вход по аккаунту


Neuropathological distinction between Parkinson's dementia and Parkinson's plus Alzheimer's disease.

код для вставкиСкачать
Neuropathological Distinction Between
Parhnson’s Dementia and Padanson’s plus
Alzheimer’s Disease
Suzanne M. de la Monte, MD, MPH, Sarah E. Wells, BS, E. Tessa Hedley-Whyte, MD,
and John H. Growdon, MD
The distinctive clinical features of dementia in Parkinson’s disease (PDD) and Parkinson’s plus Alzheimer’s disease
(PD+ AD) suggest different patterns of cerebral atrophy in these conditions. To determine the pathoanatomical substrates of dementia in PDD and PD iAD, morphometric analysis of 5 standardized coronal slices was used to identify
volumetric changes in cerebral tissue. In PDD (n = 4) there were 9 to 23% reductions in cross-sectional area of
cerebral cortex, a 38% loss of tissue in the globus pallidus + putamen, and an 18% reduction in area of the amygdala,
whereas in PD+AD (n = 6) there was seyere global atrophy of the cerebral cortex (27-29% reductions), moderate
atrophy of white matter (10-19% reductions), and 40% reductions in areas of globus pallidus+putamen and the
amygdala relative to neuropathologically intact controls (n = 14). Immunostaining with anti-glial fibrillary acidic
protein disclosed significant gliosis of all four major subdivisions of neocortex in PD+AD and gray matter of the
caudate, putamen, globus pallidus, and thalamus in both PDD and PD + AD relative to controls. The findings suggest
that dementia in PDD is mainly subcortical in origin and due to neuronal degeneration in basal ganglia, the amygdala,
and thalamus. In PD + AD the same pattern and degree of subcortical degeneration is evident, but there are clearly
superimposed lesions involving cortical neurons and long projection fibers coursing through cerebral white matter
that most likely account for the distinctive manifestations of dementia in this condition compared with PDD.
de la Monte SM, Wells SE, Hedley-Whyte ET, Growdon JH. Neuropathological distinction between
Parkinson’s dementia and Parkinson’s plus Alzheimer’s disease. Ann Neurol 1989;26:307-320
The clinical features differentiating dementia in Parkinson’s disease (PDD) from Alzheimer’s disease
(AD) suggest that the targets of neuronal degeneration
within the central nervous system differ for these two
conditions. In PDD, dementia is often characterized
by slowing of mental processes, forgetfulness, impaired cognition, apathy, and depression [1-51, as observed in dementia due to Huntington’s disease, progressive supranuclear palsy, and Wilson’s disease [3, 6,
71, all of which prominently feature neurodegeneration in subcortical structures. However, dysfunction of
the frontal cortex has also been implicated in PDD [S}.
In AD these same clinical features are manifested together with prominent aphasia, amnesia, agnosia, and
apraxia reflecting degeneration of association areas of
frontal, parietal, and temporal cortex C9-141. However, the proposed distinction between subcortical and
cortical dementia [GI has limitations and is anatomically misleading { l S , 161. Despite differences in the
neuropsychiatric manifestations of cognitive impairment, the neuropathological correlates of dementia in
PDD have remained inscrutable largely because lesions of AD are often observed in patients with PDD
{17-191, lesions of PD occur in patients with AD
{20-241, and patients with AD frequently develop
signs and symptoms of extrapyramidal dysfunction
similar to those in PDD [20,25-271. Moreover, there
is considerable overlap between PDD and AD with
respect to neurochemical and neurotransmitter deficits such that in both, immunoreactive neurofilament
protein, tubulin, the microtubule-associated proteins
MAP1 and MAP2, and ubiquitin accumulate within
the perikarya of abnormal pyramidal neurons, that is,
those containing Lewy bodies in PDD 128-32}, and
neurofibrillary tangles in AD C31-37). Both manifest
degeneration of magnocellular neurons within basal
forebrain nuclei, particularly the nucleus basalis of
Meynert [l?, 38-44] with attendant reduction in
acetylcholine and choline acetyltransferase activity in
cerebral cortex [45-497. Also, there is neurodegeneration of catecholamine systems in both PDD and AD
[?, 49, SO}. Thus, these observations suggest that
From the Massachusetts Alzheimer’s Disease Research Center and
Charles S. Kubik Laboratory of Neuropathology in the Department
School, Boston, MA.
Received Nov 17, 1988, and in revised form Feb 2, 1989. Accepted
for publication Feb 4, 1989.
Address correspondence to Dr de la Monte, Warren 325, Department of Pathology, Massachusetts General Hospital, Fruit St, Boston, MA 021 14.
Copyright 0 1987 by the American Neurological Association 309
PDD and AD involve similar pathogenetic mechanisms and similar distributions of neurodegenerative
lesions within the brain, although they do not fully
account for dementia in all forms of PDD.
Controversy over the precise etiology of cognitive
impairment in PDD pivots about the clinically distinctive features and independent occurrence of dementia
in PDD and AD 173 vis-his the frequent concurrence
of PDD and AD, together with their overlapping
neurochemical and neurotransmitter abnormalities.
One approach to this problem is to compare the distribution and degree of neurodegeneration-for example, the gross atrophy and cell loss within the cerebral hemispheres of demented patients with only
histopathological lesions of PDD with those of demented patients with lesions of both PDD and AD.
However, with the present routine methods of evaluating cerebral atrophy, that is visual inspection and
brain weight, it would be difficult to delineate and
quantitate such differences. To address this issue we
devised a simple and reproducible method of performing morphometric analysis of the cerebrum using photographs of coronal slices made at standardized levels.
With this protocol we have been able to characterize
the nature, distribution, and extent of atrophy within
the cerebrum in a number of important neurodegenerative disorders associated with dementia including
Huntington’s disease [ S 11, Alzheimer’s disease 1521,
Down’s syndrome 1531, and chronic alcoholism 1541.
In the present study we employed this technique to
elucidate the patterns of regional brain atrophy in
PDD and PD+AD. In addition, to corroborate the
neurodegeneration detected by gross morphometric
analysis we immunostained cerebral tissue with antiglial fibrillary acidic protein (GFAP) to demonstrate
ghosis in grossly atrophic structures. The degree of
ghosis determined in this way provides an indirect estimation of neuronal loss that might not be appreciated
in routine histological sections. The findings revealed
substantial degeneration of subcortical nuclei in both
PDD and PD + AD, marked cortical and white matter
atrophy in PD+AD, and mild to moderate cortical
but no white matter atrophy in PDD. While neuronal
degeneration within subcortical nuclei most likely accounts for dementia in PDD, the distinctive clinical
features of cognitive impairment in P D + A D compared with PDD are best explained by the superimposed global degeneration of cerebral cortex and white
Tissue Procurement
Brains with PDD and P D + A D were obtained from demented patients with Parkinson’s disease examined in the
MovemendMemory Disorders Unit at the Massachusetts
General Hospital or in the Clinical Research Center at the
310 Annals of Neurology Vol 26
Massachusetts Institute of Technology. Brains from patients
with no underlying neurological disease (controls) were obtained through the autopsy service at the Massachusetts General Hospital. As part of the Alzheimer’s Disease Research
Center (ADRC) brain-banking protocol the brains were
weighed fresh and then divided in the mid-sagittal plane. The
ventricular volumes were measured in both halves of the
fresh brain by determining the quantity of water required to
fill the ventricular system to capacity. This was done by first
emptying ventricles of cerebrospinal fluid, and then with the
hemisphere placed on a board, medial side upward and supported by moist towels or a pair of hands, a 60-ml syringe
containing tap water was used to fill the ventricular system to
One half of the brain (randomized) was sliced fresh in the
coronal plane at 0.5- to 1.0-cm intervals using standardized
landmarks, photographed (a ruler was included in the view to
determine magnification), and then blocked taking adjacent
sections for histopathological and immunocytochemical analysis and storage at -70°C in the ADRC brain bank for
future biochemical and molecular biological studies. The
other half was either similarly processed or fixed intact in
10% neutral buffered formaldehyde solution for 2 to 4
days and then sliced in the coronal plane, photographed,
and blocked for histopathological examination. Paraffinembedded histological sections stained with Lux01 fast bluehematoxylin and eosin, Bodian silver impregnation, and
Congo red were used for diagnostic assessment in all cases.
The diagnosis of PDD, with or without superimposed AD,
was rendered if there was evidence of neuronal loss, gliosis,
and Lewy bodies in the substantia nigra and locus ceruleus.
The diagnosis of coexisting AD was made using the
threshold criteria suggested by Khatchaturian [ 5 57. Thus, the
AD and PD components of the P D + A D diagnosis were
established independently. In contrast, both the PDD and
control brains were essentially devoid of AD lesions, except
for rare neuritic plaques or neurofibrillary tangles. In all
cases the histopathological sampling was extensive. A total of
4 3 blocks from symmetrical regions of the cerebral hemispheres, brainstem, cerebellum, and spinal cord were examined both to establish a diagnosis and to determine the presence of superimposed neuropathological processes unrelated
to the primary neurodegenerative disease. For control brains,
this degree of sampling permitted us to exclude the presence
of significant neuropathological lesions such as ischemic infarction.
Morphometric Analysis
Morphometric analysis of the cerebrum was performed using
photographs of the fresh coronal slices from one hemisphere
as described previously 15 1-54). The five standardized levels
used for morphometric analysis included the following: anterior frontal region, approximately 1 cm anterior to the temporal poles (AF); head of caudate nucleus with the putamen
and nucleus accumbens (CAP); globus pallidus with the putamen, body of caudate nucleus, and amygdala (GP); hippocampus with the lateral geniculate nucleus and centrum
medianum nucleus of the thalamus (LGN); and the parietooccipital fissure (OCP).
The cross-sectional areas of the entire coronal slice, the
No 3 September 1989
cerebral Cortex, the white matter, the SUbCOrtiCd nuclei, and
the ventricular system were measured or computed from
high contrast, 11 x 14-inch black-and-white mat photographic prints. This was done by digitizing the perimeters of
the entire slice (CTX-OUT), the innermost portion of cerebral cortex at the cortical-white matter junction (CTX-IN),
the boundaries of each subcortical nuclear structure
(NUCL), and the walls of the ventricles. Cross-sectional
areas were computed directly using a morphometry software
package that integrated area from the sum of pixels (picture
elements) generated by digitizing the boundaries of an object
(see below). The cross-sectional areas of the various smctures used in the data analysis were measured directly or
computed as follows:
Entire cerebral slice = CTX-OUT,
Cerebral cortex = (CTX-OUT) - (CTX-IN);
Total nuclear mass = (TNUCL) sum of areas of each nuclear
Ventricular system = (TVENT) sum of areas of all components of ventricles at a given level;
Cerebral white matter = (CTX-IN) - (TNUCL
ploying the protocol suggested by the manufacturer (Vectastain Burlingame, CA). The immunostaining reaction product including that located both perikaryally and within
processes in gray matter, white matter, and at cortical graywhite junctions was gauged semiquantitatively on a scale
from 0 to 3 as follows: 0 = absent; 1 = rare positive cells
(approximately 1 or 2 per ten 400 x microscopic fields) or
occasional small aggregations of GFAP+ fibrils; 2 = GFAP+
labeling intermediate in degree between grades 1and 3; 3 =
multiple GFAP+ cells in at least five of ten 400 X microscopic fields, or a dense meshwork of GFAP+ fibrils in the
neuropil. For regions too small to encompass ten 400 x
fields, grade 1 GFAP+ labeling was assigned if positive cells
and fibers were rare and scattered, grade 3 if multiple positive cells and fibers were observed in virtually every field,
and grade 2 if the degree of immunolabeling was between
that for grades 1 and 3. Grading was based on the maximum
level of immunostaining in a given region; however, the immediate subpial and subependymal layers were excluded
from this assessment. The immunostaining was graded without knowledge of the clinical history or data derived from
morphometric analysis.
Immunohistocbmical Analysis
Population Profile
Among the 24 subjects included in this study, 4 (17%)
had PDD, 6 (25%) had PD+AD, and 14 (58%) were
neurologically intact controls with no diagnostic abnormalities in their brains by gross and microscopic examination. The demographic features, underlying diseases,
and causes of death among these groups were similar
(Tables 1 and 2). Atrophy of the brain, reflected by
reduced brain weight and enlarged ventricles, was
more severe in patients with P D + A D than in those
with PDD only. The mean brain weight in PDD was
6% less than control, whereas in P D + A D the mean
brain weight was 17% less than control. Although the
differences in mean brain weght and ventricular volume among the three groups were not statistically
significant by ANOVA testing, there was a positive
trend toward reduced weight and larger ventricles in
diseased brains (0.05 < p < 0.10).
With longstanding, slowly progressive degeneration in the
central nervous system it may be difficult to detect neuronal
loss because of collapse and remodeling of the parenchyma
with attendant reestablishment of normal neuronal densities.
One way to ascertain neuronal loss indirectly is to demonstrate increased astrocytic activity, that is, gliosis in the tissue.
Immunostaining for GFAP is a sensitive assay for gliosis. To
determine whether cerebral atrophy was associated with histopathological evidence of cell loss, paraffin-embedded sections from Brodmann’s areas 8/9 in the frontal lobe, 3,1,2
and 40 in the parietal lobe, 2012 1 in the temporal lobe, and
17/18 in the occipital lobe, and the dorsal thalamus (with
centrum medianum nucleus), caudate and putamen at the
level of the nucleus accumbens, and globus pallidus with
putamen at the level of the amygdala were immunostained
with a mouse monoclonal antibody to GFAP using the
avidin-biotin-horseradish peroxidase complex method em-
Patterns of Cerebral Atrophy in PDD and PD +AD:
Changes in Absolute Cross-sectionalAreas
Overall trends regarding differences in cross-sectional
areas of the cerebrum and its components were initially analyzed using two-way ANOVA and multiple
correlation analysis to detect inhomogeneity among
the three groups. The total cerebral slice areas in
P D + A D and PDD (at the 5 levels studied) were
significantly smaller than in controls ( r = 0.35; p <
0.001). These differences were associated with significant global reductions in the absolute cross-sectional
areas of cerebral cortex (r = 0.58; p < 0.001) and
subcortical nuclei (r = 0.37; p < 0.011, a modest reduction in the mean area of white matter ( r = 0.21;
p = 0.08), and significant increases in the cross-
At corresponding levels, relative cross-sectional areas were
computed from the ratios of cortex, white matter, ventricular
system, and subcortical nuclei to area of the cerebral slice,
and the ratio of the areas of cortex and white matter.
The digitizing was performed by tracing structures with a
hand-held cursor on a digitizing pad with data transmitted
directly to a minicomputer. Computations were carried out
using the Bioquant System IV morphometry software package (R&M Biometrics, Inc, Nashville, TN) interfaced with
an IBM AT computer. Each structure was digitized twice,
and the averaged values were used for data analysis. Data are
expressed as proportions (percentages) or mean standard
error. The data were analyzed statistically with Student’s t
tests, analysis of variance (ANOVA), correlation matrixes,
and linear regression using the SYSTAT programs (1986;
Evanston, IL) interfaced with an IBM PC computer. The data
analyses were focused on characterizing and quantitating regional losses of cerebral tissue in PDD and PD AD.
de la Monte et al: Cerebral Atrophy in PDD and P D + A D
Table 1. Comparison of Brain Weight and Ventricalar V O I M
in ~PDD and PD +AD”
Brain Weight
Patient Group
No. of Subjects
73.1 t 3.6
79.3 t 2.3
78.5 % 3.7
“Values expressed as mean
Ventricular Volume
31.3 ? 3.2
37.8 & 9.2
39.8 +- 6.4
standard error.
PD = Parkinson’s disease, PDD = PD dementia, PD + AD = Parkinson’s plus Alzheimer’s disease
Table 2. Un&&ng
Diseases and Cawes of Death“
Chronic Diseases
Organ or System
(n = 14)
(n = 4 )
Causes of Death
(n = 6 )
(n = 14)
(n = 4 )
(n = 6 )
“Values expressed as frequency
None of the differences between groups was staristically significant.
PDD = PD dementia, PD + AD = Parkinson’s
+ Alzheimer’s disease.
sectional areas of the ventricular system ( r = 0.32; p
= 0.006). At the same time, the relative area of cerebral cortex-the percentage of the total cerebral area
that was cortex-was globally reduced ( r = 0.32; p =
0.003), and the relative areas of white matter and the
ventricles were increased ( r = 0.35 and p = 0.001,
and r = 0.39 andp < 0.001, respectively) in PDD and
P D + A D compared with controls (Fig 1). Detailed
data analysis was performed to characterize further the
distribution of atrophic changes in the cerebrum of
patients with PDD and PD + AD.
In the control group, mean cross-sectional area of the
cerebrum ranged from 3 1.4 to 49.8 cm’, the area of
cerebral cortex ranged from 15.5 to 21.5 cm2, white
matter 15.4 to 20.7 cm2, subcortical nuclei 4.5 to 7.5
cm’, and the ventricular system 0.5 to 1.3 cm2. In
PDD, at the AF, CAP, GP, and OCP levels the overall
cross-sectionalarea of cerebrum was reduced 1 to 9%
and the cerebral cortex, 9 to 23% (Table 3; see Fig 1).
In addition, there was striking atrophy (28%) of subcortical nuclei at the GP level where the globus pallidus and putamen were markedly atrophic (each reduced 38%), and the amygdala was moderately (18%)
atrophic. The selective nature of these atrophic
changes in PDD was manifested by the relative preservation, that is, normal size, of the caudate nucleus at
the G P level, all subcortical nuclei at the CAP and
LGN levels, and white matter at all 5 levels. The hydrocephalus detected by measuring ventricular volume
at the time of autopsy was associated with 48 to 113%
increases in mean cross-sectional area of the ventricular system at the four most anterior slice levels. At the
OCP level the mean ventricular area was only 0.09
cm’, which was 82% smaller than that in the control
In PD + A D the severity and distribution of cerebral
atrophy were more striking than in PDD (see Fig 1 and
Table 3). The mean cross-sectional area of cerebrum
was reduced by 21% at the AF and G P levels, 18% at
the CAP level, 9% at the LGN level, and 21% at the
OCP level compared with controls. Shrinkage of the
cerebral hemispheres was associated with marked generalized atrophy of cortex at a l l 5 levels (27-29%);
moderate atrophy of white matter at the AF, CAP,
GP, and OCP levels (10-1996); and varying degrees
of atrophy in subcortical nuclei at the CAP (12%), GP
(36%), and LGN (7%) levels. At the G P level, there
was marked atrophy of the globus pallidus+putamen
312 Annals of Neurology Vol 26 N o 3 September 1989
w w
+ a
I u
Fig 1. Pattemr of cerebral atrophy in demented patients with
Parkinson's diseuse (PD) or PD plus Alzheimer's disease
(PD +AD) compared with neuropatbologicalb intact controls.
Cross-sectionalareas of one cerebral hemisphere, its cortex, white
mutter, subcortical nuclei, and ventricles were determined by morphmetric analysis of coronal slices made at 5 standardized levels
including the fillwing: AF = 1 cm anterior to the temporal
poles; CAP = beau' of the caudate and nucleus accumbens; GP
= globus padidus and atrtygakla; LGN = lateral geniculate
nucleus; OCP = parietooccipitaljssure. Area (dis
)represented on the ordinate, and coronal slice h e / on the abscissa. In
PD the total cerebral areas were similar t o control at all 5 levels,
but signtjkant atropb of the cortex was presetzt at the AF,
CAP, and OCP levels. In PD A D there was significant global
atrophy of the cerebrum at all levels, which was associated with
marked atrophy o f cerebral cortex and moderate atrophy ofwhite
matter. The white mutter was not significantly atrophic in PD.
in both PD and PD + A D subcortical nuclei at the GP level
were signz$cantly atrophic due to marked reductions in the areas
of the putamen, globus pallidus, and amygdala. Cerebral atrophy
in PD and PD -k A D was as.sociated with striking enlargement
ofthe ventricular system at all or mst levels.
de la Monte et al: Cerebral Atrophy in PDD and PD+AD
T’abie 3. Patterns of Cerebral Atrophy in PDD, PD +AD, and AD: Changes in Absolute CrossSectional Areas
Lesion in PDD
Lesion in PD + AD
Lesion in AD
Mild atrophy (1-996) at all
but LGN level
Global atrophy (7-19%)
Cerebral cortex
Variable atrophy (9-23%) at
all but LGN level
Global atrophy (18-21%) at
all but LGN level; 9% atrophy at LGN level
Marked uniform atrophy
(27-29%) with gliosis
Moderate atrophy (10-19%)
at all but LGN level
White matter
Subcortical nuclei
Marked atrophy of
putamen +globus pallidus
(38%); moderate atrophy
of amygdala (18%). Gliosis
of basal ganglia, thalamus,
and amygdala
Marked dilation (48-113%)
at all but OCP level
Marked atrophy of
putamen +globus pallidus
(39%) and amygdala
(42%). Gliosis of basal ganglia, thalamus, and amygdala
Marked dilation (22-184%)
Moderate atrophy (13-2496)
at all levels
Mild-moderate atrophy (319%) throughout
Moderate atrophy of amygdala (24%), other nuclei
Massive dilation (150-29095)
PDD = PD dementia, AD = Alzheimer’s disease, P D + A D = Parkinson’s plus Alzheimer’s disease, LGN = hippocampus with the lateral
geniculate nucleus and centnun medianum nucleus of the thalamus, OCP = parietooccipital fissure.
(39% reduction measured together) and the amygdala
(42%}, whereas the area of the caudate nucleus was
normal. Brains with PD+AD also manifested increases in the mean cross-sectional areas of the ventricular system ranging from 22 to 184%, but unlike PDD
the most prominent ventricular dilation in PD+AD
was at the OCP level (184% areal increase).
Compared with PDD, in PD + AD the cerebral slice
areas were 14 to 18% smaller; the cerebral cortex, 8 to
27% smaller; white matter, 6 to 21% smaller; and
subcortical nuclei, 10 to 13% smaller at all levels.
However, at the GP level the areas of the caudate
nucleus and globus pallidus + putamen were essentially
the same in PDD and PD+AD, but the amygdala
was disproportionately more atrophic (29% smaller) in
PD+AD. With regard to the ventricular system, at
the AF, CAP, and LGN levels ventricular areas
in PD+AD were 22 to 39% smaller than in PDD,
whereas at the GP level the ventricular area was the
same as that in PDD, and at the OCP level it was 14
times larger than that in PDD.
Changes in Relative Proportions of Cerebral Structures in
PDD and PD +AD
Relative cross-sectional areas were computed from the
ratios of the area of a given structure, such as cortex, to
the overall area of the corresponding cerebral slice.
These quotients (reported in percentages) reflected the
proportion of a cerebral slice composed of cortex,
white matter, subcortical nuclei, and ventricular system, and therefore could be used to determine
whether the degree of atrophy in any given structure
was disproportionate compared with other structures
at the same or other levels. In addition, one could
determine whether the degree of ventricular enlargement was comparable to the degree of cerebral at-
rophy or disproportionate and perhaps primary in origin, as noted previously in Huntington’s disease I5l}.
cerebral cortex constituted between 42 and 56%, and
white matter 39 to 4995, of the cross-sectional area at
the 5 cerebral-slice levels studied. Correspondingly, at
the AF, GP, and LGN levels, cortex and white matter
had approximately equal cross-sectional areas (ratios of
1:l), but at the CAP and OCP levels the cortical area
exceeded white matter by 24% and 3996, respectively.
In PDD the relative area of cortex was consistently less
than that in controls ( p < O.Ol), but the degree of this
relative atrophy varied with the slice level (Table 4).
At the AF, GP, and LGN levels, the relative area of
cortex was reduced only slightly-2 to 7%-but at the
CAP and OCP levels, the relative areas of cortex were
reduced by 14 to 15% ( p < 0.05). In contrast, the
proportion of total slice area composed of white matter
was 4 to 12% greater than control at the 4 most anterior slice levels; only at the OCP level was the relative
area of white matter reduced (20%). Corresponding
with the selective atrophy of cerebral cortex and general sparing of white matter, the ratios of the areas of
cortex and white matter were decreased by 9 to 23%
at all 5 slice levels in PDD relative to control ( p <
In PD AD the relative area of cortex was reduced
by 8 to 19% compared with controls, and although the
absolute area of white matter was also reduced, its
relative area was in fact increased by 3 to 17% because
at corresponding levels the degree of cortical atrophy
exceeded the degree of white matter loss. Only at the
OCP level was the relative area of white matter less
than control by a factor of 20%. Correspondingly, the
314 Annals of Neurology Vol 26 No 3 September 1989
Table 4. Changes in Relative Froportions of Cerebral Strgctures in FDD, FD +AD, and AD
Lesion in PDD
Lesion in PD+AD
Lesion in AD
15% reduction at CAP and
OCP levels; 2-7% at AF,
GP, and LGN levels
4-12% increases at all but
OCP level; 20% reduction
at OCP level
Reduced 25% at G P level;
normal at CAP and LGN
Marked disproportionate hydrocephalus
Moderate reductions (819%) throughout
4-13% reduction at all levels
3-17% increases at all but
OCP level; 20% reduction
at OCP level
Reduced 19% at G P level;
normal at CAP and LGN
Marked disproportionate hydrocephalus
White matter
Subcortical nuclei
Marked disproportionate hydrocephalus
PDD = PD dementia; AD = Alzheimer‘s disease; PD AD = Parkinson’s plus Alzheimer’s disease; CAP = head of caudate nucleus with the
putamen and nucleus accumbens; OCP = parietooccipital fissure; AF = anterior frontal region approximately 1 cm anterior to the temporal
poles; GP = globus pallidus with the putamen, body of caudate nucleus, and amygdala; LGN = hippocampus with the lateral geniculate nucleus
and centmm medianum nucleus of the thalamus.
mean areal ratios of cortex and white matter in
PD AD were 13 to 32% lower than control at all 5
In PDD and PD AD the relative areas of cerebral
cortex were comparable at the AF, CAP, GP, and
OCP levels, but at the LGN level the cortex was relatively more atrophic (18%) in PD+AD. The relative
areas of white matter were also comparable between
PD + AD and PDD at the CAP, GP, and OCP levels,
but at the AF and LGN levels they were, respectively,
9% and 8% less in PD-tAD. Thus, the cortedwhite
matter ratio was increased in PD+AD relative to
PDD only at the AF (10%) and LGN (25%) levels,
whereas at the CAP, GP, and OCP levels this ratio was
similar for the two groups.
brains, subcortical nuclei constituted 11 to 15% of the
total cerebral slice area. In PDD and PD+AD the
relative areas of subcortical nuclei were normal at the
CAP and LGN levels, but at the GP level they were
reduced by 25% and 19%, respectively, because of
selective atrophy of the globus pallidus putamen and
the amygdala.
In the control group, the ventricles constituted between 1 and 3% of the total cross-sectional area of the
cerebral slice. In both PDD and PD + A D the ventricles constituted up to 5 or 6% of the total cerebral
slice area, representing maximum relative areal increases of 132% and 193%, respectively. However,
no consistent pattern emerged with respect to slice
level and degree of relative hydrocephalus.
Degree of Gliosis Manifested by GFAPt
In control brains, GFAPt immunolabeling was predominantly fibrillar and seldom associated with hyper-
trophic astrocytes, except within incidental microscopic foci of ischemia. In controls, immunolabelingof
glial fibrils was particularly evident in perivascular
spaces, the subpial and subependymal zones, and at the
cortical gray-white matter junction. In both PDD and
PD +AD, GFAP+ fibrils were similarly distributed,
but in addition there were scattered individual and
small foci of hypertrophic GFAP+ astrocytes within
the neuropil and central white matter. Also in PD +AD
were circumscribed pencil-like foci of &al fibrils as
described previously for amyotrophic lateral sclerosis
[56]. At times these foci were associated with neuritic
plaques, but frequently they occurred independently
of plaques.
Three-way ANOVA disclosed significant inhomogeneities among the three groups with respect to the
degree of GFAP+ immunolabeling in cerebral cortex
from Brodmann’s areas 8/9 in the frontal lobe ( p <
0.05), 3,1,2 and 40 in the parietal lobe ( p < 0.001),
and 20/21 in the temporal lobe ( p < 0.05), but not in
primary and associative visual cortex (areas 17 and 18;
p > 0.10). Similarly, the degree of GFAP+ immunolabeling in gray matter of all subcortical nuclei examined at the CAP, GP, and LGN levels, including
caudate nucleus, putamen, globus pallidus, and thalamus, varied significantly among the groups ( p < 0.05
to 0.001). In contrast, there were no significant trends
regarding the degree of gliosis in central white matter,
white matter comprising the intracortical fibers at the
gray-white junctions, or white matter among subcortical nuclear structures (all p > 0.10).
In PDD, relative to controls, strikingly greater degrees of GFAP+ immunolabeling were observed in
gray matter of the thalamus, caudate nucleus, putamen,
and both inner and outer segments of the globus pallidus ( d p < 0.05) (Figs 2 and 3). In the parietal cortex
(areas 3,1,2 and 40) there was also a positive trend
toward increased gliosis in PDD, but the differences
de la Monte et al: Cerebral Atrophy in PDD and PD+AD
Fig 2. Gliosis of cerebral structures in Parkinson’s disease dementia (POD) and Parkinson’s plus Alzbeimds disease
(PD +ADj. The degree of immunokzbelingfor glialjbrilkwy
acidicprotein (GFAP) in cortex, white mutter, and the cortical
graywhite junction in Brodmunn’s areas 8/9,3,1,2, 40,20121,
and 17118, and in both gray and white matter Of the thalamus
(lwelofthe centrum medianum nucleus) (THAL),head of cauLte nucleus (CAUD),putamen (PUTAM),and both inner
(GP-IN)and outer (GP-OUT) segments oftbe globus pallidus
was assessed semiquantitatively on a scale fm0 to 3 (see text).
The means standurd errorfor the dtflerent groups are plotted
separateb and the values are indicated along the ordinate. C m paved with controls, GFAP’ immunokzbeling was significantly
inmeased in cortical gray matter of areas 819, 3,1,2,40, and
20121 in PD +A D and in areas 3,1,2 and 40 in PDD. In
contrast, there were no consistent or significant trends with respect to the degree afGFAP+ immunolabeling in subcortical
white mutter or at cortical gray-whitejunctions. In subcortical
nuclei, gray matter exhibited simikzr degrees of increased
GFAP’ immunolabeling in PDD and PD +AD, but white
matter showed no distinct pattern of GFAP+ immunolabeling.
I t is of interest that the degree of gliosis in subcortical nuclei was
greater, that is, more extensive, than the atrophic changes detected by movphometric analysis, whereas gliosis of the cortex corresponded well with gross corticul atrophy.
316 Annals of Neurology Vol 26 No 3 September 1989
Fig 3. Gliosis ofthe globus pallidus (A-C) and Brodmann’s
area 40 o f the cerebral cortex ( 0 - F ) in Parkinson’s diseuse dementia (POD) and Parkinson’s plus Alzheimer‘s disease
(PO +AD) demonstrated b~ imunostaining with anti-glial
f i b t r h r ~acidic protein (GFAP).As illustrated, GFAP’ labeling was genera& absent in neuropil ofthe cerebral cortex (0)
and gray mutter of the globus pallidus (A)(as well as other
subcortical nuclei) in control bruins. In PDD gliosis ofthe cwe-
bral cortex was either absent or mild and foal (I?),whweas in
PD A D cortical gliosis was generally mo&rate, although also
patcby or focal in distvibution (F).Gliosis in the globu pdldus
was marked in both PDD (B) and PD + A D (C). In addition,
in both PDD and PD +A D variable degrees ofgliosis were observed in gray matter of all other subcortical nuclei, even those
not grossly atrophic b~ morphmtric anatysis. ( x 500 befoe
20% reduction.)
de la Monte et al: Cerebral Atrophy in PDD and PD+AD 317
did not reach statistical significance because of the
large variance resulting from the small number of patients included in each group. In contrast, the degrees
of GFAP’ immunolabeling in the frontal, temporal, and occipital cortex were similar for PDD and
In PD AD the degree of GFAP+ immunolabeling
was significantly increased in cerebral cortex from
Brodmann’s areas 8/9 ( p < 0.05); 3,1,2 (p < 0.001);
40 ( p < 0.001); and 20/21 ( p < 0.05),but not areas 17
and 18 relative to controls. Increased GFAP+ immunolabeling in PD AD relative to controls was also
detected in the thalamus ( p < 0.05), especially the
dorsal and lateral nuclear groups; caudate nucleus ( p <
0.005); putamen ( p < 0.001); and both inner and
outer segments of the globus pallidus ( p < 0.001) (see
Figs 2 and 3). Compared with PDD, GFAP+ immunolabeling in PD +AD was significantly increased
only in areas 3,1,2 ( p < 0.05),40 ( p < 0.005), and 201
21 ( p < 0.05), and not in either the frontal or occipital
cortex or any of the subcortical nuclei. Although not
studied systematically by GFAP immunostaining, the
routine histopathological sections disclosed fibrillary
ghosis of the amygdala in both PDD and PDt-AD,
but not in controls.
In this study the distinctive histopathological features
of PDD and PD + AD were associated with different
patterns and degrees of cerebral atrophy as demonstrated by morphometric analysis. In PDD, although
gross brain weight was reduced only 5% and cerebral
atrophy was inapparent by subjective visual inspection,
quantitation of cross-sectional areas disclosed moderate atrophy of the cerebral cortex in the anterior frontal (AF and CAP levels) and posterior parietal (OCP
level) regions and marked atrophy of subcortical nuclei
including the neostriatum, thalamus, and amygdala. In
retrospect, the only indication of cerebral atrophy
gauged by routine postmortem examination was the
increased ventricular volume (mean 2 1%) relative to
controls, but this crude measurement could not have
helped predict the distribution or severity of cerebral
degeneration, as hydrocephalus was equally severe in
PD +AD, yet the pattern of atrophy differed from that
observed in PDD. The patterns and severity of cerebral atrophy in PDD and PD+AD could not be accounted for on the basis of underlying systemic disease
or cause of death, since there were no statistically
significant differences among the groups with respect
to these variables.
In PD+AD gross brain weight was lower than in
PDD, but also compared with AD [52}. In fact, the
15% mean reduction in brain weight was approximately equal to the combined effect of cerebral atrophy independently associated with PD and AD.
318 Annals of Neurology Vol 26
Morphometric analysis disclosed diffuse and somewhat
uniform atrophy of cerebral cortex, which contrasts
with the predominantly frontal atrophy in PDD and
the posterior frontal and parietal atrophy of AD. In
both PDD and PD+AD, gliosis of cortex demonstrated by immunohistochemical labeling for GFAP
reflected cell loss corresponding with the atrophy detected by morphometric analysis. It is of interest that
the greater degree and more widespread atrophy of
the cerebral cortex in PD +AD compared with PDD
was corroborated by the more extensive gliosis of the
cortex in the former.
As in AD and unlike PDD, PD +AD was also associated with atrophy of cerebral white matter (see
Tables 3 and 4), most likely because of degeneration of
long projecting fibers manifested by rarefaction and
irregularity of axons [52, 571 and aberrant constipation
of cytoskeletal proteins such as tau within the cell soma
and dendrites rather than within axons [58}. Unlike
atrophy of the cortex, atrophy of cerebral white matter
was not associated with signlficantly increased ghosis.
This may indicate that white matter atrophy occurs
secondary to degeneration and death of neuronal cell
bodies with attendant reduction in the density of long
projection fibers rather than primary (intrinsic) disease
or glial cell death within white matter. In other words,
white matter atrophy in both PD + AD and AD may
be best explained on the basis of either anterograde or
retrograde degeneration of axons. Given the fact that
wallerian degeneration is associated with ghosis of
white matter, it is unclear why significant astroghosis
was not a component of white matter degeneration in
PD+AD, but it may be related to the slow progression of the degenerative process.
The most striking abnormality in PDD and
PD+AD was the marked atrophy and ghosis of the
lenticular nuclei, which were conspicuously absent in
AD 1521. However, neurodegeneration was not restricted to these subcortical nuclei, since marked
gliosis was also evident in the caudate nucleus and
thalamus in both PDD and PD+AD, despite the fact
that these structures were not atrophic by morphometric quantitation. That the caudate nucleus (both
head and body), putamen, globus pallidus, and thalamus exhibited similar degrees of atrophy or gliosis or
both in PDD and PD+AD implies that these structures are targets of neuronal degeneration in PDD
dementia, and moreover, no additional neurodegeneration in these subcortical nuclei arises from superimposed AD ([52); also see Tables 3 and 4). This is in
contradistinction to the more pronounced atrophy of
the amygdala and hippocampus that occurs in PD + AD
(42%) compared with PDD (18%). In Huntington’s
disease these same subcortical nuclei are atrophic, but
the overall degree of atrophy is far more severe-20
to 30% in the amygdala and thalamus and 60% in the
No 3 September 1989
caudate nucleus, putamen, and globus pallidus-and
accompanied by 20 to 30% areal reductions in both
cerebral cortex and white matter {Sl). In contrast, in
cerebral atrophy due to chronic ethanol abuse the subcortical nuclei are preserved [54}.
Because atrophy and ghosis of the cerebral cortex in
PDD were mild compared with that in the putamen,
globus pallidus, thalamus, and amygdala, the cause of
dementia in this condition is more apt to be related to
cortical dysfunction resulting from degeneration of
subcortical nuclei rather than to primary lesions in the
cortex. The design of this study did not permit determination of which particular subcortical nuclei are implicated in the pathogenesis of dementia in PDD.
Clearly, future investigations along these lines should
focus on the basal ganglia, amygdala, and thalamus, as
well as other nuclear structures not examined herein,
including the locus ceruleus, the dorsal raphe nucleus,
and nucleus basalis of Meynert, in which neurodegeneration has also been demonstrated in PDD {59]. In
PD +AD, the neurodegenerative lesions appear to be
caused by the superimposition of two disease processes, PD and AD. This point cannot be better expressed than by the comparative quantitative analysis
depicted in Tables 3 and 4. The degrees of overall
cerebral atrophy and atrophy of the cerebral cortex in
AD were precisely intermediate between those in
PDD and PD+AD. The degree of white matter atrophy in AD and PD+AD was approximately the
same, yet white matter atrophy was not observed in
PDD. In the basal ganglia where there was no atrophy
in AD, virtually identical degrees of atrophy were
present in PDD and PD AD. For the amygdala, the
degree of atrophy in AD was again intermediate between that in PDD and PD AD. Overall, the findings
in this study suggest that dementia in PD+AD is due
to the interaction of neurodegeneration in the cerebral
cortex, amygdala, and hippocampus caused by AD and
neurodegeneration of basal ganglia and the thalamus
due to PD, whereas in PDD, dementia is predominantly subcortical in nature with restricted, regional
involvement of the anterior frontal and posterior
parietal cortex. In PD + AD, as in AD, global cerebral
atrophy probably results from extensive degeneration
of both intracortical and subcortical-cortical circuitry,
whereas in PDD the cortex is relatively spared but may
have functional impairments due to deaerentation
caused by neuronal loss in subcortical nuclei. Thus,
the morphological distinction between PDD and
PD+AD is sharp and should provide an anatomical
basis for comparing and contrasting cognitive deficits
in these two conditions.
Supported by grant no. P50-AGO5 134 from the National Institute
on Aging. Dr de la Monte is the recipient of Physician Scientist
Award AGO0425 from the National Institute on Aging.
1. Sweet RD, McDowell FH, Feigenson JS, et al. Mental symptoms in Parkinson’s disease during chronic treatment with
levodopa Neurology 1976;26:305-3 10
2. Lieberman A, Dziatolowsky M, Kupersmith M, et al. Dementia
in Parkinson’s disease. Ann Neurol 1979;6:355-359
3. Cummings JL, Benson DF. Subcortical dementia. Review of an
emerging concept. Arch Neurol 1984;41:874-879
4. Growdon JH, Corkin S . Cognitive impairments in Parkinson’s
disease. Adv Neurol 1986;45:383-390
5. Korczyn AD, Inzelberg R, Treves T, et al. Dementia of Parkinson’s disease. Adv Neurol 1986;45:399-403
6. Albert ML. Subcortical dementia In: Katzman R, Terry RD,
Bick KL, eds. Alzheimer’s disease: senile dementia and related
disorders. New York: Raven, 1978
7. Ball MJ. The morphological basis of dementia in Parkinson’s
disease. Can J Neurol Sci 1984;11:180-184
8. Stern Y, Mayeux R. Intellectual impairment in Parkinson’s disease. Adv Neurol 1986;45:405-408
9. CorsellisJAN. Aging and the dementias. In: Adams JH, Corsellis JAN, Duchen LW, eds. Greenfield’s neuropathology. 4th ed.
New York: Wiley, 1984951-1025
0. Teravainen H, Hietanen M, Stoessl J, Calne DB. Dementia in
movement disorders. Can J Neurol Sci 1986;13:546-558
11. Terry RD, Katzman R. Senile dementia of the Alzheimer type:
defining a disease. In: Katman R, Terry RD, eds. Neurology of
aging, vol 22. Philadelphia: Davis Co, 1983:51-84
12. Terry RD, Peck A, DeTeresa R, et al. Some morphometric
aspects of the brain in senile dementia of the Alzheimer type.
Ann Neurol 1981;10184-192
13. Mountjoy CQ, Roth M, Evans JHR, Evans HM. Cortical
neurond counts in normal elderly controls and demented patients. Neurobiol &ng 1983;4:1-11
14. Terry RD, Davies P. Dementia of the Alzheimer type. Annu
Rev Neurosci 1980;3:77-95
15. Mayeux R, Stern Y, Rosen J, Benson DF. Is “subcortical dementia” a recognizable clinical entity? Ann Neurol 1983;14:
16. Whitehouse PJ Jr. The concept of subcortical and cortical demencia: another look. Ann Neurol 1986;19:1-6
17. Gaspar P, Gray F. Dementia in idiopathic Parkinson’s disease.
Acta Neuropathol (Berl) 1984;64:43-52
18. Hakim AM, Mathieson G. Dementia in Parkinson’s disease. A
neuropathologic study. Neurology 1979;29:1209-1214
19. Boller F, Mizutani T, Roessmann U, Gambetti P. Parkinson’s
disease, dementia, and Alzheimer’s disease: clinicopathological
correlations. Ann Neurol 1980;7:329-335
20. kverenz J, Sumi SM. Parkinson’s disease in patients with Alzheimer‘s disease. Arch Neurol 198643662-664
21. Diner SM, Mirra SS. Neuropathologic and clinical features of
Parkinson’s disease in Alzheimer’s dsease patients. Neurology
22. Woodard JS. Concentric hyaline inclusion body formation in
mental disease: analysis of 27 cases. J Neuropathol Exp Neurol
23. Kosaka K, Mehrein P. Dementia-parkinsonism syndrome with
numerous k w y bodies and senile plaques in cerebral cortex.
Arch Psychiatr Nervenkr 1979;226:241-2 50
24. Rosenblum WI, Ghatak NR. Lewy bodies in the presence of
Alzheimer’s disease. Arch Neurol 1979;36:170-171
25. Rothschild D, Kasani J. Clinicopathologic study of Alzheimer’s
disease. Arch Neurol Psychiatry 1936;35:293-32 1
26. Pearce J. The extrapyramidal disorder of Alzheimer’s disease.
Eur Neurol 1974;12:94-103
27. Sulkava R Alzheimer‘s disease and senile dementia of the Alzheimer type: a comparative study. Acta Neurol Sand 1982;
de la Monte et al: Cerebral Atrophy in PDD and PD+AD 319
28. Goldman JE, Yen S-H, Chiu F-C, Peress NS. Lewy M e s of
Parkinson’s disease contain neufofilament antigens. Science
29. Pappolla MA. Lewy bodies of Parkinson’s disease. Immune
electron microscopic demonstration of neurofilament antigens
in constituent filaments. Arch Pathol Lab Med 1986;11011601163
30. Forno LS, Strefling AL, Sternberger LA, et al. Immunocytochemical staining of neurofibrillary tangles of the periphery of
Lewy bodies with a monoclonal antibody to neurofilaments. J
Neuropathol Exp Neurol 1983;42:342 (Abstract)
31. Galloway PG, Grundke-Iqbal I, Iqbal K, Perry G. Lewy bodies
contain epitopes both shared and distinct from Alzheimer
neurofibrillary tangles. J Neuropathol Exp Neurol 1988;
32. Manetto V, Perry G, Tabaton M, et al. Ubiquitin is associated
with abnormal cytoplasmic filaments characteristic of neurodegenerative diseases. Proc Natl Acad Sci USA 1988;85:
33. Anderson BH, Breinburg D, Domes MJ, et al. Monoclonal
antibodies show that neurofibrlllary tangles and neurofilaments
share antigenic dererminants. Nature 1982;298:84-86
34. Gambetti P, Shecket G, Ghetti B, et al. Neurofibrillary changes
in human brain. An imrnunocytochemical study with a neurofilament antiserum. J Neuropathol Exp N e w 1 1983;42:69-79
35. Perry G, Rizzuto N, Autilio-Gambem L, Gambetti P. Paired
helical filaments from Alzheimer‘s disease patients contain cytoskeletal components. Proc Natl Acad Sci USA 1985;82:
36. Sternberger NH, Sternberger LA, Ulrich J. Aberrant neurofilament phosphorylation in Alzheimer’s disease. Proc Natl
Acad Sci USA 1985;82:4274-4276
37. Cork LC,Sternberger NH, Sternberger LA, et al. Phosphorylated neurofilament antigens in neurofibrillary tangles in Alzheimer‘s disease. J Neuropathol Exp Neurol 1986;45:56-64
38. Nakano I, Hirano A. Parkinson’s disease: neuron loss in the
nucleus basalis without concomitdt Alzheimer‘s disease. Ann
N e w 1 1984;15:415-418
39. Rogers JD, Brogan D, Mirra SS.The nucleus basalis of Meynert
in neurological disease: a quantitative morphological study. Ann
Neurol 1985;17:163-170
40. Taghavini F, Pilleri G, Bouras C, Constantinidis J. The basal
nucleus of Meynen in idiopathic Parkinson’s disease. Acta
Neurol Scand 1984;70:20-28
41. Whitehouse PJ, Price DL, Clark AW, et al. Alzheimer‘s &ease:
evidence~forselective loss of cholinergic neurons in the nudeus
basalis. Ann Neurol 1981;10:122-126
42. Whitehouse PJ, Price DL, Struble RG, et al. Alzheimer‘s disease and senile dementia: loss of neurons in the basal forebrain.
Science 1982;215:1237-1239
43. Candy JM, Perry RH, Perry EK, et al. Pathological changes in
the nucleus of Meynert in Alzheimer’s and Parkinson’s disease.
J Neurol Sci 1983;59:277-289
44. Wilcock GK, Esiri MM, Bowen DM, Smith CCT. The nucleus
basaLs in Alzheimer‘s disease. Cell counts and cortical biochemistry. Neuropathol Appl Neurobiol 1983;9:175-179
45. Davies P, Maloney AJF. Selective loss of central cholinergic
neurons in Alzheimer‘s disease. Lancet 1976;2:1403
46. Perry EK, Perry RH, Blessed G, Tomlinson BE. Necropsy evidence of central cholinergic deficits in senile dementia. Lancet
47. Bird TD, Stranahan S, Sumi SM, Raskind M. Alzheimer‘s disease: choline acetyltransferase activity in brain tissue from clinical and pathological subgroups. Ann Neurol 1983;14:284293
48. Dubois B, Rubert M, Javoy-Agid F, et al. A subcortical cholinergic system is affected in Parkinson’s disease. Brain Res
1983;288:213-2 18
49. Adolfsson R, Gotdries CG, Roos BE, Winblad B. Changes in
brain catecholamines in patients with dementia of Alzheimer
type. Br J Psychiatry 1979;135:216-223
50. Scatton V, Javoy-Agid F, Roquier L, et al. Reduction of cortical
dopamine, noradrenaline, serotonin and their metabolites in
Parkinson’s disease. Brain Res 1983;275:321-328
51. de la Monte SM, VonsattelJP, Richardson EP Jr. Morphometric
demonstration of atrophic changes in the cerebral cortex, white
matter, and neostriatum in Huntington’s disease.J Neuropathol
Exp Neurol 1988;47:516-525
52. de la Monte SM. Quantitation of cerebral atrophy in preclinical
and end-stage Alzheimer‘s disease. Ann Neurol 1989;25:450459
53. Wells SE, de la Monte SM, Hedley-Whyte ET. Morphometric
analysis of the cerebral hemispheres in Down’s syndrome. J
Neuropathol Exj~Neurol 1988;47:380 (Abstract)
54. de la Monte SM. Disproportionate atrophy of cerebral white
matter in chronic alcoholics. Arch Neurol 1988;45:990-%2
55. Khatchaturian 2s. Diagnosis of Alzheimer‘s disease. Arch
Neurol 1985;42:1097-1104
56. Kamo H, Haebara H, Akiguchi I, et al. A distinctive distribution of reactive astrogha in the precentral cortex in amyotrophic
lateral sclerosis. Acta Neuropathol (Berl) 1987;74:33-38
57. Brun A, Englund EL A white matter disorder in dementia of
the Alzheimer type: a pathoanatomical study. Ann Neurol
58. Kowall NW, Kosik KS. Axonal disruption and aberrant localization of tau protein characterize the neuropil pathology of Alzheimer‘s disease. Ann Neurol 1987;22:639-643
59. Jellinger K. Neuropathological substrates of Alzheimer’s disease
and Parkinson’s disease. J Neural Transm 1987;24(suppl):
109- 129
320 Annals of Neurology Vol 26 No 3 September 1989
Без категории
Размер файла
1 204 Кб
plus, neuropathological, dementia, distinction, disease, parkinson, alzheimers
Пожаловаться на содержимое документа