Plaque and tangle imaging and cognition in normal aging and Alzheimer's disease
Transcript of Plaque and tangle imaging and cognition in normal aging and Alzheimer's disease
N
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
A16
httncwcli
17
18
19
20
21
22
23
24
25
26
©27
K28
29
130
31
t32
m33
NYT
0d
TED
PR
OO
F
ARTICLE IN PRESSBA 7198 1–10
Neurobiology of Aging xxx (2008) xxx–xxx
Plaque and tangle imaging and cognition in normal agingand Alzheimer’s disease
Meredith N. Braskie a,b, Andrea D. Klunder c, Kiralee M. Hayashi c, Hillary Protas d,h,Vladimir Kepe d, Karen J. Miller e, S.-C. Huang d, Jorge R. Barrio d, Linda M. Ercoli e,
Prabha Siddarth e, Nagichettiar Satyamurthy d, Jie Liu d, Arthur W. Toga c,Susan Y. Bookheimer b,e,g, Gary W. Small e,f, Paul M. Thompson c,∗
a Helen Wills Neuroscience Institute, University of California Berkeley, CA 94720, USAb Ahmanson-Lovelace Brain Mapping Center, David Geffen School of Medicine at UCLA, CA 90095, USA
c Laboratory of Neuro Imaging, Department of Neurology, David Geffen School of Medicine at UCLA, CA 90095, USAd Department of Molecular and Medical Pharmacology, David Geffen School of Medicine at UCLA, CA 90095, USAe Department of Psychiatry and Biobehavioral Sciences, David Geffen School of Medicine at UCLA, CA 90095, USA
f Center on Aging, David Geffen School of Medicine at UCLA, CA 90095, USAg Department of Psychology, David Geffen School of Medicine at UCLA, CA 90095, USA
h Department of Biomathematics, David Geffen School of Medicine at UCLA, CA 90095, USA
Received 26 March 2008; received in revised form 9 July 2008; accepted 22 September 2008
bstract
Amyloid plaques and tau neurofibrillary tangles, the pathological hallmarks of Alzheimer’s disease (AD), begin accumulating in theealthy human brain decades before clinical dementia symptoms can be detected. There is great interest in how this pathology spreads inhe living brain and its association with cognitive deterioration. Using MRI-derived cortical surface models and four-dimensional animationechniques, we related cognitive ability to positron emission tomography (PET) signal from 2-(1-{6-[(2-[F-18]fluoroethyl)(methyl)amino]-2-aphthyl}ethylidene)malononitrile ([18F]FDDNP), a molecular imaging probe for plaques and tangles. We examined this relationship at eachortical surface point in 23 older adults (10 cognitively intact, 6 with amnestic mild cognitive impairment, 7 with AD). [18F]FDDNP-PET signalas highly correlated with cognitive performance, even in cognitively intact subjects. Animations of [18F]FDDNP signal growth with decreased
ECognition across all subjects (http://www.loni.ucla.edu/∼thompson/FDDNP/video.html) mirrored the classic Braak and Braak trajectory in
ateral temporal, parietal, and frontal cortices. Regions in which cognitive performance was significantly correlated with [18F]FDDNP signalnclude those that deteriorate earliest in AD, suggesting the potential utility of [18F]FDDNP for early diagnosis.
2008 Elsevier Inc. All rights reserved.
t 34
a 35
RReywords: Amyloid; Cerebral cortex; Cognitive aging; Memory; PET
. Introduction
UN
CO
Please cite this article in press as: Braskie, M.N., et al., Plaque and tangleNeurobiol Aging (2008), doi:10.1016/j.neurobiolaging.2008.09.012
Decades may elapse between initial cortical accumula-ions of tau neurofibrillary tangles and amyloid plaques (the
ajor pathologic hallmarks of Alzheimer’s disease (AD)) and
∗ Corresponding author at: Laboratory of Neuro Imaging, Department ofeurology, David Geffen School of Medicine at UCLA, 635 Charles E.oung Drive South, Suite 225E, Los Angeles, CA 90095-7332, USA.el.: +1 310 206 2101; fax: +1 310 206 5518.
E-mail address: [email protected] (P.M. Thompson).
t 36
l 37
H 38
t 39
o 40
2sso
197-4580/$ – see front matter © 2008 Elsevier Inc. All rights reserved.oi:10.1016/j.neurobiolaging.2008.09.012
he cognitive changes required for clinical diagnosis (Braaknd Braak, 1991; Thal et al., 2002). Historically, plaques andangles were only detectable post mortem, making the patho-ogical burden difficult to relate to cognitive performance.owever, PET probe advances may facilitate investigation of
his pathology in living humans. In vivo cortical accumulationf the ligand 2-(1-{6-[(2-[F-18]fluoroethyl)(methyl)amino]-
imaging and cognition in normal aging and Alzheimer’s disease.
-naphthyl}ethylidene)malononitrile ([18F]FDDNP) is con- 41
istent with depositions of both plaques and tangles in living 42
ubjects (Agdeppa et al., 2001b; Small et al., 2006), and previ- 43
usly has been verified via subsequent autopsy to co-localize 44
D
INNBA 7198 1–10
2 iology o
w45
a46
h47
m48
i49
c50
F51
s52
u53
a54
b55
256
a57
r58
d59
a60
(61
62
t63
c64
c65
c66
c67
t68
u69
m70
t71
c72
m73
m74
a75
o76
277
278
79
a80
e81
h82
p83
e84
f85
s86
o87
T88
d89
i90
i91
m92
w93
n94
P95
c96
s 97
R 98
l 99
e 100
i 101
( 102
b 103
104
a 105
e 106
q 107
w 108
109
o 110
I 111
L 112
2 113
114
c 115
u 116
W 117
o 118
l 119
r 120
e 121
t 122
a 123
124
t 125
o 126
i 127
A 128
w 129
p 130
o 131
132
f 133
s 134
p 135
w 136
f 137
i 138
t 139
e 140
M 141
c 142
u 143
o 144
A 145
t 146
NC
OR
RE
CTE
ARTICLEM.N. Braskie et al. / Neurob
ith plaques and tangles (Agdeppa et al., 2001b). Digitalutoradiography of AD brain specimens using [18F]FDDNPas likewise demonstrated that the pattern of ligand bindingatches the pattern of plaques and tangles in neighbor-
ng slices (as determined using immunohistochemistry andonfocal fluorescence microscopy) (Agdeppa et al., 2001a).urthermore, in a recent study, global [18F]FDDNP washown to be more accurate than FDG-PET or MRI brain vol-mes at discriminating among clinical diagnoses (Small etl., 2006). In addition to [18F]FDDNP, PET ligands Pitts-urgh Compound B (PIB) (Buckner et al., 2005; Klunk et al.,004; Mintun et al., 2006) and stilbene (SB-13) (Verhoeff etl., 2004) are both reported to visualize plaques alone. Initialeports have focused on validating the imaging probes andifferentiating diagnostic groups by examining signal aver-ged over regions of interest rather than at each cortical pointKlunk et al., 2004; Small et al., 2006; Verhoeff et al., 2004).
In this study, we compared the voxel-wise spatial rela-ionship between 3D cortical [18F]FDDNP distribution andognitive ability across subjects with diagnoses ranging fromognitively intact to mild AD. This approach empoweredortical signal detection without requiring a priori specifi-ation of regions of interest (ROIs), thus preventing biaseshat could be introduced by variations in ROI sizes. We eval-ated cognition using composite test scores as a continuouseasure, which ensured that participants close to diagnos-
ic boundaries did not obscure results as they might withategorical comparisons. Finally, we mapped cortical grayatter thickness (Thompson et al., 2004) both to separateolecular pathology from the effects of structural atrophy
nd to assess the specificity and independent predictive valuef PET versus MRI measures.
. Methods
.1. Subjects
Twenty-three community-dwelling subjects weressessed with standard neurological and psychologicalxams. Those with a history of stroke, mental illness, seriousead injury, and non-AD diseases that could affect cognitiveerformance were excluded. Although subjects were notxcluded for presence of white matter lesions, which arerequently present in AD, our composite cognitive testcore was not significantly correlated with the occurrencef at least one visible white matter hyperintensity in a2-weighted MRI image (p = 0.868). Seven subjects metiagnostic criteria for AD, 6 for amnestic mild cognitivempairment (MCI) (Petersen, 2004), and 10 were cognitivelyntact controls, although some had typical age-related
emory complaints. Individuals with AD or amnestic MCI
UPlease cite this article in press as: Braskie, M.N., et al., Plaque and tangleNeurobiol Aging (2008), doi:10.1016/j.neurobiolaging.2008.09.012
ere diagnosed using previously published standard diag-ostic criteria (American Psychological Association, 2000;etersen, 2004). Control subjects did not meet diagnosticriteria for MCI or AD. Diagnostic groups did not differ
ccao
PR
OO
F
PRESSf Aging xxx (2008) xxx–xxx
ignificantly in age, education, sex, or Hamilton Depressionating Scale (21 item) (Hamilton, 1960). To maximize the
ikelihood of finding cognitively intact subjects who havearly AD-like pathology, we included several cognitivelyntact subjects who had a family history of dementia6), were apolipoprotein E e4 (APOE4) carriers (4), oroth (2).
All but seven of the subjects included in the current studylso were reported in a study previously published (Smallt al., 2006). Subjects from that study who lacked high-uality T1-weighted MPRAGE MRI scans (Siemens 3T)ere excluded from our study.We obtained informed written consent from all subjects
r their medical proxies, and the study was approved by thenstitutional Review Board of the University of California,os Angeles (UCLA).
.2. Neuropsychological testing
Participants underwent a full battery of neuropsychologi-al tests whose scores were converted to age-adjusted Z scoressing established age-based normative data for each test.e created a composite average cognitive Z score composed
f three tests of episodic memory and three tests of frontalobe function (Table 1). Selected tasks had a wide range ofesponses and high sensitivity to AD-related changes (Jonest al., 2006; Tabert et al., 2006) and assessed a range of men-al function including memory, processing speed, attentionalnd response control, and phonemic fluency.
Among the AD patients, cognitive impairment preventedwo subjects from completing the Stroop C Interference task,ne from completing the Buschke-Fuld Selective Remind-ng task, and one from completing the WMS Verbal Pairedssociates Immediate task. In these cases, we substituted theorst Z score on that task by any study subject for the incom-lete test score. No subject failed to complete more than onef the tasks.
Age-adjusted Z scores on the well-established tests per-ormed by subjects were correlated with the compositecores they composed (Pearson’s r values range = 0.67–0.90;≤ 0.0005 for all tests), suggesting that the composite scoreas a valid indicator of memory ability and frontal lobe
unction in these participants. Averaging Z scores from thendividual tests into a composite score reduced the likelihoodhat an aberrant score on any one task would unduly influ-nce the results of the study. The composite score and theMSE score were also correlated (r = 0.73, p < 0.0001). The
omposite score, however, offered a greater range of val-es than the MMSE, allowing us to distinguish gradationsf cognitive ability even in cognitively intact older adults.
covariance matrix of the age-adjusted Z scores for eachest across all subjects showed that all the neuropsychologi-
imaging and cognition in normal aging and Alzheimer’s disease.
al scores included in the composite score were significantly 147
orrelated with one another (r = 0.42–0.83), suggesting that 148
veraging these scores provided a reasonable single measure 149
f global cognition.
ED
F
ARTICLE IN PRESSNBA 7198 1–10
M.N. Braskie et al. / Neurobiology of Aging xxx (2008) xxx–xxx 3
Table 1Demographics and clinical characteristics.
AD MCI Controls
No. of participantsa 7 6 10Average ageb 77 ± 10.4 73 ± 12.8 73 ± 10.4Males/females 3/4 2/4 3/7Years of educationb 16 ± 2 17 ± 3 17 ± 3Apolipoprotein E �4 carriersc 4 (57%) 2 (33%) 4 (40%)Known family history of dementiad 2 (29%) 5 (83%) 6 (60%)Mini-Mental State Exam scoreb 23 ± 2e 28 ± 1e 29 ± 1Hamilton Depression Rating Scale, 21 itemb 3 ± 3 4 ± 3 2 ± 2Cognitive composite Z scoreb,f −2.2 ± 0.9e −1.0 ± 0.4e 0.7 ± 0.8e
a Two participants with AD and one with MCI were excluded from the cortical thickness comparison because of technical difficulties with their MRI scans.b Listed as mean ± standard deviation.c APOE genotype was not available for two subjects with MCI.d Family history defined as having a parent, grandparent or sibling with dementia.e Significant difference between starred items.
Fuld SeW e functioA
2150
151
r152
(153
e154
0155
2156
2157
158
d159
O160
S161
a162
m163
t164
t165
a166
t167
i168
a169
s170
t171
f172
173
a174
v175
h176
g177
m178
a179
180
a181
t182
t183
T 184
w 185
186
2 187
e 188
o 189
O 190
g 191
m 192
193
i 194
t 195
a 196
a 197
a 198
r 199
( 200
2 201
202
[ 203
( 204
v 205
f 206
K 207
s 208
( 209
a 210
5 211
a 212
2 213
NC
OR
RE
CT
f Mean age-adjusted Z scores for three tests of verbal memory (Buschke-MS verbal paired associates immediate total) and three tests of frontal lobssociation Test: FAS).
.3. MRI protocol
We obtained sagittal T1-weighted magnetization preparedapid acquisition gradient-echo (MPRAGE) volumetric scans3T Siemens Allegra MRI): repetition time (TR) 2300 ms;cho time (TE) 2.93 ms; 160 slices; slice thickness 1 mm/skip.5 mm; in-plane voxel size 1.3 mm × 1.3 mm; field of view56 × 256; flip angle 8◦.
.4. MRI image processing
MRI scans were processed using a sequence of stepsescribed previously (Thompson et al., 2004). We used thexford Centre for Functional MRI of the Brain (FMRIB)oftware Library (FSL) Brain Extraction Tool (Smith, 2002)nd “FSL view” to create and manually refine individual brainasks, which were then applied to exclude non-brain mat-
er. The raw data were scaled and spatially normalized tohe International Consortium for Brain Mapping ICBM53verage brain imaging template with a nine-parameter linearransformation (Collins et al., 1994). Magnetic susceptibil-ty artifacts and image non-uniformities were reduced usingregularized tricubic B-spline approach. We automatically
egmented each resulting image into gray matter, white mat-er, and cerebrospinal fluid using a Gaussian mixture modelor their MRI signal values (Shattuck et al., 2001).
A three-dimensional cortical surface model was extractedutomatically from each subject’s scan as described pre-iously (MacDonald et al., 2000). Three-dimensionalemispheric reconstructions were created, onto which a sin-le trained researcher, blind to [18F]FDDNP-PET results,anually traced neuroanatomical landmarks. Inter-rater reli-
bility has been reported previously (Sowell et al., 2000).
UPlease cite this article in press as: Braskie, M.N., et al., Plaque and tangleNeurobiol Aging (2008), doi:10.1016/j.neurobiolaging.2008.09.012
We created an inter-subject 3D average cortical model,s detailed previously (Thompson et al., 2003), by flatteninghe cortex and gyral landmarks into two-dimensional space,hen warping all landmarks into alignment across subjects.
f(
PR
OOlective Reminding total recall, WMS logical memory II delayed total, and
n (Stroop C interference, WAIS-R digit symbol, and Controlled Oral Word
his method allows voxel by voxel averaging of these imagesithin each delineated region across subjects.After extracting gray matter volumes (Shattuck et al.,
001) and spatially registering them to the hemispheric mod-ls, we calculated cortical gray matter thickness at each pointf the brain surface (Sapiro, 2001; Thompson et al., 2004).ur approach defines thickness as the distance from the innerray–white boundary to the closest point on the outer grayatter surface.After supersampling the image data to create 0.33 mm
sotropic voxels, a 3D Eikonal equation was applied onlyo gray matter voxels, and a smoothing kernel was used toverage gray matter thickness values within a 15-mm spheret each cortical surface point (Hayashi et al., 2002; Sowell etl., 2004). Test–retest reliability of cortical thickness usingepeated scanning and analysis has been reported previouslySowell et al., 2004).
.5. [18F]FDDNP-PET scanning
Each subject received a bolus injection of 320–550 MBq18F]FDDNP, prepared at very high specific activities>37 GBq/mmol) (Liu et al., 2007), through an in-dwellingenous catheter. Consecutive dynamic PET scans were per-ormed for 2 h on an EXACT HR+ tomograph (Siemens-CTI,noxville, TN) while participants lay supine. Sixty-three
lices were collected parallel to the orbito-meatal line2.24 mm plane separation). The scans were decay correctednd reconstructed using filtered back-projection (Hann filter,.5 mm FWHM) with correction for scatter and measuredttenuation.
.6. [18F]FDDNP-PET image analysis
imaging and cognition in normal aging and Alzheimer’s disease.
We performed Logan graphical analysis (using PET 214
rames 30–125 min) to create distribution volume ratio 215
DVR) parametric images of relative [18F]FDDNP-PET 216
D P
RO
OF
IN PRESSNBA 7198 1–10
4 iology of Aging xxx (2008) xxx–xxx
b217
i218
l219
d220
d221
e222
223
a224
d225
v226
a227
s228
c229
230
v231
m232
a233
r234
2235
2236
237
b238
p239
t240
p241
q242
a243
a244
i245
o246
w247
c248
c249
2250
c251
252
i253
d254
o255
t256
a257
t258
a259
e260
2261
262
t263
[264
b265
f
Fig. 1. Increasing [18F]FDDNP and declining cognition. Cortical maps showthat as [18F]FDDNP signal (DVR) increased at each cortical surface point(across all subjects), cognitive performance decreased (see Section 2 forcalculation of composite test scores). Regions of significant correlation(p < 0.05), color-coded in red, follow the profile of tangle and plaque accu- Q2
mulation characteristic of mild AD, as established in large-scale post mortemmapping studies (Braak and Braak, 1991). Corresponding maps did not showregions where both [18F]FDDNP signal and cognition increased together (notstv
3 266
3 267
p 268
269
w 270
c 271
w 272
r 273
p 274
f 275
m 276
o 277
a 278
w 279
c 280
a 281
NC
OR
RE
CTE
ARTICLEM.N. Braskie et al. / Neurob
inding (manifested as PET signal). In these parametricmages, the value at each point represented the slope of theinear portion of the Logan plot (or DVR), calculated as theistribution volume of the tracer at that point divided by theistribution volume in the cerebellar reference region (Logant al., 1996), where little [18F]FDDNP binding was expected.
PET images were co-registered to the MRI images usingmutual information-based rigid body transformation. As
escribed previously (Protas et al., 2005), we assigned PETalues to each cortical surface vertex by computing the aver-ge value of the DVR image in a kernel of radius 7 mmurrounding each cortical mesh point, while excluding extra-ortical voxels.
For each subject, we also obtained average [18F]FDDNPalues within ROIs that included bilateral frontal, parietal,edial temporal, and lateral temporal brain regions as well
s posterior cingulate gyrus. These were traced on the co-egistered MRI scans as described previously (Small et al.,006).
.7. Statistical analysis
Statistical maps were generated indicating the correlationetween [18F]FDDNP-PET signal at each cortical surfaceoint and each subject’s composite test score. Different sta-istical models were fitted at each surface vertex, as detailedreviously (Thompson et al., 2004), including linear anduadratic terms in the model, and retaining only terms withsignificant fit. The resulting significance values associ-
ting [18F]FDDNP signal and cognitive performance werendicated by a color code plotted at each surface pointn the average cortex. After obtaining average ROI valuese additionally performed non-parametric Spearman’s rank
orrelations between regional [18F]FDDNP values and theomposite cognitive scores.
.8. Permutation testing and multiple comparisonsorrection
The significance map was corrected for multiple compar-sons by permutation testing using a threshold of p < 0.05 toefine a suprathreshold region. The area of this suprathresh-ld region was compared with a null distribution of statisticshat occurred by chance when test scores were randomlyssigned to subjects in 100,000 random simulations. Permu-ation testing has been used widely in imaging and providesglobal p value for the observed pattern of effects (Bullmoret al., 1999; Thompson et al., 2003).
.9. Hypotheses
We used one-sided hypothesis testing, predicting a priori
UPlease cite this article in press as: Braskie, M.N., et al., Plaque and tangleNeurobiol Aging (2008), doi:10.1016/j.neurobiolaging.2008.09.012
hat subjects with lower composite scores would show greater18F]FDDNP signal, particularly in those cortical regions thatoth degenerate early in Alzheimer’s disease and are criticalor cognitive performance.
c
dH
hown here; maps are blue with p > 0.05 at all voxels). (For interpretation ofhe references to color in this figure legend, the reader is referred to the webersion of the article.)
. Results
.1. [18F]FDDNP-PET signal and cognitiveerformance
[18F]FDDNP-PET signal was significantly higher acrossidespread cortical regions in subjects with poorer neuropsy-
hological test performance (Fig. 1). Strong correlationsere seen in the entorhinal, orbitofrontal, and lateral tempo-
al cortices, temporoparietal and perisylvian language areas,arietal association cortices, and much of the dorsolateral pre-rontal cortex. Correlated regions were bilateral, except in theedial wall, where above threshold correlations were visible
nly in the left hemisphere in restricted frontal pole regions,lthough the difference in signal between hemispheresas not significant. As expected, primary sensorimotor
ortices (e.g., central and pre-central gyri) did not shown association between [18F]FDDNP signal variation and
imaging and cognition in normal aging and Alzheimer’s disease.
ognition. 282
Notably, cingulate and paralimbic belts did not show 283
etectable correlations on the medial hemispheric surface. 284
owever, significant correlations were found in the medial 285
TED
PR
OO
F
ARTICLE IN PRESSNBA 7198 1–10
M.N. Braskie et al. / Neurobiology of Aging xxx (2008) xxx–xxx 5
Fig. 2. [18F]FDDNP and cognition in cognitively intact subjects. Frontalregions of the right lateral cortical surface showed elevated [18F]FDDNPsignal in healthy normal subjects with poorer cognitive performance whencorrelation analyses were restricted to cognitively intact subjects (controlsonly). Cortical maps show similar correlations between [18F]FDDNP signaland composite cognitive score as in Fig. 1, but in a more restricted region.Dorsolateral prefrontal regions are implicated in executive function, whichis among the functions required for optimal performance on the neuropsy-chological tests. Additional correlations were found in parietal associationareas; the left lateral hemisphere did not show broad regions with correla-tions (shown on the right panels here). Corroborating these results, mapso 18
s
a286
t287
288
o289
a290
o291
(292
i293
294
[295
t296
a297
t298
r299
p300
o301
a302
w303
(304
Fig. 3. Time-lapse films calibrating maps of [18F]FDDNP signal versus cog-nition. Projected mean [18F]FDDNP signal (DVR) can be calculated forvarious cognition scores based on the relationship of [18F]FDDNP signalwith cognition in the subjects studied. Here we show projected signal forsubjects who scored (A) high normal (2 standard deviations above age-normal), (B) at age-normal levels (Z score = 0), (C) low normal (2 standarddeviations lower than age-normal), and (D) at impaired levels (4 standarddeviations below age-normal). Red colors denote regions in which greaterpredicted [18F]FDDNP signal is associated with lower cognitive Z scoresat each cortical point based on a nonlinear spatially varying model. Theparameterization of disease stages is based on cross-sectional data, but it isplausible that a comparable trajectory would be followed for each cognitives(i
305
w 306
t 307
[ 308
o 309
j 310
v 311
d 312
M 313
v 314
o 315
( 316
ing two standard deviations from the mean lies within the 317
UNCO
RR
ECf regions in which [ F]FDDNP signal and cognition in cognitively intact
ubjects increased together showed no effects, as expected (not shown).
nd lateral temporal cortices, where plaques and tangles arehought to be deposited first (Braak and Braak, 1997).
To determine whether these patterns could have beenbserved by chance, permutation tests were conducted tossign global p values to the maps. At a voxel-level thresholdf p = 0.05, the corrected significance values were p = 0.004left hemisphere) and p = 0.008 (right hemisphere), suggest-ng that the results were unlikely to be due to chance.
When only cognitively intact subjects were considered,18F]FDDNP signal was higher in restricted right frontal cor-ices, and in some parietal association areas in those havinglower composite cognitive score (Fig. 2). At a voxel-level
hreshold of p = 0.05, the corrected significance value in theight hemisphere was p = 0.031 after permutation testing waserformed. In the left hemisphere, correlations were detectednly in isolated anterior prefrontal regions on the medial wall
Please cite this article in press as: Braskie, M.N., et al., Plaque and tangleNeurobiol Aging (2008), doi:10.1016/j.neurobiolaging.2008.09.012
nd in some of the occipital lobe medial surface, but theseere not significant after correction for multiple comparisons
p = 0.129).
9ns
tage in an individual subject, albeit with variable timing across individuals.For interpretation of the references to color in this figure legend, the readers referred to the web version of the article.)
Given that the estimation of a nonlinear model relationas significant both pointwise and overall after permuta-
ion testing, it is possible to predict the cortical profile of18F]FDDNP signal that would be expected at each levelf cognitive deterioration. Based on the 65,536 fitted tra-ectories for [18F]FDDNP signal at the cortical surfaceertices across all subjects, we created an image of pre-icted [18F]FDDNP signal for each composite score value.aps of mean [18F]FDDNP signal were generated for indi-
iduals at each end of the normal range and in the middlef the normal range (maps corresponding to Z = +2, 0, −2)Fig. 3). These Z scores were selected because a person scor-
imaging and cognition in normal aging and Alzheimer’s disease.
5th percentile for normal performance. Notably, there is a 318
euroanatomical spread in the areas of higher [18F]FDDNP 319
ignal even within the normal range, with greatest signal 320
D P
RO
F
ARTICLE IN PRESSNBA 7198 1–10
6 M.N. Braskie et al. / Neurobiology of Aging xxx (2008) xxx–xxx
F tionshii terior ch
i321
e322
s323
m324
o325
a326
i327
h328
329
w330
t331
t332
r333
334
o335
c336
a337
p338
p339
w340
w341
c342
t343
c344
a345
i346
[347
(348
l349
i350
3351
352
[353
t 354
signal with cortical thickness. A map of the pointwise rela- 355
tionship between cortical thickness and [18F]FDDNP signal 356
(Fig. 5) showed that such correlations were low and non- 357
Fig. 5. PET-MRI correlation. (A) Maps of the pointwise significancebetween [18F]FDDNP signal and cortical thickness are shown in millimeters,in 20 subjects. (B) Maps of the pointwise correlation (Pearson’s r value) areshown for the same relationship. These maps show that there is no detectableresidual correlation between PET and cortical thickness, and therefore thePET effects observed in Figs. 1 and 2 are not confounded by the effects of
NC
OR
RE
CTE
ig. 4. ROI analysis. Scatterplot graphs and linear regressions show the relan (A) frontal lobe, (B) parietal lobe, (C) lateral temporal lobe, and (D) posemispheres was pooled for these analyses.
n the medial temporal, entorhinal, orbitofrontal, and lat-ral temporal cortices. The increase in frontal [18F]FDDNPignal is characteristic of subjects that are outside the nor-al range of cognition (see map corresponding to Z = −4,
r scores 4 standard deviations below the mean). Thedvancement of pathology can be seen in the accompany-ng Supplementary Online Data, observable on the internet atttp://www.loni.ucla.edu/∼thompson/FDDNP/video.html.
In this film, frames corresponding to Z scores +2 to −4ere estimated from the statistical model fitted at each cor-
ical surface vertex, and were concatenated at 30 frames/so create a digital animation of the path of pathology withespect to different levels of cognitive performance.
To supplement our cortical surface map results with thosebtained using an ROI analysis, we performed Spearman rankorrelations between composite cognitive scores and aver-ge [18F]FDDNP signal DVR values in several ROIs: frontal,arietal, medial temporal and lateral temporal regions, and theosterior cingulate gyrus. As hypothesized, when all subjectsere included, average [18F]FDDNP values in each regionere significantly lower in those having higher composite
ognitive scores (p < 0.05). Graphs in Fig. 4 demonstratehese relationships in four of the ROIs. In contrast, when onlyognitively intact subjects were considered using the ROInalysis (rather than the voxel-wise approach), the compos-te cognitive score was significantly correlated with average18F]FDDNP signal only in the posterior cingulate gyrusrho = −0.76; p = 0.01) (not shown). The posterior cingu-ate gyrus is a region thought to show metabolic changesn subjects at risk for AD (Small et al., 2000).
U
Please cite this article in press as: Braskie, M.N., et al., Plaque and tangleNeurobiol Aging (2008), doi:10.1016/j.neurobiolaging.2008.09.012
.2. Correlations with cortical thickness
In order to determine the independent predictive value of18F]FDDNP versus MRI-derived measures, we examined
aiFca
Op between the composite cognitive scores and average [18F]FDDNP signalingulate gyrus across all subjects. Average [18F]FDDNP signal from both
he separate relationships of both cognition and [18F]FDDNP
imaging and cognition in normal aging and Alzheimer’s disease.
trophy. Thickness and PET signal are orthogonal (not correlated) exceptn very small, scattered regions not overlapping with the main effects inigs. 1 and 2. The left frontal pole effects cover no more than 5% of theortex, so they are likely false positives (5% false positives are expected inny null map thresholded at p = 0.05).
ED
PR
OO
F
ARTICLE IN PRESSNBA 7198 1–10
M.N. Braskie et al. / Neurobiology of Aging xxx (2008) xxx–xxx 7
Fig. 6. Maps of post mortem amyloid load and cortical thickness. (A) Images, which were adapted from Fig. 1 of a previously published paper (Braak and Braak,1991) (copyright Springer-Verlag, 1991), with kind permission of the authors, Springer Science, and Business Media, display the classic post mortem amyloiddeposition pattern. Images displayed in (B) also have been previously published (Thompson et al., 2003) (copyright 2003 by the Society for Neuroscience),and show thinner average gray matter in AD patients (mean MMSE = 18) compared with controls at an initial scan (top right) and a follow-up scan 1.5 yearsl ). The pp se regio[ from th
s358
m359
c360
e361
i362
fi363
t364
[365
t366
a367
(368
369
r370
i371
T372
m373
374
n375
r376
4377
378
[379
i 380
n 381
r 382
a 383
g 384
i 385
n 386
m 387
B 388
[ 389
r 390
fi 391
392
w 393
t 394
w 395
p 396
p 397
p 398
v 399
r 400
NC
OR
RE
CT
ater, when the mean MMSE for AD patients dropped to 13 (bottom rightreviously in patients with mild to moderate AD (upper panels) mirrors tho18F]FDDNP signal in Fig. 1 of the current study [adapted with permission
ignificant throughout the cortex. Note that there is no knownethod to ascribe a global significance value to a map of
orrelations between two spatially varying signals, as thexchangeability assumption, required for permutation test-ng, does not apply, and there is no associated Gaussianeld theory yet developed to give such a global p value for
wo correlated random processes. Interestingly, the pattern of18F]FDDNP signal in this study did match the pattern of cor-ical thinning found previously in a population having moredvanced AD than the subjects in the current study (Fig. 6B)Thompson et al., 2003).
Because reduced cortical gray matter thickness has beeneported to be associated with AD, MCI, and poorer cognitionn prior studies (Apostolova et al., 2006; Singh et al., 2006;hompson et al., 2004), we also correlated cortical thicknesseasures with cognitive performance in our sample.As with [18F]FDDNP signal, cognitive performance was
ot associated with cortical thickness in this sample (cor-ected p > 0.05; both brain hemispheres).
U
Please cite this article in press as: Braskie, M.N., et al., Plaque and tangleNeurobiol Aging (2008), doi:10.1016/j.neurobiolaging.2008.09.012
. Discussion
Cognitive performance was significantly correlated with18F]FDDNP signal in right frontal and parietal regions
p 401
ots
attern of cortical thinning and post mortem amyloid deposition displayedns demonstrating a significant correlation between cognitive function and
e authors and publishers].
n cognitively intact subjects, suggesting that some cog-itive aging that is considered age-normal actually mayeflect pathological brain changes, particularly in subjectst risk for AD. That is not to say that plaque and tan-le accumulations are a result of aging or are presentn all older adults, but rather that in some people diag-osed as having normal cognitive aging, plaques and tanglesay be associated with their subtle cognitive decline.oth the lack of significant positive correlations between
18F]FDDNP signal and composite cognitive scores and theigor of permutation testing supported the validity of ourndings.
It is important to note that even though [18F]FDDNP signalas elevated in the medial temporal lobe in some cogni-
ively intact adults, it was primarily in the frontal cortexhere [18F]FDDNP signal distinguished those controls whoerformed better on certain cognitive tasks from those whoerformed worse. Permutation testing without a restricted ariori search region appropriately applies a somewhat conser-ative correction for false positives, so the correlation in theight hemisphere but not the left may reflect limited statisticalower rather than true hemispheric specificity. The specificity
imaging and cognition in normal aging and Alzheimer’s disease.
f the relationship between [18F]FDDNP signal and cogni- 402
ion for right frontal cortex will be tested specifically in future 403
tudies having larger sample sizes. 404
D
INNBA 7198 1–10
8 iology o
405
s406
t407
C408
w409
o410
t411
b412
1413
i414
t415
416
c417
n418
b419
i420
s421
[422
p423
r424
i425
r426
t427
m428
s429
v430
n431
e432
s433
n434
b435
w436
t437
a438
439
v440
f441
r442
e443
t444
e445
N446
d447
b448
H449
n450
t451
a452
s453
t454
d455
c456
c457
g458
f459
a460
b 461
2 462
p 463
t 464
t 465
466
c 467
i 468
b 469
i 470
t 471
i 472
s 473
n 474
m 475
l 476
c 477
p 478
j 479
e 480
c 481
n 482
m 483
T 484
b 485
n 486
f 487
t 488
d 489
b 490
M 491
f 492
d 493
494
i 495
t 496
a 497
l 498
t 499
t 500
n 501
d 502
c 503
o 504
a 505
v 506
e 507
s 508
n 509
t 510
a 511
n 512
NC
OR
RE
CTE
ARTICLEM.N. Braskie et al. / Neurob
The correlations we observed in the cognitively intactubjects alone were maintained and expanded when cogni-ively impaired subjects were also included in the sample.ognitive performance across all subjects was correlatedith [18F]FDDNP signal in inferior and lateral temporal,rbitofrontal, dorsolateral prefrontal, and parietal associa-ion cortices—regions with the greatest plaque and tangleurden in histopathological studies of AD (Braak and Braak,991). The anatomical agreement is striking between thesen vivo maps and the well-established post mortem maps forhe staging of AD (Braak and Braak, 1991).
ROI analyses yielded significant relationships between theomposite cognitive scores and the average [18F]FDDNP sig-al in all regions examined when all subjects were included,ut only in the posterior cingulate gyrus when cognitivelyntact subjects alone were considered. Given the reducedample size and the restricted range of variability in the18F]FDDNP signal within the control group, it is not sur-rising that the controls alone did not demonstrate significantelationships between [18F]FDDNP signal and the compos-te cognitive score in several of the ROIs. When there is aestricted range of variability in the [18F]FDDNP signal (ashere is within the control group), a voxel-wise approach
ay provide advantages over an ROI approach (in whichignal is averaged across significant and non-significantoxels) for detecting relationships between [18F]FDDNP sig-al and cognition. It is interesting to note, however, thatven using these averaged ROIs to determine [18F]FDDNPignal, the graphs in Fig. 4 demonstrate that there doot appear to be outlying data points in the relationshipsetween [18F]FDDNP signal and composite cognitive scoresithin the controls, lending support to the significant rela-
ionships we found using a voxel-wise statistical mappingpproach.
The relationship of plaques and tangles to AD is contro-ersial. Both must accompany specific cognitive impairmentor a definitive AD diagnosis (McKhann et al., 1984). Someesearchers believe that plaques or tangles cause the dis-ase (Binder et al., 2005; Selkoe, 2001); others contend thathese merely tend to co-occur with other more causative dis-ase processes (Castellani et al., 2006; Watson et al., 2005).eurofibrillary tangle density correlates more strongly withisease severity and neuronal death than does total plaqueurden (Berg et al., 1998; Giannakopoulos et al., 2003).owever, even in studies in which total plaque load didot correlate with disease severity, a simple comparison ofotal plaque load in AD subjects versus controls (rather than
correlation with severity of dementia) showed that ADubjects had on average more extensive plaques than con-rols (Bouras et al., 1994; Gomez-Isla et al., 1996). Theseata suggest that both plaques and tangles are good indi-ators of disease processes, regardless of whether they are
UPlease cite this article in press as: Braskie, M.N., et al., Plaque and tangleNeurobiol Aging (2008), doi:10.1016/j.neurobiolaging.2008.09.012
ausative factors in AD. Although we are unable to distin-uish [18F]FDDNP signal associated with amyloid plaquesrom that associated with tau neurofibrillary tangles in vivo,
recent study compared [18F]FDDNP-PET scanning and
t 513
d
f
PR
OO
F
PRESSf Aging xxx (2008) xxx–xxx
rain autopsy assessment in the same patient (Small et al.,006). In that study, [18F]FDDNP signal in the medial tem-oral lobe was mainly associated with tau pathology whereashat in other areas of the brain was overwhelmingly relatedo amyloid plaque deposition.
We did not find a significant correlation between corti-al thickness and either [18F]FDDNP signal or cognitionn the current study. However, the significant correlationsetween [18F]FDDNP signal distribution and cognitive abil-ty in the current study were evident in the same regionshat showed cortical thinning related to more advanced ADn our prior studies (Fig. 5B) (Thompson et al., 2003). Asuch, the pattern of cortical [18F]FDDNP signal in this cog-itively intact and mildly affected population very closelyatches the topography of cortical thinning known to appear
ater, albeit with a substantial time-lag. Previous studies thatonsidered cognition and cortical thickness either focusedrimarily on cognitively impaired subjects or had larger sub-ect samples than we had in the current study (Apostolovat al., 2006; Lerch et al., 2005; Thompson et al., 2004). Inontrast, nearly half of the subjects in the current study wereormal controls who would not be expected to show anythingore than the most subtle pathology-related cortical thinning.his early distribution of plaques and tangles may thereforee followed by MRI-detectable cortical thinning only wheneuronal damage has become more extensive than is typicallyound in cognitively intact older adults. Our results suggesthat plaque and tangle deposition occurs early and precedesetectable changes in cortical structure, so [18F]FDDNP maye more sensitive to early cognitive changes than structuralRI measures are, and therefore may offer greater power
or disease detection, at least during the early stages of theisease process.
Without partial volume correction, PET measures arenfluenced by cortical atrophy, which reduces the gray mat-er volume emitting radioisotope signals, resulting in signalttenuation. However, partial volume correction is arguablyess critical for interpretation of [18F]FDDNP-PET scanshan for metabolic or perfusion PET images, as the diseaseends to elevate [18F]FDDNP signal and reduce cortical thick-ess. Therefore, any atrophic effect works against finding aisease-associated PET signal increase, and PET increasesannot reasonably be attributed to cortical thinning; usef uncorrected values is, therefore, a slightly conservativepproach. It also avoids the risk of overcorrecting the signalalues, which could occur if the partial volume model was notxactly correct, and ensures that any observed [18F]FDDNPignal increase can be interpreted as related to the ligand andot to structural atrophy. Future empirical estimation of par-ial volume models for different gray/white matter fractionsnd local cortical geometries may increase the signal-to-oise ratio for detecting correlations with cognition withhis ligand, so the current approach should be considered as
imaging and cognition in normal aging and Alzheimer’s disease.
eliberately conservative. 514
Eighteen of our 23 subjects were APOE4+, had a known 515
amily history of dementia, or both. Because our subjects 516
ED
INNBA 7198 1–10
iology o
w517
h518
a519
a520
f521
i522
s523
l524
s525
C526
527
P528
A529
l530
B531
i532
N533
i534
A535
536
M537
B538
a539
m540
H541
(542
M543
J544
t545
C546
R547
R548
b549
O550
J551
T552
R553
v554
f555
o556
f557
A558
M559
560
D561
r562
G563
t564
H565
F
R 566
A 567
568
569
570
571
A 572
Q1 573
574
575
A 576
577
A 578
579
580
581
B 582
583
584
585
586
587
B 588
589
590
B 591
592
593
594
595
B 596
597
B 598
599
B 600
601
602
603
604
B 605
606
607
608
C 609
610
611
C 612
613
614
G 615
616
617
618
G 619
620
621
622
H 623
624
H 625
626
NC
OR
RE
CT
ARTICLEM.N. Braskie et al. / Neurob
ere at high risk for AD and were highly educated, those whoad lower memory ability on certain tasks than their same-ge peers were more likely than the general population to beffected by a pathological condition. Including participantsrom backgrounds representative of the general populationn future studies may help to further elucidate these relation-hips. Finally, the results presented here are cross-sectional;ongitudinal follow-up is needed to determine which controlubjects will eventually develop AD.
onflicts of interest
The University of California, Los Angeles, owns a U.S.atent (6,274,119) entitled “Methods for Labeling Beta-myloid Plaques and Neurofibrillary Tangles,” which is
icensed to Siemens. Drs. Small, Huang, Satyamurthy, andarrio are among the inventors and each receives royalties
n regard to the application of the FDDNP-PET radioligand.one of the other authors have real or perceived conflicts of
nterest.
cknowledgements
National Institute on Aging, the National Library ofedicine, the National Institute for Biomedical Imaging andioengineering, the National Center for Research Resources,nd the National Institute for Child Health and Develop-ent (AG016570, LM05639, EB01651, RR019771, andD050735) to PMT; National Institutes of Health grants
P01-AG024831, AG13308, P50 AG 16570, MH/AG58156,H52453, AG10123, and M01-RR00865) to GWS and
RB; the Department of Energy (DE-FC03-87-ER60615),he General Clinical Research Centers Program, the RotaryART Fund, the Alzheimer’s Association, the Fran anday Stark Foundation Fund for Alzheimer’s Diseaseesearch, the Ahmanson Foundation, the Larry L. Hill-lom Foundation, the Lovelace Foundation, the Judithlenick Elgart Fund for Research on Brain Aging, the
ohn D. French Foundation for Alzheimer’s Research, theamkin Foundation, and the National Center for Researchesources grants (RR13642 and RR021813) to AWT; Indi-idual National Research Service Award (F31 NS45425)rom the National Institutes of Health/National Institutef Neurological Disorders and Stroke and a scholarshiprom ARCS Foundation, Inc./The John Douglas Frenchlzheimer Foundation (with the Erteszek Foundation) toNB.We are indebted to Ms. Andrea Kaplan, Ms. Deborah
orsey, and Ms. Teresann Crowe-Lear for their help inecruiting volunteers and coordinating the study, and to Ms.
UPlease cite this article in press as: Braskie, M.N., et al., Plaque and tangleNeurobiol Aging (2008), doi:10.1016/j.neurobiolaging.2008.09.012
wendolyn Byrd for her help with scheduling of volun-eers and collection of their personal data. We also thank Dr.eiko Braak for allowing us to include his images as part ofig. 6.
627
J
PR
OO
F
PRESSf Aging xxx (2008) xxx–xxx 9
eferences
gdeppa, E.D., Kepe, V., Liu, J., Flores-Torres, S., Satyamurthy, N., Petric,A., Cole, G.M., Small, G.W., Huang, S.C., Barrio, J.R., 2001a. Bindingcharacteristics of radiofluorinated 6-dialkylamino-2-naphthylethylidenederivatives as positron emission tomography imaging probes for beta-amyloid plaques in Alzheimer’s disease. J. Neurosci. 21, RC189.
gdeppa, E.D., Kepe, V., Shoghi-Jadid, K., Satyamurthy, N., Small, G.W.,Petric, A., Vinters, H.V., Huang, S.-C., Barrio, J.R., 2001b. In vivo andin vitro labeling of plaques and tangles in the brain of an Alzheimer’sdisease patient: a case study. J. Nuclear Med. (Suppl 42), 65P.
merican Psychological Association, 2000. Diagnosis and Statistical Man-ual of Mental Disorders DSM-IV-TR (Text Revision). Washington, DC.
postolova, L.G., Lu, P.H., Rogers, S., Dutton, R.A., Hayashi, K.M.,Toga, A.W., Cummings, J.L., Thompson, P.M., 2006. 3D mapping ofmini-mental state examination performance in clinical and preclinicalAlzheimer disease. Alzheimer Dis. Assoc. Disord. 20, 224–231.
erg, L., McKeel Jr., D.W., Miller, J.P., Storandt, M., Rubin, E.H., Morris,J.C., Baty, J., Coats, M., Norton, J., Goate, A.M., Price, J.L., Gearing,M., Mirra, S.S., Saunders, A.M., 1998. Clinicopathologic studies in cog-nitively healthy aging and Alzheimer’s disease: relation of histologicmarkers to dementia severity, age, sex, and apolipoprotein E genotype.Arch. Neurol. 55, 326–335.
inder, L.I., Guillozet-Bongaarts, A.L., Garcia-Sierra, F., Berry, R.W., 2005.Tau, tangles, and Alzheimer’s disease. Biochim. Biophys. Acta 1739,216–223.
ouras, C., Hof, P.R., Giannakopoulos, P., Michel, J.P., Morrison, J.H., 1994.Regional distribution of neurofibrillary tangles and senile plaques inthe cerebral cortex of elderly patients: a quantitative evaluation of aone-year autopsy population from a geriatric hospital. Cereb. Cortex 4,138–150.
raak, H., Braak, E., 1991. Neuropathological stageing of Alzheimer-relatedchanges. Acta Neuropathol. (Berl) 82, 239–259.
raak, H., Braak, E., 1997. Frequency of stages of Alzheimer-related lesionsin different age categories. Neurobiol. Aging 18, 351–357.
uckner, R.L., Snyder, A.Z., Shannon, B.J., LaRossa, G., Sachs, R., Fotenos,A.F., Sheline, Y.I., Klunk, W.E., Mathis, C.A., Morris, J.C., Mintun,M.A., 2005. Molecular, structural, and functional characterization ofAlzheimer’s disease: evidence for a relationship between default activity,amyloid, and memory. J. Neurosci. 25, 7709–7717.
ullmore, E.T., Suckling, J., Overmeyer, S., Rabe-Hesketh, S., Taylor, E.,Brammer, M.J., 1999. Global, voxel, and cluster tests, by theory and per-mutation, for a difference between two groups of structural MR imagesof the brain. IEEE Trans. Med. Imag. 18, 32–42.
astellani, R.J., Lee, H.G., Perry, G., Smith, M.A., 2006. Antioxidant pro-tection and neurodegenerative disease: the role of amyloid-beta and tau.Am. J. Alzheimers Dis. Other Demen. 21, 126–130.
ollins, D.L., Neelin, P., Peters, T.M., Evans, A.C., 1994. Automatic 3Dintersubject registration of MR volumetric data in standardized Talairachspace. J. Comput. Assist. Tomogr. 18, 192–205.
iannakopoulos, P., Herrmann, F.R., Bussiere, T., Bouras, C., Kovari, E.,Perl, D.P., Morrison, J.H., Gold, G., Hof, P.R., 2003. Tangle and neuronnumbers, but not amyloid load, predict cognitive status in Alzheimer’sdisease. Neurology 60, 1495–1500.
omez-Isla, T., Price, J.L., McKeel Jr., D.W., Morris, J.C., Growdon, J.H.,Hyman, B.T., 1996. Profound loss of layer II entorhinal cortex neu-rons occurs in very mild Alzheimer’s disease. J. Neurosci. 16, 4491–4500.
amilton, M., 1960. A rating scale for depression. J. Neurol. Neurosurg.Psychiat. 23, 56–62.
ayashi K.M., Thompson, P.M., Mega, M.S., Zoumalan, C.I., 2002.Medial hemispheric surface gyral pattern delineation in 3D: sur-face curve protocol http://www.loni.ucla.edu/∼khayashi/Public/Medial
imaging and cognition in normal aging and Alzheimer’s disease.
suface/MedialLinesProtocol. 628
ones, S., Laukka, E.J., Backman, L., 2006. Differential verbal fluency 629
deficits in the preclinical stages of Alzheimer’s disease and vascular 630
dementia. Cortex 42, 347–355. 631
D
INNBA 7198 1–10
1 iology o
K632
633
634
635
636
637
L638
639
640
L641
642
643
644
645
L646
647
648
M649
650
651
M652
653
654
655
656
M657
658
659
660
P661
662
P663
664
665
666
S667
668
S669
670
671
S672
673
674
S675
676
S 677
678
679
680
681
682
683
S 684
685
686
687
688
S 689
690
S 691
692
693
S 694
695
696
T 697
698
699
700
701
T 702
703
704
T 705
706
707
708
T 709
710
711
712
713
V 714
715
716
717
TE
ARTICLE0 M.N. Braskie et al. / Neurob
lunk, W.E., Engler, H., Nordberg, A., Wang, Y., Blomqvist, G., Holt, D.P.,Bergstrom, M., Savitcheva, I., Huang, G.F., Estrada, S., Ausen, B., Deb-nath, M.L., Barletta, J., Price, J.C., Sandell, J., Lopresti, B.J., Wall, A.,Koivisto, P., Antoni, G., Mathis, C.A., Langstrom, B., 2004. Imagingbrain amyloid in Alzheimer’s disease with Pittsburgh Compound-B.Ann. Neurol. 55, 306–319.
erch, J.P., Pruessner, J.C., Zijdenbos, A., Hampel, H., Teipel, S.J., Evans,A.C., 2005. Focal decline of cortical thickness in Alzheimer’s diseaseidentified by computational neuroanatomy. Cereb. Cortex 15, 995–1001.
iu, J., Kepe, V., Zabjek, A., Petric, A., Padgett, H.C., Satyamurthy, N.,Barrio, J.R., 2007. High-yield, automated radiosynthesis of 2-(1-{6-[(2-[18F]fluoroethyl)(methyl)amino]-2-naphthyl}ethylidene)malononitrile([18F]FDDNP) ready for animal or human administration. Mol. Imag.Biol. 9, 6–16.
ogan, J., Fowler, J.S., Volkow, N.D., Wang, G.J., Ding, Y.S., Alexoff, D.L.,1996. Distribution volume ratios without blood sampling from graphicalanalysis of PET data. J. Cereb. Blood Flow Metab. 16, 834–840.
acDonald, D., Kabani, N., Avis, D., Evans, A.C., 2000. Automated 3Dextraction of inner and outer surfaces of cerebral cortex from MRI.Neuroimage 12, 340–356.
cKhann, G., Drachman, D., Folstein, M., Katzman, R., Price, D., Stad-lan, E.M., 1984. Clinical diagnosis of Alzheimer’s disease: report of theNINCDS-ADRDA Work Group under the auspices of Department ofHealth and Human Services Task Force on Alzheimer’s disease. Neu-rology 34, 939–944.
intun, M.A., Larossa, G.N., Sheline, Y.I., Dence, C.S., Lee, S.Y., Mach,R.H., Klunk, W.E., Mathis, C.A., DeKosky, S.T., Morris, J.C., 2006.[11C]PIB in a nondemented population: potential antecedent marker ofAlzheimer disease. Neurology 67, 446–452.
etersen, R.C., 2004. Mild cognitive impairment as a diagnostic entity. J.Intern. Med. 256, 183–194.
rotas, H., Thompson, P.M., Hayashi, K.M., Huang, S.C., 2005. Corticalbrain surface mapping for studying partial volume effects in brain FDG-PET images. In: Proceedings of the 52nd Annual Meeting of the Societyfor Nuclear Medicine (SNM), Toronto, Canada, June 18–22, 2005.
apiro, G., 2001. Geometric Partial Differential Equations and Image Anal-ysis. Cambridge University Press.
elkoe, D.J., 2001. Alzheimer’s disease results from the cerebral accumu-lation and cytotoxicity of amyloid beta-protein. J. Alzheimers Dis. 3,75–80.
hattuck, D.W., Sandor-Leahy, S.R., Schaper, K.A., Rottenberg, D.A.,
UN
CO
RR
EC
Please cite this article in press as: Braskie, M.N., et al., Plaque and tangleNeurobiol Aging (2008), doi:10.1016/j.neurobiolaging.2008.09.012
Leahy, R.M., 2001. Magnetic resonance image tissue classification usinga partial volume model. Neuroimage 13, 856–876.
ingh, V., Chertkow, H., Lerch, J.P., Evans, A.C., Dorr, A.E., Kabani, N.J.,2006. Spatial patterns of cortical thinning in mild cognitive impairmentand Alzheimer’s disease. Brain 129, 2885–2893.
W
PR
OO
F
PRESSf Aging xxx (2008) xxx–xxx
mall, G.W., Ercoli, L.M., Silverman, D.H., Huang, S.C., Komo, S.,Bookheimer, S.Y., Lavretsky, H., Miller, K., Siddarth, P., Rasgon, N.L.,Mazziotta, J.C., Saxena, S., Wu, H.M., Mega, M.S., Cummings, J.L.,Saunders, A.M., Pericak-Vance, M.A., Roses, A.D., Barrio, J.R., Phelps,M.E., 2000. Cerebral metabolic and cognitive decline in persons atgenetic risk for Alzheimer’s disease. Proc. Natl. Acad. Sci. U.S.A. 97,6037–6042.
mall, G.W., Kepe, V., Ercoli, L.M., Siddarth, P., Bookheimer, S.Y., Miller,K.J., Lavretsky, H., Burggren, A.C., Cole, G.M., Vinters, H.V., Thomp-son, P.M., Huang, S.C., Satyamurthy, N., Phelps, M.E., Barrio, J.R.,2006. PET of brain amyloid and tau in mild cognitive impairment. N.Engl. J. Med. 355, 2652–2663.
mith, S.M., 2002. Fast robust automated brain extraction. Hum. BrainMapp. 17, 143–155.
owell E.R., Thompson, P.M., Mega, M.S., Zoumalan, C.I., Lindshield, C.J.,Rex, D.E., 2000. Gyral pattern delineation in 3D: surface curve protocol.www.loni.ucla.edu/∼esowell/new sulcvar.html.
owell, E.R., Thompson, P.M., Leonard, C.M., Welcome, S.E., Kan, E.,Toga, A.W., 2004. Longitudinal mapping of cortical thickness and braingrowth in normal children. J. Neurosci. 24 (38), 8223–8231.
abert, M.H., Manly, J.J., Liu, X., Pelton, G.H., Rosenblum, S., Jacobs,M., Zamora, D., Goodkind, M., Bell, K., Stern, Y., Devanand, D.P.,2006. Neuropsychological prediction of conversion to Alzheimer dis-ease in patients with mild cognitive impairment. Arch. Gen. Psychiat.63, 916–924.
hal, D.R., Rub, U., Orantes, M., Braak, H., 2002. Phases of A beta-deposition in the human brain and its relevance for the developmentof AD. Neurology 58, 1791–1800.
hompson, P.M., Hayashi, K.M., de Zubicaray, G., Janke, A.L., Rose, S.E.,Semple, J., Herman, D., Hong, M.S., Dittmer, S.S., Doddrell, D.M.,Toga, A.W., 2003. Dynamics of gray matter loss in Alzheimer’s disease.J. Neurosci. 23, 994–1005.
hompson, P.M., Hayashi, K.M., Sowell, E.R., Gogtay, N., Giedd, J.N.,Rapoport, J.L., de Zubicaray, G.I., Janke, A.L., Rose, S.E., Semple, J.,Doddrell, D.M., Wang, Y., van Erp, T.G., Cannon, T.D., Toga, A.W.,2004. Mapping cortical change in Alzheimer’s disease, brain develop-ment, and schizophrenia. Neuroimage 23 (Suppl 1), S2–18.
erhoeff, N.P., Wilson, A.A., Takeshita, S., Trop, L., Hussey, D., Singh, K.,Kung, H.F., Kung, M.P., Houle, S., 2004. In-vivo imaging of Alzheimerdisease beta-amyloid with [11C]SB-13 PET. Am. J. Geriatr. Psychiat.12, 584–595.
imaging and cognition in normal aging and Alzheimer’s disease.
atson, D., Castano, E., Kokjohn, T.A., Kuo, Y.M., Lyubchenko, Y., Pinsky, 718
D., Connolly Jr., E.S., Esh, C., Luehrs, D.C., Stine, W.B., Rowse, L.M., 719
Emmerling, M.R., Roher, A.E., 2005. Physicochemical characteristics 720
of soluble oligomeric A beta and their pathologic role in Alzheimer’s 721
disease. Neurol. Res. 27, 869–881. 722