[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Figure 1.
Mean T807 Uptake in Preclinical Stages of Alzheimer Disease (AD)
Mean T807 Uptake in Preclinical Stages of Alzheimer Disease (AD)

Suspected non-AD pathophysiology (SNAP) and preclinical AD stage 0 show indistinguishable levels of tau across all regions. Preclinical AD stages 1 and 2 show elevated tau levels in both medial temporal regions. Stage 2 shows significantly higher levels of tau in the inferior temporal gyrus, whereas levels in stage 1 are intermediate compared with SNAP. Error bars indicate SD. Stages are described in the Introduction.

Figure 2.
Associations of Hippocampal Volume (HV) vs Inferior Temporal Lobe T807 Uptake
Associations of Hippocampal Volume (HV) vs Inferior Temporal Lobe T807 Uptake

Plotted variables are shown as residuals for age. No association was observed between inferior temporal tau levels and HV volume within the group with β-amyloid (Aβ)–negative findings. There was a significant association in the group with Aβ-positive findings that accounted for 15% of the variance above and beyond age. Diagonal lines indicate fit from the linear regression model. SUVR indicates standardized uptake value ratio.

Figure 3.
Longitudinal Change in the Preclinical Alzheimer Cognitive Composite (PACC) by Preclinical Stage
Longitudinal Change in the Preclinical Alzheimer Cognitive Composite (PACC) by Preclinical Stage

Preclinical Alzheimer disease (AD) stage 2 shows decline compared with all other groups. Suspected non-AD pathophysiology (SNAP) shows worse performance over time compared with stage 0. Stages are described in the Introduction.

Table.  
Group Characteristics
Group Characteristics
1.
Sperling  RA, Aisen  PS, Beckett  LA,  et al.  Toward defining the preclinical stages of Alzheimer’s disease: recommendations from the National Institute on Aging–Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease.  Alzheimers Dement. 2011;7(3):280-292.PubMedGoogle ScholarCrossref
2.
Hardy  J, Selkoe  DJ.  The amyloid hypothesis of Alzheimer’s disease: progress and problems on the road to therapeutics.  Science. 2002;297(5580):353-356.PubMedGoogle ScholarCrossref
3.
Jack  CR  Jr, Knopman  DS, Jagust  WJ,  et al.  Tracking pathophysiological processes in Alzheimer’s disease: an updated hypothetical model of dynamic biomarkers.  Lancet Neurol. 2013;12(2):207-216.PubMedGoogle ScholarCrossref
4.
Jack  CR  Jr, Knopman  DS, Weigand  SD,  et al.  An operational approach to National Institute on Aging–Alzheimer’s Association criteria for preclinical Alzheimer disease.  Ann Neurol. 2012;71(6):765-775.PubMedGoogle ScholarCrossref
5.
Jack  CR  Jr, Knopman  DS, Chételat  G,  et al.  Suspected non-Alzheimer disease pathophysiology: concept and controversy.  Nat Rev Neurol. 2016;12(2):117-124.PubMedGoogle ScholarCrossref
6.
Chételat  G.  Alzheimer disease: Aβ-independent processes—rethinking preclinical AD.  Nat Rev Neurol. 2013;9(3):123-124.PubMedGoogle ScholarCrossref
7.
Jack  CR  Jr, Wiste  HJ, Weigand  SD,  et al.  Amyloid-first and neurodegeneration-first profiles characterize incident amyloid PET positivity.  Neurology. 2013;81(20):1732-1740.PubMedGoogle ScholarCrossref
8.
Knopman  DS, Jack  CR  Jr, Wiste  HJ,  et al.  Brain injury biomarkers are not dependent on β-amyloid in normal elderly.  Ann Neurol. 2013;73(4):472-480.PubMedGoogle ScholarCrossref
9.
Wirth  M, Villeneuve  S, Haase  CM,  et al.  Associations between Alzheimer disease biomarkers, neurodegeneration, and cognition in cognitively normal older people.  JAMA Neurol. 2013;70(12):1512-1519.PubMedGoogle Scholar
10.
Crary  JF, Trojanowski  JQ, Schneider  JA,  et al.  Primary age-related tauopathy (PART): a common pathology associated with human aging.  Acta Neuropathol. 2014;128(6):755-766.PubMedGoogle ScholarCrossref
11.
Jack  CR  Jr.  PART and SNAP.  Acta Neuropathol. 2014;128(6):773-776.PubMedGoogle ScholarCrossref
12.
Jellinger  KA, Alafuzoff  I, Attems  J,  et al.  PART, a distinct tauopathy, different from classical sporadic Alzheimer disease.  Acta Neuropathol. 2015;129(5):757-762.PubMedGoogle ScholarCrossref
13.
Braak  H, Braak  E.  Frequency of stages of Alzheimer-related lesions in different age categories.  Neurobiol Aging. 1997;18(4):351-357.PubMedGoogle ScholarCrossref
14.
Nelson  PT, Alafuzoff  I, Bigio  EH,  et al.  Correlation of Alzheimer disease neuropathologic changes with cognitive status: a review of the literature.  J Neuropathol Exp Neurol. 2012;71(5):362-381.PubMedGoogle ScholarCrossref
15.
Duyckaerts  C, Braak  H, Brion  JP,  et al.  PART is part of Alzheimer disease.  Acta Neuropathol. 2015;129(5):749-756.PubMedGoogle ScholarCrossref
16.
Chien  DT, Bahri  S, Szardenings  AK,  et al.  Early clinical PET imaging results with the novel PHF-tau radioligand [F-18]-T807.  J Alzheimers Dis. 2013;34(2):457-468.PubMedGoogle Scholar
17.
Johnson  KA, Schultz  A, Betensky  RA,  et al.  Tau positron emission tomographic imaging in aging and early Alzheimer disease.  Ann Neurol. 2016;79(1):110-119.PubMedGoogle ScholarCrossref
18.
Knopman  DS, Jack  CR  Jr, Wiste  HJ,  et al.  Short-term clinical outcomes for stages of NIA-AA preclinical Alzheimer disease.  Neurology. 2012;78(20):1576-1582.PubMedGoogle ScholarCrossref
19.
Mormino  EC, Betensky  RA, Hedden  T,  et al.  Synergistic effect of β-amyloid and neurodegeneration on cognitive decline in clinically normal individuals.  JAMA Neurol. 2014;71(11):1379-1385.PubMedGoogle ScholarCrossref
20.
Vos  SJ, Xiong  C, Visser  PJ,  et al.  Preclinical Alzheimer’s disease and its outcome: a longitudinal cohort study.  Lancet Neurol. 2013;12(10):957-965.PubMedGoogle ScholarCrossref
21.
Morris  JC.  The Clinical Dementia Rating (CDR): current version and scoring rules.  Neurology. 1993;43(11):2412-2414.PubMedGoogle ScholarCrossref
22.
Wechsler  D.  Wechsler Memory Scale–Revised. San Antonio, TX: The Psychological Corp; 1987.
23.
Folstein  MF, Folstein  SE, McHugh  PR.  “Mini-mental state”: a practical method for grading the cognitive state of patients for the clinician.  J Psychiatr Res. 1975;12(3):189-198.PubMedGoogle ScholarCrossref
24.
Fischl  B, Salat  DH, Busa  E,  et al.  Whole brain segmentation: automated labeling of neuroanatomical structures in the human brain.  Neuron. 2002;33(3):341-355.PubMedGoogle ScholarCrossref
25.
Landau  SM, Harvey  D, Madison  CM,  et al; Alzheimer’s Disease Neuroimaging Initiative.  Associations between cognitive, functional, and FDG-PET measures of decline in AD and MCI.  Neurobiol Aging. 2011;32(7):1207-1218.PubMedGoogle ScholarCrossref
26.
Mormino  EC, Betensky  RA, Hedden  T,  et al; Alzheimer’s Disease Neuroimaging Initiative; Australian Imaging Biomarkers and Lifestyle Flagship Study of Ageing; Harvard Aging Brain Study.  Amyloid and APOE ε4 interact to influence short-term decline in preclinical Alzheimer disease.  Neurology. 2014;82(20):1760-1767.PubMedGoogle ScholarCrossref
27.
Amariglio  RE, Mormino  EC, Pietras  AC,  et al.  Subjective cognitive concerns, amyloid-β, and neurodegeneration in clinically normal elderly.  Neurology. 2015;85(1):56-62.PubMedGoogle ScholarCrossref
28.
Papp  KV, Amariglio  RE, Mormino  EC,  et al.  Free and cued memory in relation to biomarker-defined abnormalities in clinically normal older adults and those at risk for Alzheimer’s disease.  Neuropsychologia. 2015;73:169-175.PubMedGoogle ScholarCrossref
29.
Donohue  MC, Sperling  RA, Salmon  DP,  et al; Australian Imaging, Biomarkers, and Lifestyle Flagship Study of Ageing; Alzheimer’s Disease Neuroimaging Initiative; Alzheimer’s Disease Cooperative Study.  The preclinical Alzheimer Cognitive Composite: measuring amyloid-related decline.  JAMA Neurol. 2014;71(8):961-970.PubMedGoogle ScholarCrossref
30.
Dickerson  BC, Bakkour  A, Salat  DH,  et al.  The cortical signature of Alzheimer’s disease: regionally specific cortical thinning relates to symptom severity in very mild to mild AD dementia and is detectable in asymptomatic amyloid-positive individuals.  Cereb Cortex. 2009;19(3):497-510.PubMedGoogle ScholarCrossref
31.
Rowe  CC, Ellis  KA, Rimajova  M,  et al.  Amyloid imaging results from the Australian Imaging, Biomarkers and Lifestyle (AIBL) study of aging.  Neurobiol Aging. 2010;31(8):1275-1283.PubMedGoogle ScholarCrossref
32.
Dawe  RJ, Bennett  DA, Schneider  JA, Arfanakis  K.  Neuropathologic correlates of hippocampal atrophy in the elderly: a clinical, pathologic, postmortem MRI study.  PLoS One. 2011;6(10):e26286.PubMedGoogle ScholarCrossref
33.
Josephs  KA, Whitwell  JL, Tosakulwong  N,  et al.  TAR DNA-binding protein 43 and pathological subtype of Alzheimer’s disease impact clinical features.  Ann Neurol. 2015;78(5):697-709.PubMedGoogle ScholarCrossref
34.
Raz  N, Gunning-Dixon  F, Head  D, Rodrigue  KM, Williamson  A, Acker  JD.  Aging, sexual dimorphism, and hemispheric asymmetry of the cerebral cortex: replicability of regional differences in volume.  Neurobiol Aging. 2004;25(3):377-396.PubMedGoogle ScholarCrossref
35.
Knopman  DS, Jack  CR  Jr, Wiste  HJ,  et al.  18F-fluorodeoxyglucose positron emission tomography, aging, and apolipoprotein E genotype in cognitively normal persons.  Neurobiol Aging. 2014;35(9):2096-2106.PubMedGoogle ScholarCrossref
36.
Jack  CR  Jr, Dickson  DW, Parisi  JE,  et al.  Antemortem MRI findings correlate with hippocampal neuropathology in typical aging and dementia.  Neurology. 2002;58(5):750-757.PubMedGoogle ScholarCrossref
37.
Thal  DR, Beach  TG, Zanette  M,  et al.  [(18)F]flutemetamol amyloid positron emission tomography in preclinical and symptomatic Alzheimer’s disease: specific detection of advanced phases of amyloid-β pathology.  Alzheimers Dement. 2015;11(8):975-985.PubMedGoogle ScholarCrossref
Original Investigation
October 2016

Heterogeneity in Suspected Non–Alzheimer Disease Pathophysiology Among Clinically Normal Older Individuals

Author Affiliations
  • 1Department of Neurology, Massachusetts General Hospital, Harvard Medical School, Charlestown
  • 2Center for Alzheimer Research and Treatment, Department of Neurology, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts
  • 3Athinoula A. Martinos Center for Biomedical Imaging, Department of Radiology, Massachusetts General Hospital, Charlestown
  • 4Department of Radiology, Massachusetts General Hospital, Harvard Medical School, Boston
  • 5Division of Nuclear Medicine and Molecular Imaging, Department of Radiology, Massachusetts General Hospital, Harvard Medical School, Boston
 

Copyright 2016 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.

JAMA Neurol. 2016;73(10):1185-1191. doi:10.1001/jamaneurol.2016.2237
Key Points

Question  Do cognitively normal older individuals with suspected non–Alzheimer disease pathophysiology (SNAP) show evidence of early Alzheimer disease processes?

Findings  In this longitudinal study, the SNAP group did not show elevated levels of tau or cognitive decline. A small subset of the SNAP group with subthreshold Pittsburgh Compound B values and small hippocampus showed rapid cognitive decline.

Meaning  Those with SNAP are heterogeneous, and further biomarker refinement will be necessary to characterize this group.

Abstract

Importance  A substantial proportion of clinically normal (CN) older individuals are classified as having suspected non–Alzheimer disease pathophysiology (SNAP), defined as biomarker negative for β-amyloid (Aβ−) but positive for neurodegeneration (ND+). The etiology of SNAP in this population remains unclear.

Objective  To determine whether CN individuals with SNAP show evidence of early Alzheimer disease (AD) processes (ie, elevated tau levels and/or increased risk for cognitive decline).

Design, Setting, and Participants  This longitudinal observational study performed in an academic medical center included 247 CN participants from the Harvard Aging Brain Study. Participants were classified into preclinical AD stages using measures of Aβ (Pittsburgh Compound B [PIB]–labeled positron emission tomography) and ND (hippocampal volume or cortical glucose metabolism from AD-vulnerable regions). Classifications included stages 0 (Aβ−/ND−), 1 (Aβ+/ND−), and 2 (Aβ+/ND+) and SNAP (Aβ−/ND+). Continuous levels of PiB and ND, tau levels in the medial and inferior temporal lobes, and longitudinal cognition were examined. Data collection began in 2010 and is ongoing. Data were analyzed from 2015 to 2016.

Main Outcomes and Measures  Evidence of amyloid-independent tau deposition and/or cognitive decline.

Results  Of the 247 participants (142 women [57.5%]; 105 men [42.5%]; mean age, 74 [range, 63-90] years), 64 (25.9%) were classified as having SNAP. Compared with the stage 0 group, the SNAP group was not more likely to have subthreshold PiB values (higher values within the Aβ− range), suggesting that misclassification due to the PiB cutoff was not a prominent contributor to this group (mean [SD] distribution volume ratio, 1.08 [0.05] for the SNAP group; 1.09 [0.05] for the stage 1 group). Tau levels in the medial and inferior temporal lobes were indistinguishable between the SNAP and stage 0 groups (entorhinal cortex, β = −0.005 [SE, 0.036]; parahippocampal gyrus, β = −0.001 [SE, 0.027]; and inferior temporal lobe, β = −0.004 [SE, 0.027]; P ≥ .88) and were lower in the SNAP group compared with the stage 2 group (entorhinal cortex, β = −0.125 [SE, 0.041]; parahippocampal gyrus, β = −0.074 [SE, 0.030]; and inferior temporal lobe, β = −0.083 [SE, 0.031]; P ≤ .02). The stage 2 group demonstrated greater cognitive decline compared with all other groups (stage 0, β = −0.239 [SE, 0.042]; stage 1, β = −0.242 [SE, 0.051]; and SNAP, β = −0.157 [SE, 0.044]; P ≤ .001), whereas the SNAP group showed a diminished practice effect over time compared with the stage 0 group (β = −0.082 [SE, 0.037]; P = .03).

Conclusions and Relevance  In this study, clinically normal adults with SNAP did not exhibit evidence of elevated tau levels, which suggests that this biomarker construct does not represent amyloid-independent tauopathy. At the group level, individuals with SNAP did not show cognitive decline but did show a diminished practice effect. SNAP is likely heterogeneous, with a subset of this group at elevated risk for short-term decline. Future refinement of biomarkers will be necessary to subclassify this group and determine the biological correlates of ND markers among Aβ− CN individuals.

Introduction

Quiz Ref IDIn 2011, the National Institute on Aging and the Alzheimer’s Association workgroup published criteria for classifying clinically normal (CN) older individuals thought to be on the trajectory into stages of preclinical Alzheimer disease (AD).1 This staging framework postulated a sequence that begins with β-amyloid (Aβ) accumulation, followed by neurodegeneration (ND) and eventual cognitive decline.2,3 Individuals in preclinical stage 1 have positive Aβ findings (Aβ+) but negative findings for ND (ND−); stage 2, Aβ+/ND+; and stage 3, Aβ+/ND+ and show subtle cognitive impairment. Soon after the publication of these criteria, Quiz Ref IDJack et al4 described an additional category of CN individuals who were Aβ−/ND+ (ie, suspected non–Alzheimer disease pathophysiology [SNAP]). Quiz Ref IDInterestingly, the proportion of CN individuals with SNAP has been remarkably consistent at approximately 25% across multiple independent cohorts.5

The relevance of SNAP in CN individuals to the conceptualization of preclinical AD is currently unclear.5,6 Among Aβ− CN individuals, baseline ND markers are not associated with subsequent accumulation of Aβ,7 suggesting that CN individuals with SNAP are not at elevated risk for entering the AD cascade compared with individuals in stage 0. In addition, markers of non-AD pathologic processes, such as cerebrovascular disease and α-synucleinopathy, are not more prevalent in SNAP8 despite findings by Wirth et al.9 A remaining possibility is that SNAP in CN individuals, or at least a portion of the SNAP CN group, reflects amyloid-independent tauopathy.10-12Quiz Ref ID Tau aggregation in the medial temporal lobe is ubiquitous in aging (approximately 94% of individuals in their 70s are Braak stage I and higher13). Although tau aggregation beyond the medial temporal lobe is coupled with Aβ accumulation,13,14 a subset of individuals with low Aβ levels show this more extensive pattern of tau deposition. The presence of tau aggregation in the medial temporal lobe and beyond among Aβ− participants has recently been labeled primary age-related tauopathy (PART) and is currently under discussion.10,12,15 An intriguing possibility is that CN individuals with SNAP are the in vivo analogue of this postmortem group. We can test this possibility using positron emission tomographic (PET) imaging that enables assessment of the spatial distribution of tau aggregates.16,17

Another clarification regarding the relevance of SNAP is whether this group shows cognitive decline over time. If PART is a contributing cause of SNAP, individuals with SNAP should show cognitive decline, given that cases with PART demonstrate worse cognitive scores than their counterparts with low Aβ and tau levels.10 Although most studies to date have found elevated decline in individuals in stage 2, results in CN individuals with SNAP vary.18-20 A previous investigation19 reported intermediate levels of longitudinal change during 2 years in CN individuals with SNAP on a global cognitive composite measure. We sought to expand on this finding by examining a larger sample with a longer follow-up and additionally exploring correlates of decline within the SNAP group. Overall, the primary goal of the present study was to explore whether CN individuals with SNAP show evidence of amyloid-independent tau deposition and/or cognitive decline.

Methods
Participants

Quiz Ref IDParticipants in the Harvard Aging Brain Study (HABS) undergo baseline magnetic resonance imaging (MRI) and PET scanning and annual neuropsychological testing. Data collection started in 2010 and is ongoing. Study protocols were approved by the institutional review board of Partners Healthcare, and all participants provided written informed consent.

At baseline, participants had a global Clinical Dementia Rating score of 0,21 performed within educational level–adjusted norms on the Logical Memory Delayed Recall score,22 and had a Mini-Mental State Examination (MMSE) score of at least 27 (range, 27 to 30, with higher scores indicating better performance).23 Two hundred forty-seven participants included in the present analyses completed carbon 11–labeled Pittsburgh Compound B (PiB) PET, fludeoxyglucose F 18 (FDG) PET, and structural MRI scanning within 1 year of baseline (Table). Eighty of these participants additionally underwent T807 PET within 1 year of the other imaging procedures.

Magnetic Resonance Imaging

Magnetic resonance imaging was completed at the Athinoula A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital, Charlestown, on a 3-T system with a 12-channel head coil (TIM Trio; Siemens). Structural T1-weighted, volumetric magnetization-prepared, rapid acquisition gradient-echo scans (repetition time, 6400 milliseconds; echo time, 2.8 milliseconds; inversion time, 900 milliseconds; flip angle, 8°; and 1 × 1 × 1.2-mm resolution) were used to extract hippocampal volume (HV) with FreeSurfer (version 5.1).24 Total bilateral HV was adjusted for estimated total intracranial volume.19

Positron Emission Tomography

Positron emission tomographic scanning using PiB, T807, and FDG radioligands was completed at the Massachusetts General Hospital PET facility, Boston, using a high-resolution scanner (ECAT EXACT HR+; Siemens) in a 3-dimensional mode (63 image planes; 15.2-cm axial field of view; 5.6-mm transaxial resolution; and 2.4-mm section interval). Carbon 11–labeled PiB and fluoride 18–labeled T807 were synthesized using previously published protocols.17

Ten-minute transmission scans for attenuation correction were collected before PET data. For PiB, 8.5 to 15.0 mCi were injected and 60 minutes of dynamic data were acquired in 69 frames (12 × 15 seconds and 57 × 60 seconds, respectively). Acquisition of T807 occurred from 80 to 100 minutes after a 9.0- to 11.0-mCi injection in 4 × 5-minute frames. For FDG, 5.0 to 10.0 mCi was injected, and the images were acquired across 6 × 5-minute frames 45 minutes after the injection.

Preprocessing for PET was performed using SPM8 (Wellcome Trust Centre for Neuroimaging). The PiB images were realigned, and the mean of the first 8 minutes was calculated and used for normalization to the MNI (Montreal Neurological Institute) FDG template. Distribution volume ratio (DVR) images were created with Logan plotting (40-60 minutes, gray matter cerebellar reference). The PiB signal from a global cortical aggregate was extracted for each participant.19 We realigned, summed, and coregistered T807 PET data to each participant’s MRI. The T807 was extracted from FreeSurfer-defined bilateral entorhinal, parahippocampal, and inferior temporal gyrus and expressed as the standardized uptake value ratio (SUVR) relative to a gray matter cerebellar reference.17 The FDG was extracted from a region of interest reflecting AD-vulnerable cortical regions from a previously published meta-analysis of hypometabolic regions in AD (MetaROI) and normalized using a pons-vermis reference region.25

Classification of Preclinical AD Stages

A Gaussian mixture modeling approach was used to classify HABS CN individuals as Aβ+ or Aβ− (cutoff value, DVR of 1.20).26 Although ND markers vary across studies,5 the most commonly used ND markers are adjusted HV (aHV), MetaROI FDG, and cerebrospinal fluid (CSF) tau level. Previous studies19,27,28 have used aHV and FDG to classify participants into preclinical stages because these data are available for most of the HABS participants (CSF tau levels are only collected on a subset) and are consistent with classification procedures used by Jack and colleagues4 and Mormino and colleagues.19 Specifically, participants were classified as ND+ if positive for aHV or MetaFDG findings (using a cutoff of 6723 mm3 for aHV and SUVR of 1.249 for MetaROI FDG).19 Further details regarding classification into ND groups are presented in the eMethods and eTable 1 in the Supplement. Based on joint Aβ and ND status, CN individuals were classified as being in stage 0 (Aβ−/ND−), stage 1 (Aβ+/ND−), stage 2 (Aβ+/ND+), or SNAP (Aβ−/ND+).4

Neuropsychological Testing

We assessed cognition using the Preclinical Alzheimer Cognitive Composite (PACC),29 which consists of the (1) Free and Cued Selective Reminding Test Cued Recall, (2) Logical Memory delayed recall, (3) Digit Symbol Coding, and (4) MMSE. Measures were z transformed based on the mean and SD from baseline data, and a mean was calculated.

Statistical Analysis

Data were analyzed from 2015 to 2016. Analyses were performed using R software (version 3.2; https://www.r-project.org/). Differences in demographics across preclinical stages were examined with 2-sided t tests for continuous variables and χ2 tests for dichotomous variables. Multiple linear regression models were used to examine biomarker differences, controlled for age and sex.

We used linear mixed models to examine change in the PACC. All models included covariates for preclinical stage, age, sex, and educational level, as well as each covariate’s interaction with time from baseline. A random intercept and slope were included for each participant. We were specifically interested in contrasting SNAP with stages 0, 1, and 2. To further explore variation within SNAP, we examined the association between continuous values of PiB and ND markers within the SNAP group. All P values were 2 sided, and no correction for multiple comparisons was performed.

Results
Continuous Levels of PiB and ND

Two hundred forty-seven participants (142 women [57.5%]; 105 men [42.5%]; mean age, 74 [range, 63-90] years) were included in the analyses. Participant characteristics are found in the Table. Given that biomarker cutoff selection may affect group classification, it is possible that CN individuals with SNAP are more likely to have greater levels of subthreshold PiB compared with the stage 0 group (values of PiB below the cutoff DVR of 1.20) and/or lower levels of ND markers compared with the stage 2 group (values of ND markers just above the cutoff for ND positivity). We therefore contrasted continuous levels of subthreshold PiB between SNAP and stage 0 (below the Aβ cutoff) and ND markers between SNAP and stage 2 (above the ND cutoff). Although the difference did not reach statistical significance, CN individuals with SNAP showed less PiB uptake compared with the stage 0 group (P = .07), suggesting that no evidence existed for higher levels of subthreshold PiB in SNAP compared with stage 0 (mean [SD] DVR, 1.08 [0.05] for SNAP; 1.09 [0.05] for stage 1). In addition, no differences were found between SNAP and stage 2 for continuous levels of aHV (mean [SD], 6966 [753] mm3 for SNAP; 6781 [973] mm3 for stage 2; P = .27) or metaFDG (mean [SD] SUVR, 1.23 [0.10] for SNAP; 1.21 [0.08] for stage 2; P = .27). To determine whether SNAP shows a distinct ND pattern, we contrasted vertexwise gray matter thickness between SNAP and stage 2 and did not find any regions showing reduced thickness in SNAP (eFigure 1 in the Supplement).

Tau Imaging Across Preclinical Stages

Examination of regional tau levels as measured with T807 PET revealed less tau in SNAP compared with stages 1 and 2 in the entorhinal cortex (SNAP vs stage 1: β = −0.106 [SE, 0.043]; P = .02; SNAP vs stage 2: β = −0.125 [SE, 0.041]; P = .003) and the parahippocampal gyrus (SNAP vs stage 1: β = −0.063 [SE, 0.032]; P = .049; SNAP vs stage 2: β = −0.074 [SE, 0.030]; P = .02). The T807 signal in the inferior temporal gyrus was significantly lower in SNAP compared with stage 2 (β = −0.083 [SE, 0.031]; P = .01), but this difference did not reach statistical significance compared with stage 1 (β = −0.044 [SE, 0.033]; P = .18). The T807 signal between SNAP and stage 0 was indistinguishable across all 3 regions (SNAP vs stage 0 for entorhinal cortex: β = −0.005 [SE, 0.036]; parahippocampal gyrus: β = −0.001 [SE, 0.027]; and inferior temporal lobe: β = −0.004 [SE, 0.027]; P ≥ .88). Stages 2 and 1 showed similar levels of tau in medial temporal regions (P ≥ .68). Although not statistically significant from stage 2, stage 1 showed intermediate values of T807 in the inferior temporal gyrus (P = .28) (Figure 1). Examination of continuous T807 vs PiB levels confirmed that the T807 signal among Aβ− participants was not related to SNAP (eFigure 2 in the Supplement). Furthermore, a higher T807 signal in the inferior temporal lobe was associated with a smaller aHV within Aβ+ (P = .04) but not within Aβ− (P = .81) groups (Figure 2). The association between metaROI FDG and T807 values did not reach significance within the Aβ+ (P = .11) or Aβ− (P = .61) groups.

Longitudinal Change in Cognitive Performance on PACC

We found no baseline differences on the PACC scores between the SNAP group and any other group (P ≥ .87; eTable 2 in the Supplement). Examination of longitudinal change revealed that the SNAP group showed better performance over time on the PACC compared with the stage 2 group (β = 0.157 [SE, 0.044]; P < .001). The SNAP group showed worse performance over time compared with the stage 0 group (β = −0.082 [SE, 0.037]; P = .03), which was primarily driven by a diminished practice effect. The difference between SNAP and stage 1 was not statistically significant (β = −0.085 [SE, 0.047]; P = .07) (Figure 3). The stage 2 group demonstrated greater cognitive decline compared with all other groups (stage 2 vs stage 0: β = −0.239 [SE, 0.042]; stage 2 vs stage 1: β = −0.242 [SE, 0.051]; and stage 2 vs SNAP: β = −0.157 [SE, 0.044]; P ≤ .001).

Examination of individual trajectories across preclinical stages revealed 2 participants with SNAP who showed rapid cognitive decline (eFigure 3 in the Supplement). We therefore repeated longitudinal models excluding these 2 participants and found a marginally significant difference between SNAP and stage 0 (β = −0.053 [0.038]; P = .08) and no difference between SNAP and stage 1 (β = −0.045 [0.038]; P = .24) (eFigure 4 in the Supplement). Thus, the diminished practice effect observed in SNAP compared with stage 0 was not solely driven by the 2 participants with rapid decline.

To further understand cognitive change within SNAP, we examined continuous levels of ND and subthreshold PiB (continuous values of PiB below the cutoff DVR of 1.20) on the longitudinal PACC scores within the SNAP group. This analysis revealed that higher subthreshold PiB levels (P = .001) and reduced aHV (P = .002) were associated with worse PACC performance over time in the SNAP group. We found no significant contribution of MetaROI FDG (P = .60). These effects were no longer significant after excluding the 2 participants with rapid decline in the SNAP group (effect of subthreshold PiB: P = .48; effect of reduced aHV: P = .41). A similar analysis within stage 0 did not reveal any significant associations between cognitive change and subthreshold PiB levels (P = .20), MetaROI FDG (P = .95), or aHV (P = .08).

Discussion

Among the CN HABS participants, 64 (25.9%) were classified as having SNAP (Aβ−/ND+). Using T807 PET imaging, we found that the SNAP group had lower levels of medial temporal lobe tau compared with the stages 1 and 2 groups and levels similar to those of the stage 0 group. Furthermore, the SNAP group had similar levels of inferior temporal lobe tau as the stage 0 group but significantly lower tau levels compared with the stage 2 group. At the group level, CN individuals with SNAP showed a diminished practice effect over time compared with the stage 0 group and better performance over time compared with the stage 2 group. Examination within the SNAP group revealed that subthreshold PiB values and reduced HV were associated with decline, an effect that was driven by 2 individuals with SNAP and rapid decline. Overall, these results highlight that patterns of ND in AD-vulnerable regions as detected with HV and cortical glucose metabolism are not specific to AD processes among CN individuals. Instead, multiple causes likely contribute to the biomarker construct of SNAP.

The presence of small hippocampi and reduced cortical metabolism in AD-vulnerable regions among Aβ− CN individuals highlights factors beyond Aβ that influence variability among ND markers in aging. Although ND markers have been associated with Aβ,30,31 they are also influenced by cerebrovascular disease,9 hippocampal sclerosis,32 and the 43-kDa transactive response DNA-binding protein.33 Likewise, associations between chronological age and gray matter34 and cortical metabolism35 are present throughout the lifespan, well before the age of Aβ accumulation.14 Given that ND markers used in our analyses are cross-sectional, these markers may also be influenced by early-life brain reserve factors (suggesting that the term neurodegeneration is a misnomer in at least some cases). Thus, abnormal ND levels measured with HV and cortical glucose metabolism do not appear to be specific to AD processes36 but are also likely influenced by a number of age-related pathologic processes, the normal aging process, and interindividual differences.

Given the high prevalence of medial temporal tangle abnormalities in aging14 and the presence of tau aggregates extending into inferior temporal cortex in a subset of participants with low levels of Aβ,10 one hypothesis is that elevated tau levels constitute a pathologic substrate of CN individuals with SNAP.11 However, examination of the T807 signal did not reveal higher levels of tau in the medial temporal or inferior temporal lobe in SNAP compared with stage 0. In fact, the T807 signal was elevated in stages 1 and 2 compared with SNAP within medial temporal lobe regions. Although the inferior temporal T807 signal was only statistically higher in stage 2 compared with SNAP, stage 1 showed intermediate levels in this region, whereas the mean values between stage 0 and SNAP were nearly identical. Thus, these data do not support the hypothesis that the biomarker construct of SNAP is analogous to the postmortem construct of PART.10

In the sample described by Crary et al,10 182 of 434 participants (41.9%) would, if they had undergone amyloid imaging with PET, likely be classified as Aβ− because they were Thal Aβ phase 0-2.37 Of these 182 Aβ− cases, 77 (42.3%) were Braak stage III or IV, were approximately 90 years of age, and had MMSE scores of approximately 23.10 In the HABS cohort, the proportion of Aβ− participants is much larger (181 of 247 [73.3%]), which is expected given the younger mean age of 74 years in the HABS cohort. Among the 181 Aβ− participants in HABS, 64 (35.4%) had abnormal ND biomarkers, a mean age of 76.5 years, and MMSE scores of 29. Given the restrictive enrollment requirements in a study of CN individuals (ie, baseline MMSE score, ≥27), cases with tau aggregation extending into Braak stage III or IV would likely be excluded from the HABS, which is consistent with our finding of a low tau PET signal in SNAP. More concordance may exist between SNAP and PART in older populations and/or in mild cognitive impairment but not within a CN sample with a mean age of 74 years.

Nevertheless, our analyses emphasize discordance between ND markers used to define SNAP and tau PET imaging. A higher T807 signal in the inferior temporal lobe was significantly associated with a smaller HV within Aβ+ participants but not within Aβ− participants (the association between inferior temporal T807 signal and HV among Aβ+ participants only accounted for 15% of the variance, suggesting that these markers are not completely aligned even within the Aβ+ group). This discordance has important implications for staging criteria of preclinical AD given that classifying participants using tau imaging will differ from approaches that use HV or MetaROI FDG.

The SNAP group showed a diminished practice effect over time compared with the stage 0 group and better performance than the stage 2 group. This finding is consistent with a previous publication19 examining longitudinal cognition for a shorter duration and in fewer participants, such that short-term decline was most prominent in Aβ+/ND+ participants. Although the group level effect was influenced by 2 SNAP participants showing rapid decline (3% of the entire SNAP sample of 64 participants), the difference between SNAP and stage 0 was marginally significant after excluding these 2 participants (the β describing the difference between SNAP and stage 0 was reduced from −0.085 to −0.053 after excluding both). Within the SNAP group, we found an association between subthreshold PiB retention and decline in the PACC score, but this effect was driven by the 2 participants with rapid decline. The finding that subthreshold PiB values were associated with cognitive change within SNAP is notable given that the SNAP group did not show greater levels of subthreshold PiB compared with the stage 0 group and that subthreshold PiB values were not associated with cognitive change within the stage 0 group. This finding suggests that preexisting ND puts a small subset of SNAP participants who are additionally confronted with early Aβ accumulation at elevated risk for cognitive decline.

Our study has several limitations. We only examined cross-sectional markers of ND, and longitudinal change in these markers might give a better estimate of the ND processes. The lack of concordance between ND status and T807 signal among Aβ− participants may reflect the spatial distribution of the ND markers used in our analyses, whereas more focal structural atrophy and/or hypometabolism may be a better correlate of amyloid-independent tauopathy. Although we did not find evidence of elevated neocortical or medial temporal lobe tau levels in SNAP compared with stage 0, tau species not detectable with the T807 radioligand may be present within SNAP. Finally, our analyses examining cognitive decline within SNAP are limited by the reduced variation in cognitive change among these groups, warranting future analyses with a longer follow-up.

Conclusions

Approximately 25% of CN participants in the HABS are classified as having SNAP. The lack of group-level associations between the SNAP group and AD processes (subthreshold PiB retention and elevated tau levels) suggests that ND markers are influenced by multiple causes. Postmortem studies will be critical to determine pathologic correlates of SNAP, and the development of novel molecular biomarkers will help to subclassify this group in vivo.

Back to top
Article Information

Corresponding Author: Elizabeth C. Mormino, PhD, Athinoula A. Martinos Center for Biomedical Imaging, Department of Neurology, Massachusetts General Hospital, Harvard Medical School, 149 13th St, Charlestown, MA 02129 (bmormino@nmr.mgh.harvard.edu).

Accepted for Publication: May 12, 2016.

Published Online: August 22, 2016. doi:10.1001/jamaneurol.2016.2237

Author Contributions: Dr Mormino had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Mormino, Rentz, Johnson, Sperling.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Mormino, Papp, Johnson, Sperling.

Critical revision of the manuscript for important intellectual content: Papp, Rentz, Schultz, LaPoint, Amariglio, Hanseeuw, Marshall, Hedden, Johnson, Sperling.

Statistical analysis: Mormino, Schultz, Hanseeuw, Johnson.

Obtained funding: Mormino, Hedden, Johnson, Sperling.

Administrative, technical, or material support: Papp, LaPoint, Johnson, Sperling.

Study supervision: Rentz, Hedden, Johnson, Sperling.

Conflict of Interest Disclosures: Dr Mormino reports receiving funding from grants from the National Institutes of Health (NIH). Dr Papp reports receiving funding through a grant from the NIH and the Charles King Trust Foundation. Dr Rentz reports receiving research support from grants from the NIH, the Alzheimer’s Association, and funding from Fidelity Biosciences and serving as a paid consultant for Eli Lilly and Co, Janssen Pharmaceuticals, and Neurotrack, relationships that are not related to the content in the manuscript. Dr Amariglio reports receiving grants from the Alzheimer’s Association and the NIH. Dr Hanseeuw reports receiving support from the Belgian American Education Foundation. Dr Marshall reports receiving research support from the NIH; salary support from Eisai, Inc, and Eli Lilly and Co; and serving as a paid consultant for Halloran/GliaCure, relationships that are not related to the content in the manuscript. Dr Hedden reports receiving grants from the NIH. Dr Johnson reports serving as a paid consultant for Bayer, GE Healthcare, Janssen Alzheimer’s Immunotherapy, Siemens Medical Solutions, Genzyme, Novartis, Biogen, Roche, ISIS Pharma, AZTherapy, GEHC, Lundberg, and AbbVie; serving as a site coinvestigator for Eli Lilly/Avid, Pfizer, Janssen Immunotherapy, and Navidea; speaking at symposia sponsored by Janssen Alzheimer’s Immunotherapy and Pfizer, relationships that are not related to the content in the manuscript; and receiving grants from the NIH and the Alzheimer’s Association. Dr Sperling reports serving as a paid consultant for AbbVie, Biogen, Bracket, Genentech, Lundbeck, Roche, and Sanofi; serving as a coinvestigator for Avid, Eli Lilly and Co, and Janssen Alzheimer Immunotherapy clinical trials; speaking at symposia sponsored by Eli Lilly and Co, Biogen, and Janssen; receiving research support from Janssen Pharmaceuticals and Eli Lilly and Co, relationships that are not related to the content in the manuscript; and receiving research support from the NIH, Fidelity Biosciences, Harvard NeuroDiscovery Center, and the Alzheimer’s Association. No other disclosures were reported.

Funding/Support: The study was supported grants P01 AG036694 and R01 AG046396 from the NIH and grants F32AG044054 and K24AG035007 from the National Institute on Aging.

Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, or interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

References
1.
Sperling  RA, Aisen  PS, Beckett  LA,  et al.  Toward defining the preclinical stages of Alzheimer’s disease: recommendations from the National Institute on Aging–Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease.  Alzheimers Dement. 2011;7(3):280-292.PubMedGoogle ScholarCrossref
2.
Hardy  J, Selkoe  DJ.  The amyloid hypothesis of Alzheimer’s disease: progress and problems on the road to therapeutics.  Science. 2002;297(5580):353-356.PubMedGoogle ScholarCrossref
3.
Jack  CR  Jr, Knopman  DS, Jagust  WJ,  et al.  Tracking pathophysiological processes in Alzheimer’s disease: an updated hypothetical model of dynamic biomarkers.  Lancet Neurol. 2013;12(2):207-216.PubMedGoogle ScholarCrossref
4.
Jack  CR  Jr, Knopman  DS, Weigand  SD,  et al.  An operational approach to National Institute on Aging–Alzheimer’s Association criteria for preclinical Alzheimer disease.  Ann Neurol. 2012;71(6):765-775.PubMedGoogle ScholarCrossref
5.
Jack  CR  Jr, Knopman  DS, Chételat  G,  et al.  Suspected non-Alzheimer disease pathophysiology: concept and controversy.  Nat Rev Neurol. 2016;12(2):117-124.PubMedGoogle ScholarCrossref
6.
Chételat  G.  Alzheimer disease: Aβ-independent processes—rethinking preclinical AD.  Nat Rev Neurol. 2013;9(3):123-124.PubMedGoogle ScholarCrossref
7.
Jack  CR  Jr, Wiste  HJ, Weigand  SD,  et al.  Amyloid-first and neurodegeneration-first profiles characterize incident amyloid PET positivity.  Neurology. 2013;81(20):1732-1740.PubMedGoogle ScholarCrossref
8.
Knopman  DS, Jack  CR  Jr, Wiste  HJ,  et al.  Brain injury biomarkers are not dependent on β-amyloid in normal elderly.  Ann Neurol. 2013;73(4):472-480.PubMedGoogle ScholarCrossref
9.
Wirth  M, Villeneuve  S, Haase  CM,  et al.  Associations between Alzheimer disease biomarkers, neurodegeneration, and cognition in cognitively normal older people.  JAMA Neurol. 2013;70(12):1512-1519.PubMedGoogle Scholar
10.
Crary  JF, Trojanowski  JQ, Schneider  JA,  et al.  Primary age-related tauopathy (PART): a common pathology associated with human aging.  Acta Neuropathol. 2014;128(6):755-766.PubMedGoogle ScholarCrossref
11.
Jack  CR  Jr.  PART and SNAP.  Acta Neuropathol. 2014;128(6):773-776.PubMedGoogle ScholarCrossref
12.
Jellinger  KA, Alafuzoff  I, Attems  J,  et al.  PART, a distinct tauopathy, different from classical sporadic Alzheimer disease.  Acta Neuropathol. 2015;129(5):757-762.PubMedGoogle ScholarCrossref
13.
Braak  H, Braak  E.  Frequency of stages of Alzheimer-related lesions in different age categories.  Neurobiol Aging. 1997;18(4):351-357.PubMedGoogle ScholarCrossref
14.
Nelson  PT, Alafuzoff  I, Bigio  EH,  et al.  Correlation of Alzheimer disease neuropathologic changes with cognitive status: a review of the literature.  J Neuropathol Exp Neurol. 2012;71(5):362-381.PubMedGoogle ScholarCrossref
15.
Duyckaerts  C, Braak  H, Brion  JP,  et al.  PART is part of Alzheimer disease.  Acta Neuropathol. 2015;129(5):749-756.PubMedGoogle ScholarCrossref
16.
Chien  DT, Bahri  S, Szardenings  AK,  et al.  Early clinical PET imaging results with the novel PHF-tau radioligand [F-18]-T807.  J Alzheimers Dis. 2013;34(2):457-468.PubMedGoogle Scholar
17.
Johnson  KA, Schultz  A, Betensky  RA,  et al.  Tau positron emission tomographic imaging in aging and early Alzheimer disease.  Ann Neurol. 2016;79(1):110-119.PubMedGoogle ScholarCrossref
18.
Knopman  DS, Jack  CR  Jr, Wiste  HJ,  et al.  Short-term clinical outcomes for stages of NIA-AA preclinical Alzheimer disease.  Neurology. 2012;78(20):1576-1582.PubMedGoogle ScholarCrossref
19.
Mormino  EC, Betensky  RA, Hedden  T,  et al.  Synergistic effect of β-amyloid and neurodegeneration on cognitive decline in clinically normal individuals.  JAMA Neurol. 2014;71(11):1379-1385.PubMedGoogle ScholarCrossref
20.
Vos  SJ, Xiong  C, Visser  PJ,  et al.  Preclinical Alzheimer’s disease and its outcome: a longitudinal cohort study.  Lancet Neurol. 2013;12(10):957-965.PubMedGoogle ScholarCrossref
21.
Morris  JC.  The Clinical Dementia Rating (CDR): current version and scoring rules.  Neurology. 1993;43(11):2412-2414.PubMedGoogle ScholarCrossref
22.
Wechsler  D.  Wechsler Memory Scale–Revised. San Antonio, TX: The Psychological Corp; 1987.
23.
Folstein  MF, Folstein  SE, McHugh  PR.  “Mini-mental state”: a practical method for grading the cognitive state of patients for the clinician.  J Psychiatr Res. 1975;12(3):189-198.PubMedGoogle ScholarCrossref
24.
Fischl  B, Salat  DH, Busa  E,  et al.  Whole brain segmentation: automated labeling of neuroanatomical structures in the human brain.  Neuron. 2002;33(3):341-355.PubMedGoogle ScholarCrossref
25.
Landau  SM, Harvey  D, Madison  CM,  et al; Alzheimer’s Disease Neuroimaging Initiative.  Associations between cognitive, functional, and FDG-PET measures of decline in AD and MCI.  Neurobiol Aging. 2011;32(7):1207-1218.PubMedGoogle ScholarCrossref
26.
Mormino  EC, Betensky  RA, Hedden  T,  et al; Alzheimer’s Disease Neuroimaging Initiative; Australian Imaging Biomarkers and Lifestyle Flagship Study of Ageing; Harvard Aging Brain Study.  Amyloid and APOE ε4 interact to influence short-term decline in preclinical Alzheimer disease.  Neurology. 2014;82(20):1760-1767.PubMedGoogle ScholarCrossref
27.
Amariglio  RE, Mormino  EC, Pietras  AC,  et al.  Subjective cognitive concerns, amyloid-β, and neurodegeneration in clinically normal elderly.  Neurology. 2015;85(1):56-62.PubMedGoogle ScholarCrossref
28.
Papp  KV, Amariglio  RE, Mormino  EC,  et al.  Free and cued memory in relation to biomarker-defined abnormalities in clinically normal older adults and those at risk for Alzheimer’s disease.  Neuropsychologia. 2015;73:169-175.PubMedGoogle ScholarCrossref
29.
Donohue  MC, Sperling  RA, Salmon  DP,  et al; Australian Imaging, Biomarkers, and Lifestyle Flagship Study of Ageing; Alzheimer’s Disease Neuroimaging Initiative; Alzheimer’s Disease Cooperative Study.  The preclinical Alzheimer Cognitive Composite: measuring amyloid-related decline.  JAMA Neurol. 2014;71(8):961-970.PubMedGoogle ScholarCrossref
30.
Dickerson  BC, Bakkour  A, Salat  DH,  et al.  The cortical signature of Alzheimer’s disease: regionally specific cortical thinning relates to symptom severity in very mild to mild AD dementia and is detectable in asymptomatic amyloid-positive individuals.  Cereb Cortex. 2009;19(3):497-510.PubMedGoogle ScholarCrossref
31.
Rowe  CC, Ellis  KA, Rimajova  M,  et al.  Amyloid imaging results from the Australian Imaging, Biomarkers and Lifestyle (AIBL) study of aging.  Neurobiol Aging. 2010;31(8):1275-1283.PubMedGoogle ScholarCrossref
32.
Dawe  RJ, Bennett  DA, Schneider  JA, Arfanakis  K.  Neuropathologic correlates of hippocampal atrophy in the elderly: a clinical, pathologic, postmortem MRI study.  PLoS One. 2011;6(10):e26286.PubMedGoogle ScholarCrossref
33.
Josephs  KA, Whitwell  JL, Tosakulwong  N,  et al.  TAR DNA-binding protein 43 and pathological subtype of Alzheimer’s disease impact clinical features.  Ann Neurol. 2015;78(5):697-709.PubMedGoogle ScholarCrossref
34.
Raz  N, Gunning-Dixon  F, Head  D, Rodrigue  KM, Williamson  A, Acker  JD.  Aging, sexual dimorphism, and hemispheric asymmetry of the cerebral cortex: replicability of regional differences in volume.  Neurobiol Aging. 2004;25(3):377-396.PubMedGoogle ScholarCrossref
35.
Knopman  DS, Jack  CR  Jr, Wiste  HJ,  et al.  18F-fluorodeoxyglucose positron emission tomography, aging, and apolipoprotein E genotype in cognitively normal persons.  Neurobiol Aging. 2014;35(9):2096-2106.PubMedGoogle ScholarCrossref
36.
Jack  CR  Jr, Dickson  DW, Parisi  JE,  et al.  Antemortem MRI findings correlate with hippocampal neuropathology in typical aging and dementia.  Neurology. 2002;58(5):750-757.PubMedGoogle ScholarCrossref
37.
Thal  DR, Beach  TG, Zanette  M,  et al.  [(18)F]flutemetamol amyloid positron emission tomography in preclinical and symptomatic Alzheimer’s disease: specific detection of advanced phases of amyloid-β pathology.  Alzheimers Dement. 2015;11(8):975-985.PubMedGoogle ScholarCrossref
×