using VBM.
Participants were prospectively recruited from the department of Neurology at a university hospital. We enrolled 20 patients with clinical diagnosis of probable PDD and 18 patients with probable DLB. The diagnosis of DLB was made according to the consensus criteria for DLB.2 Those patients diagnosed as probable DLB who had two or more core features were included. PD and PDD were diagnosed according to the clinical diagnostic criteria of the UK Parkinson's Disease Society Brain Bank12 and the clinical diagnostic criteria for probable PDD,13 respectively. The onset of PD preceded the development of dementia by at least 12 months in all patients with PDD. Motor symptoms were assessed using the Unified PD Rating Scale Part III (UPDRS-III). We used the Seoul Neuropsychological Screening Battery (SNSB),14 an extensive neuropsychological battery test, to determine cognitive subsets in the diagnosis of probable PDD. The SNSB covers the cognitive subsets of attention (forward and backward digit span and letter cancellation tests), language and related functions (reading, writing, comprehension, repetition, confrontational naming using the Korean version of the Boston Naming Test (K-BNT),15 finger naming, right–left orientation, body part identification, calculation, ideomotor, and buccofacial praxis), visuospatial function test (drawing an interlocking pentagon and the Rey Complex Figure Test [RCTF]), verbal memory test (three- word registration and recall, and the Seoul Verbal Learning Test), visual memory test (the RCFT; immediate recall, 20-min delayed recall and recognition); frontal executive function test (motor impersistence, contrasting program, go-no-go test, fist-edge-palm, alternating hand movement, alternating square and triangle, luria loop, phonemic and semantic Controlled Oral Word Association Test (COWAT), and the Stroop test). Exclusion criteria included evidence of focal brain lesions by magnetic resonance imaging (MRI) or the presence of other neurodegenerative diseases that might account for the dementia. Possible medical comorbidities were also excluded by laboratory tests, including a thyroid function test, vitamin B12 and folic acid levels, and a VDRL test. Healthy age-matched and gender-matched, elderly volunteers were used as controls for the VBM analysis. They were recruited from advertisements about the project, or they were healthy relatives of patients with movement disorders or dementia (n = 18, age = 71.2 ± 6.5 years). The controls had no active neurological disorder. They had no cognitive complaints and a minimum score of 28 on the Korean version of the Mini-Mental State Examination (K-MMSE). Informed consent was obtained from all patients and control subjects. This study was approved by the Institutional Review Board of our hospital.
All scans of patients with PDD and DLB were acquired using a Philips 3-T scanner (Philips Intera; Philips Medical System, Best, The Netherlands) with a SENSE head coil (SENSE factor = 2). Head motion was minimized using restraining foam pads provided by the manufacturer. A high-resolution T1-weighted MRI volume data set was obtained from all subjects using a 3D T1-TFE sequence configured with the following acquisition parameters: axial acquisition with a 224 × 256 matrix; 256 × 256 reconstructed matrix with 182 slices; 220 mm field of view; 0.98 × 0.98 × 1.2 mm3 voxels; TE, 4.6 ms; TR, 9.6 ms; flip angle, 8; slice gap, 0 mm.
Voxel-Based Morphometry of GM and WM
VBM was conducted using DARTEL16 implemented in SPM8 software (Institute of Neurology, University College London, England). A group of GM and WM template was generated from control groups, to which all individual GM was spatially normalized. Spatially normalized GM maps and WM maps were modulated by the Jacobian determinant of the deformation field to adjust volume changes during non-linear transformation.17 These modulated GM and WM maps were smoothed using a 6-mm full-width half-maximum isotropic Gaussian kernel. Regional volume differences were determined using t statistics at every voxel in the GM, and WM from patients with PDD and DLB and healthy controls. Statistical significance was determined at the uncorrected p < 0.001 level, with a cluster size >50 mm3.
The Mann–Whitney U-test and Fisher's exact test were used to compare SNSB subscores between patients groups for categorical and continuous variables, respectively. A two-sided P < 0.05 was considered statistically significant. Statistical analyses were performed using commercially available software (SPSS, version 13.0).
The demographic characteristics of the patients are shown in Table 1. No significant differences were observed between patients with PDD and DLB in age, gender, years of education, duration of cognitive dysfunction, K-MMSE score, or the sum of the clinical dementia rating. The mean of UPDRS III scores was higher in patients with PDD than in those with DLB (P = 0.01). The duration of parkinsonism was significantly longer in patients with PDD (74.9 ± 60.3 months) than in those with DLB (16.7 ± 11.6 months; P < 0.001). At the time of this neuropsychological test, 19 patients with PDD and 7 patients with DLB took dopaminergic medications. For each patient, the levodopa equivalent dose, calculated as described previously,18 was significantly higher in patients with PDD (586.4 mg) than in those with DLB (171.7 mg; P < 0.001). No significant difference was found in the number of patients taking cholinesterase inhibitors (PDD, 20.0% and DLB, 27.7%). At the time the neuropsychological test was administered, 16 patients with DLB had psychiatric symptoms (4 had delusions and 15 had hallucinations) and 7 patients with PDD had psychiatric symptoms (1 had delusions and 6 had hallucinations). Visual hallucinations were more prevalent in patients with DLB than in those with PDD (P = 0.001).
Table 1. Demographic characteristics between patients with PDD and DLB
PDD (n = 20)
DLB (n = 18)
PDD: Parkinson's disease with dementia, DLB: dementia with Lewy bodies, UPDRS: Unified Parkinson's Disease Rating Scale, K-MMSE: the Korean version of the Mini-Mental State Examination, SOB: the sum of box score of the Clinical Dementia Scale, Values are expressed as mean (standard deviation), NS; not significant.
Gender (number of men)
Education durations (yr)
Parkinsonism duration (mo)
Cognitive impairment duration (mo)
UPDRS ||| scores
Psychiatric symptoms, n (%)
Levodopa equivalent dose (mg)
Use of cholinesterase inhibitors, n (%)
GM density was significantly decreased in the bilateral dorsolateral prefrontal, temporal, occipital, posterior cingular, and right parietal cortical areas in patients with PDD compared with controls (Fig. 1A). The pattern of decreased GM density in patients with DLB was similar to that in patients with PDD; however, the GM atrophy in the patients with DLB was more pronounced, especially in the occipital cortex, and included the bilateral orbitofrontal and left parietal cortices, as well as the bilateral lentiform nucleus (Fig. 1B). In a comparison between patients with PDD and DLB, GM density was significantly decreased in the left occipital, parietal, and striatal areas in DLB compared with PDD (Table 2; Fig. 1C). No area was observed where the density of GM was more severely decreased in PDD than in DLB.
Table 2. Anatomic location of areas of reduced gray matter in dementia with Lewy bodies compared with Parkinson disease with dementia
Middle occipital gyrus
Inferior parietal lobule
In patients with PDD, the decrease in WM density compared with controls was localized to the left posterior temporal, occipital, and prefrontal areas (Fig. 2A). In contrast, patients with DLB exhibited decreased WM density throughout the bilateral temporal, occipital, parietal, and frontal regions compared with controls (Fig. 2B). Thus, the decrease in WM density relative to GM density was more pronounced in patients with DLB. In a comparison between patients with PDD and DLB, WM density was significantly decreased in the bilateral occipital and left occipito-parietal areas in DLB than in PDD (Fig. 2C). No area was observed where the density of WM was more severely decreased in PDD than in DLB.
The change in WM density relative to that of GM density in patients with DLB compared with those with PDD is schematically illustrated in Figure 3. The area of WM atrophy (blue color) in the occipital areas was more extensive than that of GM atrophy (red color) in patients with DLB compared with those with PDD.
Our study demonstrated that patients with DLB had more severe GM and WM atrophy in the occipito-parietal areas than did patients with PDD. In addition, the change in WM density relative to GM density was more pronounced in patients with DLB. Since no significant difference was found in the overall severity of dementia or demographic characteristics between groups, the different patterns of GM and WM density seen in our study may reflect a difference in the underlying nature of these two diseases.
In a recent VBM study, Beyer et al.8 reported that patients with DLB had more pronounced cortical atrophy in the temporal, occipital, and parietal areas relative to patients with PDD. Our findings are consistent with those of Beyer and colleagues with the exception that the area of atrophy in our study was not as extensive. Recently, Sanchez-Castaneda et al. also demonstrated more severe cortical atrophy in DLB than in PDD, although the atrophic area was different.19 However, our results are different from those of Burton et al.11 who, in a VBM study, found no difference in the degree of cortical atrophy between PDD and DLB. Differences in demographic data, such as age, duration of dementia, duration of disease onset, or treatment, may account for discrepancies between our findings and those of previous VBM studies. In the Beyer et al. study,8 the patients with PDD had parkinsonism longer (∼12 years) before developing dementia than did the patients in the Burton et al. study (∼7 years).11 Furthermore, the duration of dementia was significantly shorter in patients with PDD than in those with DLB, which may have led to less pronounced cortical atrophy in patients with PDD relative to those with DLB, because in PDD a shorter duration of parkinsonism before dementia is associated with more severe pathological changes.20 In this study, no difference in the duration of dementia was found between the PDD and DLB groups, and the duration of parkinsonism in patients with PDD was shorter than in the Beyer et al. group. This may explain why the area of atrophy in DLB relative to PDD was not as pronounced as in our study as in that by Beyer and colleagues. In addition, we used a 3-T MR scanner and a later version of SPM than did previous studies, and these differences in MR equipment and software may also have contributed to discrepancies between our findings and those of previous VBM studies.
Controversy still exists as to whether PDD and DLB are indeed the same disease entity. The few neuropsychological studies directly comparing the cognitive profiles between patients with PDD and those with DLB have demonstrated that those with DLB showed a trend toward poorer performance in executive function, attention, or visual recognition memory,10, 21, 22 whereas other studies reported no difference in cognitive profiles.23–25 More severe atrophy in occipito-parietal areas and the lentiform nucleus in our patients with DLB relative to those with PDD may explain why patients with DLB had poorer performance in visual-related cognitive subsets or executive function. In addition, our finding that visual hallucinations were more prevalent in patients with DLB than in those with PDD may be consistent with more severe atrophyin occipital GM and WM in patients with DLB, as recent neuropathological studies have suggested that functional or pathological alterations in the temporo-occipital area may underlie visual hallucinations. Yamamoto et al.26 suggested that severe LB pathology in the secondary visual pathway and the inferior temporal cortex in DLB patients may be the cause of visual stimulus-related cognition and visual misidentification. In addition, Harding et al.7 showed that DLB patients had higher LB densities in the inferior temporal cortex than did PDD patients, and well-formed visual hallucinations are highly correlated with temporal lobe LB.
Another interesting finding in our study is that the degree of atrophy in WM relative to GM differed between patients with PDD and those with DLB. The degree of WM and GM atrophy was similar in patients with DLB, but patients with PDD exhibited less WM atrophy than GM atrophy. Additionally, the area of WM atrophy relative to that of GM atrophy in patients with DLB compared with those with PDD was more extensive in the occipital areas. WM analysis using VBM has a lower sensitivity to detect WM abnormalities because the correlation between WM T1 signal intensities and WM integrity is poor27, 28 However, simultaneous application of the same VBM technique using a 3-T MR in our patients suggests that the differences in the degree of WM atrophy relative to GM density in patients with PDD and DLB may reflect differences in the underlying pathomechanism between the two diseases, rather than technical issues. Regarding the WM pathology in patients with DLB, Higuchi et al.29 reported that WM spongiform pathology and gliosis occurred predominantly in the occipital area, and these WM pathologies were an important pathological substrate for decreased glucose metabolism in patients with DLB. Furthermore, in a study using diffusion tensor imaging, Bozzali et al.30 showed WM abnormalities in the frontal, parietal, and occipital areas in patients with DLB. Therefore, it is possible that an unknown factor, which may determine PDD and DLB distinctly, may contribute to differences in the degree of WM pathology in PDD and DLB.
Some limitations of our study need to be addressed. First, the diagnosis of PDD and DLB was based on clinical consensus criteria rather than on histopathological confirmation. This raises the possibility of misdiagnosis, especially for DLB. However, it is generally agreed that the specificity of the clinical diagnosis of DLB is high when consensus diagnostic criteria for DLB are used.2 We used the clinical diagnostic criteria for PDD suggested by the Movement Disorders Society rather than DSM-IV criteria for the PDD diagnosis. Second, even though the demographic characteristics and overall severity of dementia severity were similar between groups, the sample size was small and large intragroup variability may have limited the detection of group differences.
In summary, a VBM analysis comparing patients with PDD and DLB demonstrated that GM and WM atrophy were more severe in patients with DLB, and that WM atrophy relative to GM atrophy was less severe in patients with PDD. These data may reflect a difference in the underlying nature of PDD and DLB.
This study was supported by a grant of the Korea Healthcare technology R&D Project, Ministry for Health, Welfare & Family Affairs, Republic of Korea (A091159).
Financial disclosure: Nothing to report.
Ji E. Lee was involved in research project execution, statistical analysis execution, and writing of the first draft for the manuscript. Bosuk Park was involved in research project execution, statistical analysis execution, and writing of the first draft for the manuscript. Sook K. Song was involved in research project execution and writing of the first draft for the manuscript. Young H. Sohn was involved in statistical analysis execution, and review and critique of the manuscript. Hae-Jeong Park was involved in statistical analysis execution, and review and critique of the manuscript. Phil Hyu Lee was involved in research project conception, organization, statistical analysis review and critique of the manuscript and writing of the first draft and review and critique of the manuscript.