VIEWS & REVIEWS

A/T/N: An unbiased descriptive

classification scheme for Alzheimer disease
biomarkers

Clifford R. Jack, Jr., MD ABSTRACT

David A. Bennett, MD Biomarkers have become an essential component of Alzheimer disease (AD) research and
Kaj Blennow, MD, PhD because of the pervasiveness of AD pathology in the elderly, the same biomarkers are used in
Maria C. Carrillo, PhD cognitive aging research. A number of current issues suggest that an unbiased descriptive clas-
Howard H. Feldman, sification scheme for these biomarkers would be useful. We propose the A/T/N system in which 7
MD major AD biomarkers are divided into 3 binary categories based on the nature of the pathophys-
Giovanni B. Frisoni, MD iology that each measures. A refers to the value of a b-amyloid biomarker (amyloid PET or CSF
Harald Hampel, MD, Ab42); T, the value of a tau biomarker (CSF phospho tau, or tau PET); and N, biomarkers of
PhD neurodegeneration or neuronal injury ([18F]-fluorodeoxyglucosePET, structural MRI, or CSF total
William J. Jagust, MD tau). Each biomarker category is rated as positive or negative. An individual score might appear as
Keith A. Johnson, MD A1/T1/N2, or A1/T2/N2, etc. The A/T/N system includes the new modality tau PET. It is agnos-
David S. Knopman, MD tic to the temporal ordering of mechanisms underlying AD pathogenesis. It includes all individuals
Ronald C. Petersen, MD, in any population regardless of the mix of biomarker findings and therefore is suited to population
PhD studies of cognitive aging. It does not specify disease labels and thus is not a diagnostic classi-
Philip Scheltens, MD, fication system. It is a descriptive system for categorizing multidomain biomarker findings at the
PhD individual person level in a format that is easy to understand and use. Given the present lack of
Reisa A. Sperling, MD consensus among AD specialists on terminology across the clinically normal to dementia spec-
Bruno Dubois, MD, PhD trum, a biomarker classification scheme will have broadest acceptance if it is independent from
any one clinically defined diagnostic scheme. Neurology 2016;87:539547

Correspondence to
Dr. Jack: GLOSSARY
[email protected] Ab 5 b-amyloid; AD 5 Alzheimer disease; FDG 5 [18F]-fluorodeoxyglucose; IWG 5 International Working Group; MCI 5 mild
cognitive impairment; NIA-AA 5 National Institute on AgingAlzheimers Association; p-tau 5 phosphorylated tau; SNAP 5
suspected non-Alzheimer pathophysiology; t-tau 5 total tau.

By providing measures of relevant pathophysiology in living persons, biomarkers have become

increasingly important to understanding the biology of Alzheimer disease (AD). Biomarkers are
also used in all modern research diagnostic criteria across the AD clinical spectrum,16 in thera-
peutic trials, and by regulatory agencies.7 Because AD pathology is so frequent in the elderly, AD
biomarkers are also commonly used in cognitive aging research. However, several current issues
suggest that a different approach to biomarkers used in AD research might be useful.
First, available evidence points to recently developed tau PET tracers as useful measures of
neurofibrillary tangles in AD, while utility in non-AD tauopathies has not yet been clarified.8,9
Elevated tau PET tracer signal, particularly in neocortical regions, is highly associated with the
Editorial, page 456
From the Departments of Radiology (C.R.J.) and Neurology (D.S.K., R.C.P.), Mayo Clinic, Rochester, MN; Rush Alzheimers Disease Center (D.A.B.), Rush
University Medical Center, Chicago, IL; Clinical Neurochemistry Lab (K.B.), Department of Neuroscience and Physiology, University of Gothenburg,
Mlndal Hospital, Sahlgrenska University Hospital, Mlndal, Sweden; Alzheimers Association (M.C.C.), Chicago, IL; Division of Neurology (H.H.F.), UBC
Hospital Clinic for Alzheimers Disease and Related Disorders, University of British Columbia, Vancouver, Canada; Memory Clinic (G.B.F.), University
Hospitals and University of Geneva, Switzerland; IRCCS Fatebenefratelli (G.B.F.), The National Centre for Alzheimers Disease, Brescia, Italy; Sorbonne
Universits (H.H.), Universit Pierre et Marie Curie, Paris; Institut de la Mmoire et de la Maladie dAlzheimer (IM2A) and Institut du Cerveau et de la Moelle
pinire (ICM) (H.H.), Dpartement de Neurologie, Hpital de la Piti-Salptrire, Paris, France; Helen Wills Neuroscience Institute (W.J.J.), University of
California, Berkeley; Departments of Radiology and Neurology (K.A.J.), Massachusetts General Hospital, Harvard Medical School, Boston; Alzheimer Center
and Department of Neurology (P.S.), Vrije Universiteit Amsterdam, the Netherlands; Center for Alzheimer Research and Treatment (R.A.S.), Brigham and
Womens Hospital and Massachusetts General Hospital, Harvard Medical School, Boston; Centre des Maladies Cognitives et Comportementales (B.D.),
Institut du Cerveau et de la Moelle pinire, Paris; and Universit Pierre et Marie Curie-Paris 6 (B.D.), AP-HP, Hpital de la Salptrire, Paris, France.
Go to Neurology.org for full disclosures. Funding information and disclosures deemed relevant by the authors, if any, are provided at the end of the article.
The Article Processing Charge was paid by the authors.
This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC
BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used
commercially.

2016 American Academy of Neurology 539

2016 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 presence of positive amyloid PET scans as scheme we propose is applicable across all clinical
would be expected in a ligand that binds to diagnostic states, and is thus independent of cog-
the tau deposits in AD.10,11 Tau PET ligand nitive status.
binding correlates well with clinical impair- Fourth, the prevalence of AD, non-AD,
ment in individuals who lie along the AD clin- and mixed brain pathologies increases with
ical spectrum.10,11 When easily characterized age in both individuals who are clinically
off-target and nonspecific binding is ac- impaired and those nonimpaired.20 To be an
counted for, the topographic patterns of ligand effective tool in cognitive aging research, a bio-
uptake match quite well what is expected from marker classification system must include all
Braak staging of neurofibrillary tangles.12 possible biomarker profiles and all individuals
However, because of its recent introduction, in the population. For example, classification
tau PET is not yet integrated into any current schemes that require Ab positivity do not clas-
AD diagnostic schemes. sify individuals who are Ab negative but pos-
Second, many details of AD pathogenesis itive on tau or neurodegenerative biomarkers,
remain uncertain.13,14 One of the most conten- yet the latter biomarker profile is common.21
tious issues, and one of the oldest, is which The objective of this report is to propose an
proteinopathy causes the disease in the elderly. unbiased descriptive classification scheme for
Some propose that AD pathogenesis follows biomarkers commonly used in AD research
a specific cause and effect order of events, where that addresses each of these 4 issues.
b-amyloidosis potentiates the spread of tauop-
athy, tauopathy is associated with neurodegen- AD BIOMARKERS We propose that the major bio-
eration, which is the immediate cause of clinical markers used in AD research can be divided into 3
symptoms.15 Others argue for different, less binary categories based on the nature of the underly-
b-amyloid (Ab) centric pathways to clinically ing pathophysiology each measures. Biomarkers of
fibrillary Ab deposition are high ligand retention
symptomatic AD.13,14,16 Interdependent path-
on amyloid PET22 or low CSF Ab42.2325 Biomarkers
ways have also been proposed.17 A biomarker
of tau pathology (neurofibrillary tangles) are elevated
classification system will have the broadest use CSF phosphorylated tau (p-tau) and tau PET.24,26
if it makes no assumptions about temporal Biomarkers of AD-like neurodegeneration or
ordering of biomarkers or their putative causal neuronal injury are CSF total tau (t-tau), [18F]-
relationships. fluorodeoxyglucose (FDG)-PET hypometabolism,
Third, current biomarker classification sys- and atrophy on structural MRI in regions characteristic
tems are linked to disease or syndromic labels of AD.27
and are based on consensus diagnostic criteria CSF biomarkers report a single absolute value re-
rather than certainty of disease pathogenesis. flecting degree of abnormality but that does not indi-
cate topographic extent of pathology. In contrast,
The 2 major such diagnostic schema are those
imaging biomarkers contain information about both
of the International Working Group (IWG),1,18
the severity and topographic extent of the abnormal-
which first proposed the use of a common bio- ity. Typical AD (i.e., amnestic multidomain demen-
marker algorithm for all clinical stages of the tia) is associated with a characteristic topographic
disease, and the National Institute on Aging pattern of FDG-PET hypometabolism that appears
Alzheimers Association (NIA-AA).36 Although earliest and most severely in the medial parietal and
there are areas of agreement, important disagree- lateral temporal-parietal isocortex (figure 1A).28 Typ-
ments exist concerning staging, nomenclature, ical AD likewise is associated with a pattern of atro-
and interpretations of biomarker findings. We phy on MRI that appears earliest and most severely in
recognize that associating biomarkers with clini- the medial temporal allocortex and the basal-lateral
temporal isocortex (figure 1C).15 Ab deposition on
cal findings is important. However, given the
PET does not seem to follow a sequential topographic
confusion that has arisen over the use of com-
progression through the isocortex (figure 1E). Avail-
peting definitions of clinical impairment in able evidence indicates that in the typical AD spec-
the AD spectrum,19 a biomarker classification trum, tau PET captures the topography of tau spread
scheme will have broadest acceptance if it is inde- as described at autopsy by Braak12medial temporal
pendent from any one clinically operationalized to basal and lateral temporal, then to other isocortical
diagnostic scheme. The biomarker classification areas (figure 2).

540 Neurology 87 August 2, 2016

2016 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 abnormal) in some non-AD conditions such as HIV
Figure 1 Images of clinically normal individuals and participants with AD
encephalitis33 and multiple system atrophy.34
Typical AD shows an increase in both t-tau and
p-tau and a tight correlation between these bio-
markers.35 Neither CSF t-tau nor p-tau shows any
change in primary tauopathies such as frontotemporal
dementia, progressive supranuclear palsy, or cortico-
basal degeneration.36,37 However, findings on CSF
t-tau and p-tau diverge in conditions with acute brain
damage. There is a marked temporary increase in t-tau
with normal p-tau levels in traumatic brain injury and
stroke that correlates with the severity of neuronal dam-
age.38,39 It is the same for Creutzfeldt-Jakob disease,
which shows a large increase in t-tau (reflecting rapid
neurodegeneration) but normal p-tau since there are no
neurofibrillary tangles in this condition.40,41 The CSF
level of p-tau correlates with severity of tau pathology
postmortem,30 and high p-tau has not been found in
disorders other than AD. Taken together, these data
indicate that CSF t-tau reflects the intensity of neuronal
degeneration in AD at a specific point, while p-tau
seems more specific for the burden of AD-type tau
pathology accumulated over time.42 Tau PET is being
actively investigated, but initial data indicate that tau
PET ligands have high binding affinity for paired heli-
cal filament tau in AD but have much weaker affinity
in non-AD tauopathies, especially those with straight
filament tangles.9
Atrophy and hypometabolism involving AD-like re-
gions occur in a variety of disorders and are the least
specific for AD.43 Atrophy in the anterior/medial/basal
temporal lobes occurs in a wide variety of pathologic
conditions including AD but also cerebrovascular
disease, epilepsy, anoxia, hippocampal sclerosis, TDP-
43-opathy, primary age-related tauopathy,44 chronic
traumatic encephalopathy, argyrophilic grain disease,
and non-AD primary tauopathies such as progressive
Individuals with AD dementia are clinically diagnosed participants in Mayo Alzheimers Dis- supranuclear palsy and Pick disease.43 Temporoparietal
ease Research Center study while clinically normal individuals are participants in the Mayo hypometabolism can be found in non-AD conditions,
Clinic Study of Aging. (A) FDG-PET of 75-year-old man with AD dementia. Hypometabolism
such as corticobasal degeneration, primary progressive
in medial parietal and lateral temporal-parietal isocortex with relative preservation of fron-
tal metabolism, which is characteristic of typical (multidomain amnestic) AD. (B) FDG-PET of aphasia,45 and cerebrovascular disease.46 This nonspeci-
clinically normal 71-year-old man. Uniform FDG uptake is present throughout the isocortex. ficity is the explanation for the frequent and consistently
(C) MRI of 71-year-old man with AD dementia. Atrophy is present in the medial temporal observed finding of abnormal FDG-PET and structural
allocortex and the basal-lateral temporal isocortex, which is characteristic of typical (mul-
MRI (and CSF t-tau) in non-AD conditions
tidomain amnestic) AD. (D) MRI of clinically normal 71-year-old woman without atrophy. (E)
Amyloid PET with Pittsburgh compound B of 71-year-old woman with AD dementia. Ligand a state that has been labeled suspected non-
uptake is seen throughout the isocortex. (F) Amyloid PET of clinically normal 93-year-old Alzheimer pathophysiology (SNAP).21
man showing no ligand uptake in the isocortex. AD 5 Alzheimer disease; FDG 5 [18F]- Interpreting biomarker data is confounded by the
fluorodeoxyglucose.
common coexistence of AD pathology, and therefore
positive AD biomarkers, with other age-related
While each of the 7 commonly used core bio- pathologies. Multidomain amnestic dementia and
markers is associated with AD, they are not equally mild cognitive impairment (MCI) are most com-
specific. Biomarkers of b-amyloidosis are specific monly associated with multiple pathology,20 espe-
for AD pathology.2931 However, amyloid PET trac- cially with advancing age.
ers do bind to Ab deposits in vessel walls, and
increased tracer binding can be found following acute THE A/T/N CLASSIFICATION SYSTEM In the
traumatic brain injury.32 CSF Ab42 is decreased (i.e., proposed A/T/N classification system, the 7 major

Neurology 87 August 2, 2016 541

2016 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 biomarkers are strongly associated with worse cogni-
Figure 2 Tau and amyloid PET imaging in AD
tive symptoms throughout the clinical spectrum,
which is not the case for amyloid biomarkers.15 The
topography of hypometabolism and atrophy maps
well onto the expression of clinical symptoms, which
is not the case for amyloid PET.47
We recognize that having both T and N categories
adds complexity in comparison to simply grouping
tau PET, FDG-PET, MRI atrophy, CSF t-tau, and
p-tau into a single catch-all N category. However,
the simpler approach would fail to take advantage
of information that is available. Some (perhaps quite
a bit) of the neuronal injury/neurodegeneration pres-
ent in elderly individuals is related to non-AD etiolo-
gies. Separating biomarkers of neurofibrillary tangles
(tau PET and CSF p-tau) from markers of neuronal
injury/neurodegeneration might differentiate neuro-
nal injury/neurodegeneration that is attributable to
AD from non-AD causes.
In the A/T/N system, each biomarker is rated as pos-
itive or negative. An individual score might appear as
A1/T1/N2, or A1/T2/N2, etc. In the event that
a biomarker class was unavailable, it would be denoted
u. Conflicting results within a category would be
labeled c. For example, if an individual had conflicting
results from amyloid PET and CSF Ab, he or she might
A 79-year-old man with a clinical diagnosis of AD dementia. He is a participant in the Mayo be labeled Ac. While we regard atrophy, hypometabo-
Alzheimers Disease Research Center study. (A, B) Coronal and axial tau PET images
(AV1451) superimposed on MRI. (C, D) Coronal and axial Pittsburgh compound B PET images
lism, and total tau each as exemplars of neuronal injury/
superimposed on MRI. The tau PET images (top) illustrate extensive tracer uptake in basal neurodegeneration, we do not believe these 3 measures
lateral temporal, parietal, and frontal isocortex with sparing of sensory motor and primary should be equated. Analyses show that neurodegenera-
visual cortices. Off-target binding is seen in the basal ganglia, which is characteristic of this
tion biomarkers are only modestly correlated with one
tracer. Although areas of spatial overlap between the tau and amyloid tracers are present,
abundant amyloid tracer uptake is seen in the frontal lobes, but not with the tau tracer. another.48 Therefore, if possible, within a given research
Conversely, abundant uptake is seen in the medial temporal lobes with the tau ligand but study, the N category should be described by only one
not with the amyloid ligand. AD 5 Alzheimer disease. N marker, not either/or mixtures of the 3. Individuals
can be fully classified by CSF alone, imaging alone, or
AD biomarkers are divided into 3 binary classes. A combinations.
refers to the value of an Ab biomarker (amyloid PET
or CSF Ab42); T, the value of a tau pathology bio- POSITIVE/NEGATIVE OR NORMAL/ABNORMAL
marker (CSF p-tau or tau PET); and N, a quantita- BIOMARKER CUTPOINTS Although every bio-
tive or topographic biomarker of neurodegeneration or marker exists on a continuous scale, normal vs abnor-
neuronal injury (CSF t-tau, FDG-PET, or structural mal cutpoints exist in most disease categories to make
MRI). diagnostic categorization of individuals practical and
The A/T/N classification system is related to the inform clinical decision-making. We recognize that
biomarker classification proposed in recent consensus cutpoints can be arbitrary, and many individuals will
diagnostic criteria. In both IWG and NIA-AA diag- have biomarker values close to cutpoints. This is true
nostic criteria, A refers to Ab (PET or CSF Ab). for any disease and is not unique to biomarkers used
Segregation of MRI and FDG-PET from CSF tau in AD research. The presence of positive/negative
biomarkers was proposed by the IWG (2014),1while cutpoints is not inherently problematic provided that
in the NIA-AA criteria, MRI, FDG-PET, and CSF values close to cutpoints are interpreted and used
tau proteins were grouped together as biomarkers of properly.
neurodegeneration or neuronal injury. The rationale Amyloid biomarkers have been bimodally distrib-
for grouping CSF tau, FDG-PET, and MRI atrophy uted in some research samples where participants are
into a single category in the NIA-AA criteria had highly selected,49 while neurodegenerative biomarkers
strong support from numerous observations that the typically are not. This has led to the suggestion
3 behaved in a similar manner relative to clinical that cutpoints are not valid for neurodegenerative
symptoms. More abnormal values in all 3 of these biomarkers. However, most physiologic measures

542 Neurology 87 August 2, 2016

2016 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 are continuously, not bimodally, distributedblood a continuous or semicontinuous scale. An example of
pressure, for example.50 This has not prevented med- this, termed the centiloid scale,56 has been proposed for
icine from identifying a cutpoint and designating amyloid PET. The process requires empirically estab-
those above hypertensive and those below normoten- lishing reference values for the abnormal part of the
sive. It seems illogical to treat AD differently from distribution using individuals with AD dementia and
other diseases. for the normal part of the distribution using young,
We emphasize that binary, 1/2 categorization clinically normal individuals. Biomarker values are then
does not imply that individuals who fall below the scaled linearly from 0 (normal) to 100 (abnormal). The
cutpoint for a particular biomarker have no pathology centiloid scale is related to but not identical to percen-
in the brain. For example, an individual designated as tiles because centiloid values below 0 (i.e., below the
A may well have amyloid plaques in the brain, but not mean for the normal group) and above 100 (i.e., above
at a sufficient level to cross the in vivo detection thresh- the mean for the abnormal group) are possible. An
old of amyloid PET. Current biomarker readouts are example of how this might appear is A80/T50/N20,
not sensitive to low but perhaps biologically important where this individual ranks at the 80th centiloid for
levels of early pathology.51,52 The 1/2 designation is amyloid, 50th for tau, and 20th for neurodegeneration.
a convenient shorthand to facilitate communication, This approach would require a 0 to 100 scale that was
ease of use, and understanding. standardized across all biomarkers.
Several different approaches to selecting normal/ A second alternative to binary, 1/2 biomarker
abnormal cutpoints in the continuous range of bio- scoring is topographic staging. This approach would
marker values exist. These include selecting values only be applicable to imaging and might best apply to
that best separate clinically normal individuals from tau PET. Topographic tau PET staging would mirror
those with dementia, values that have predictive pathologic Braak neurofibrillary tangle stage whereby
power for future clinical decline,53 or using autopsied individuals would be assigned to 1 of 3 stages based
individuals with antemortem biomarker studies to on anatomical locations of tracer uptakei.e., stage
guide selection of cutpoints.51,52 In laboratory medi- 0, limbic stage, isocortical stage. Since cortical atro-
cine, the 95th percentile based on a healthy control phy on MRI closely mirrors Braak stage and tau den-
population is commonly used.54 A popular approach sity in imagingautopsy correlation studies,57 and the
has been to select cutpoints based on the (most nor- topographic spread of atrophy within individuals over
mal) 10th percentile of values seen in typical AD time also mirrors progressive Braak stages, this Braak-
dementia.21,55 Validating specific cutpoints will be like topographic staging approach might be valid for
an ongoing exercise for research groups working in MRI as well.
each separate modality. Generalizability of AD biomarker categorization is
Experience labeling tau PET scans positive vs neg- dependent on standardization and reproducibility of
ative is limited at this point. Tauopathy confined to the measures.58
the medial temporal lobes with minimal or no neo-
cortical b-amyloidosis is very common at autopsy ATYPICAL AD, CEREBROVASCULAR DISEASE, AND
and whether this should be regarded as AD or simply EXPANSION OF THE CORE A/T/N SYSTEM The de-
an aging phenomenon is controversial.12,13,44 One scriptions of FDG-PET, tau PET, and MRI topography
possible approach to categorizing tau PET is to label outlined above reflect patterns of typical multidomain
scans that have tracer uptake exceeding an analytically amnestic AD (figures 1 and 2). Because MRI, FDG,
determined threshold in AD-like isocortical areas as and tau PET topographic patterns map onto clinical
AD positive (figure 2). Where tau is located and its phenotype, a different set of imaging signatures is
topographic progression over time will be important. needed to describe atypical variants of AD. Although
But for purposes of labeling a scan abnormal, we pro- not detailed in the present report, the A/T/N system is
visionally propose that a scan with tracer uptake con- equally applicable to atypical AD variants by modifying
fined to the medial temporal lobes would not be the topographic search pattern. For example, pattern
considered positive44 but a scan with uptake in AD- recognition algorithms could easily be modified to
like isocortical areas would.10 recognize atrophy, hypometabolism, and tau PET
deposition in the frontal, dominant temporal lobe, or
ALTERNATIVES TO POSITIVE/NEGATIVE SCORING posterior cortical regions for atypical AD.
OF BIOMARKERS Advantages of binary, 1/2 catego- Our primary focus is the core A/T/N system,
rization of each biomarker class include economy, con- which addresses biomarkers of AD; however, bio-
ceptual clarity, and ease of use. Thus, we recommend markers of other proteinopathies could be added if/
binary, 1/2 categorization. Alternatives exist however when they become available. The category of synaptic
and are briefly discussed for completeness. One alterna- dysfunction (S) may be a useful future addition and
tive is to score the severity of each biomarker on this might include FDG-PET, task-free functional

Neurology 87 August 2, 2016 543

2016 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 APPLICATION OF THE A/T/N SYSTEM IN
Table 1 Clinically normal individuals
COGNITIVE AGING AND DEMENTIA RESEARCH:
A/T/N/C The A/T/N system provides a common
NIA-AA classification
A/T/N classification preclinical AD 2014 IWG classification framework by which investigators can describe and
A2/T2/N2 Not defined Not defined communicate multidomain biomarker profiles at
A1/T2/N2 Stage 1 Asymptomatic at risk of AD (if A1
the individual person level. Application in cognitive
established by amyloid PET) aging and dementia research, however, will require
A1/T1/N2 Stage 2/3 Asymptomatic at risk of AD the inclusion of clinical information about each indi-
A1/T2/N1 a
Asymptomatic at risk of AD (if A1
vidual. Clinical status (C) could be denoted in several
established by amyloid PET) possible ways resulting in A/T/N/C notation. Setting
A1/T1/N1 Stage 2/3 Asymptomatic at risk of AD aside detailed syndromic descriptions, cognitive
A2/T1/N2 b
Not defined Not defined function can be thought of categorically or as a con-
A2/T2/N1 b
Not defined Not defined
tinuum. The cognitive/functional continuum could
be divided into normal for age (n), mildly impaired
A2/T1/N1b Not defined Not defined
(m), or demented (d), resulting in designations of
Abbreviations: AD 5 Alzheimer disease; FDG 5 [18F]-fluorodeoxyglucose; IWG 5 International A/T/N/Cn, A/T/N/Cm, or A/T/N/Cd. The cogni-
Working Group; NIA-AA 5 National Institute on AgingAlzheimers Association.
a
tive continuum could also be expressed as a contin-
This combination was not addressed in NIA-AA preclinical AD criteria on the assumption that
neurodegeneration on MRI and FDG-PET that is specifically attributable to AD was tau-related. uous variable on a 0 (normal) to 100 (abnormal)
b
Described as SNAP (suspected non-Alzheimer pathophysiology) in several publications. scale; individuals are assigned values along this 0 to
100 scale.
MRI, EEG, MEG (magnetoencephalography), as
well as synapse-specific proteins in CSF.59 However, HOW DOES THE A/T/N SYSTEM RELATE TO
EXISTING AD CLINICAL CLASSIFICATION
if neurodegeneration is defined as progressive loss and
SYSTEMS? The IWG and NIA-AA criteria both
shrinkage of neurons and processes with a correspond-
integrate 5 biomarkers into the diagnostic classification
ing impairment in neuronal function,15 then synaptic
process: CSF Ab42 and tau proteins, amyloid PET,
dysfunction is subsumed within the category of
FDG-PET, and MRI. In the most recent version of
neurodegeneration.
the IWG criteria,1 CSF Ab, tau, and amyloid PET are
The core A/T/N system could also be supple-
regarded as pathophysiologic biomarkers of AD while
mented by adding a cerebrovascular disease category.
FDG-PET and MRI are considered topographic
A limitation is the absence of an agreed on pathologic
downstream biomarkers. Biomarker support for AD
summary scoring system for cerebrovascular disease
pathophysiology consists of a positive amyloid PET
on which to base an imaging counterpart. Nonethe-
scan or both depressed Ab42 and elevated t-tau or p-tau
less, systems in which the MRI findings of ischemic
in CSF. In the NIA-AA criteria (2011), separate
cerebrovascular disease are combined to form a vascu-
guidelines were outlined for 3 clinical phases
lar (V) summary score have been described.60 The A/
preclinical, MCI, and AD dementia.36 The NIA-AA
T/N system would then be extended to A/T/N/V by
preclinical AD criteria were predicated on the concept
compressing the vascular index into V1 or V2.
that AD biomarkers follow a specific temporal
ordering where b-amyloidosis occurs before tau-related
Table 2 Individuals who meet clinical criteria for MCI
neurodegeneration, which in turn is the proximate
correlate of clinical symptoms. By contrast, NIA-AA
A/T/N score NIA-AA classification 2014 IWG classification criteria for MCI and AD dementia attribute equal
A2/T2/N2 MCI, unlikely due to AD Not defined diagnostic weight to all AD biomarkers included in the
A1/T2/N2 MCI, core clinical criteriaa
Typical AD (if A1 established
criteria.
by amyloid PET) Tables 1 to 3 outline how the A/T/N system maps
A1/T1/N2 MCI, core clinical criteriaa Typical AD onto these 2 existing diagnostic classification systems.
A1/T2/N1 MCI, core clinical criteriaa Typical AD (if A1 established Conceptual differences both between and within these
by amyloid PET) sets of diagnostic criteria are evident. For example,
A1/T1/N1 MCI due to AD, high likelihood Typical AD a clinically asymptomatic individual with a positive
A2/T1/N2 b
Not defined Not defined CSF Ab42 and negative CSF tau profile is labeled pre-
A2/T2/N1b Not defined Not defined
clinical AD stage 1 by the NIA-AA criteria but is clas-
b
sified as not being in the AD pathway by IWG. A
A2/T1/N1 Not defined Not defined
clinically asymptomatic individual with positive CSF
Abbreviations: AD 5 Alzheimer disease; IWG 5 International Working Group; MCI 5 mild Ab42 and tau is labeled preclinical AD stage 2 by the
cognitive impairment; NIA-AA 5 National Institute on AgingAlzheimers Association.
a
NIA-AA criteria and as asymptomatic at risk for AD
In the event of conflicting results, biomarkers are regarded as uninformative and therefore
do not alter the individuals diagnostic classification based on clinical assessment alone. by IWG. Disagreement in how these different diagnos-
b
Described as MCI-SNAP (suspected non-Alzheimer pathophysiology) in several publications. tic criteria treat biomarkers creates uncertainty for

544 Neurology 87 August 2, 2016

2016 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 profiles will emerge as the A/T/N system is applied
Table 3 Individuals who meet clinical criteria for probable AD dementia
empirically.
A/T/N score NIA-AA classification 2014 IWG classification
AUTHOR CONTRIBUTIONS
A2/T2/N2 Dementia, unlikely due to AD Not defined
Dr. Jack: study concept and design and critical revision of the manuscript for
A1/T2/N2 Intermediate likelihood; probable AD Typical AD (if A1 established important intellectual content. Dr. Bennett: critical revision of the manuscript
dementia; based on clinical criteriaa by amyloid PET) for important intellectual content. Dr. Blennow: critical revision of the man-
A1/T1/N2 High likelihood probable AD dementia; Typical AD
uscript for important intellectual content. Dr. Carrillo: critical revision of the
based on clinical criteriaa manuscript for important intellectual content. Dr. Feldman: critical revision
of the manuscript for important intellectual content. Dr. Frisoni: critical revi-
A1/T2/N1 High likelihood; probable AD dementia; Typical AD (if A1 established
based on clinical criteriaa by amyloid PET) sion of the manuscript for important intellectual content. Dr. Hampel: critical
revision of the manuscript for important intellectual content. Dr. Jagust:
A1/T1/N1 High likelihood AD pathophysiology Typical AD critical revision of the manuscript for important intellectual content.
A2/T1/N2 Probable AD dementia; based on Not defined Dr. Johnson: critical revision of the manuscript for important intellectual con-
clinical criteriaa tent. Dr. Knopman: critical revision of the manuscript for important intellec-
tual content. Dr. Petersen: critical revision of the manuscript for important
A2/T2/N1 Intermediate likelihood; probable Not defined
AD dementia; based on clinical criteriaa intellectual content. Dr. Scheltens: critical revision of the manuscript for
important intellectual content. Dr. Sperling: critical revision of the manuscript
A2/T1/N1 Intermediate likelihood; probable Not defined for important intellectual content. Dr. Dubois: critical revision of the manu-
AD dementia; based on clinical criteriaa
script for important intellectual content.
Abbreviations: AD 5 Alzheimer disease; IWG 5 International Working Group; NIA-AA 5
National Institute on AgingAlzheimers Association. STUDY FUNDING
a
In the event of conflicting results, biomarkers are regarded as uninformative and therefore No targeted funding reported.
do not alter the individuals diagnostic classification based on clinical assessment alone.
DISCLOSURE
C. Jack has provided consulting services for Eli Lilly Co. He receives
investigators, regulators, and pharmaceutical companies research funding from the NIH (R01 AG011378, U01 HL096917,
designing trials. Thus, there is a need for a unifying U01 AG024904, RO1 AG041851, R01 AG037551, R01 AG043392,
conceptual approach to biomarkers used in AD U01 AG006786) and the Alexander Family Alzheimers Disease Research
Professorship of the Mayo Foundation. D. Bennett reports no disclosures
research.
relevant to the manuscript. K. Blennow has served as a consultant for Eli
We recognize that the A/T/N biomarker system Lilly, Novartis, Roche Diagnostics, and Sanofi-Aventis and on advisory
will almost always be used in conjunction with a com- boards for Amgen and IBL International, and has given lectures for
plementary system to classify clinical impairment. By Fujirebio Europe and Lundbeck. Dr. Blennows research team has
received grants for collaborative research projects from Eli Lilly and
taking an unbiased descriptive approach and avoiding Roche Diagnostics. M. Carrillo is a full time employee of the Alzheimers
syndromic or disease labels, the A/T/N system can be Association. H. Feldman reports no disclosures relevant to the manu-
used in any set of diagnostic criteria that exist cur- script. G. Frisoni has served on advisory boards for Lilly, BMS, Bayer,
Lundbeck, Elan, AstraZeneca, Pfizer, TauRx, Wyeth, GE, Baxter. He is
rently or are developed in the future. In addition,
a member of the editorial boards of Lancet Neurology, Aging Clinical &
the A/T/N system includes all individuals in any pop- Experimental Research, Alzheimers Diseases & Associated Disorders, and
ulation and thus is suited to cognitive aging research Neurodegenerative Diseases. He is imaging section editor of Neurobiology
where AD constitutes only part of the etiologic land- of Aging. Dr. Frisoni has received grants from Wyeth International, Lilly
International, Lundbeck Italia, GE International, Avid/Lilly, Roche,
scape. We leave it to expert panels such as the IWG
Piramal, and the Alzheimers Association. Research of industrial interest
and NIA-AA to decide what disease or syndromic la- has touched: memantine, PET amyloid ligands, diagnostic and tracking
bels (i.e., AD, not AD, SNAP, preclinical AD, at risk Alzheimers biomarkers. Lecture fees when speaking at the invitation of
of AD, MCI due to AD, prodromal AD, etc.) should a commercial sponsor (in the past 2 years): Lundbeck, Piramal, and GE.
H. Hampel serves as senior associate editor for the journal Alzheimers &
apply to individuals with varying clinical presenta- Dementia; he is a scientific advisor/consultant for Axovant, Anavex, Eli
tions who are classified into different biomarker states Lilly and Company, GE Healthcare, Cytox, Jung Diagnostics, Roche,
by this system. Biogen, Takeda-Zinfandel, Oryzon; and receives research support from
the Alzheimer Foundation (Paris), Pierre and Marie Curie University,
Finally, while we anticipate N and T will often
Pfizer, Avid; and has (partly pending) patent applications, but receives
concur, i.e., either both will be positive or both neg- no royalties. W. Jagust is a consultant to BioClinica, Banner Alzheimers
ative, this will not always be the case and informative InstituteGenentech, and Novartis. K. Johnson has consulted for Novar-
profiles will be identified that are not captured by tis, AbbVie, Janssen, GEHC, Avid/Lilly, Piramal, Roche, and Isis Phar-
maceuticals, is an investigator in clinical trials sponsored by Biogen,
either the NIA-AA or the IWG system. For example, Merck, Janssen, Eisai, Navidea, Avid/Lilly, and receives research support
an A2T2N1 profile would be expected with from the NIH, Fidelity Biosciences, and the Alzheimers Association.
pathologies such as ischemic cerebrovascular disease D. Knopman serves as deputy editor for Neurology; serves on a data
safety monitoring board for Lundbeck Pharmaceuticals and for the DIAN
or hippocampal sclerosis whereas an A2T1N1 pro-
Study; is an investigator in clinical trials sponsored by TauRX Pharma-
file would be expected with PART (primary age- ceuticals, Lilly Pharmaceuticals, and the Alzheimers Disease Cooperative
related tauopathy).44 An A1T2N1 profile might Study; and receives research support from the NIH. R. Petersen is the
indicate an individual in the earliest stage of preclin- chair of the data monitoring committee for Pfizer, Inc., and Janssen
Alzheimer Immunotherapy, consultant for Hoffmann-La Roche, Inc.,
ical AD (accounting for the A1T2 status) who also
Merck, Inc., Genentech, Inc., Biogen, Inc., and Eli Lilly & Co. He
has a non-AD pathology such as hippocampal sclero- receives royalties from Oxford University Press for Mild Cognitive Impair-
sis (accounting for the N1 status). Other informative ment. P. Scheltens has received grant support (for the institution) from

Neurology 87 August 2, 2016 545

2016 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 GE Healthcare, Danone Research, Piramal, and Merck. In the past 2 15. Jack CR Jr, Holtzman DM. Biomarker modeling of Alz-
years he has received consultancy/speaker fees (paid to the institution) heimers disease. Neuron 2013;80:13471358.
from Lilly, GE Healthcare, Novartis, Forum, Sanofi, Nutricia, Probio- 16. Chetelat G. Alzheimer disease: Abeta-independent pro-
drug, and EIP Pharma. R. Sperling has been a consultant for Janssen
cessesrethinking preclinical AD. Nat Rev Neurol
Eisai, Lundbeck, Isis, Boehringer Ingelheim, Roche, and Genentech; and
2013;9:123124.
receives research support from the Alzheimers Association, Fidelity Bio-
sciences, Janssen, BrightFocus Foundation, and the National Institute on
17. Small SA, Duff K. Linking Abeta and tau in late-onset
Aging. B. Dubois has received consulting fees from Eli Lilly and Pfizer Alzheimers disease: a dual pathway hypothesis. Neuron
grant support from industry in 2012. Go to Neurology.org for full 2008;60:534542.
disclosures. 18. Dubois B, Feldman HH, Jacova C, et al. Revising the
definition of Alzheimers disease: a new lexicon. Lancet
Received December 16, 2015. Accepted in final form March 16, 2016. Neurol 2010;9:11181127.
19. Vos SJ, Verhey F, Frolich L, et al. Prevalence and prog-
REFERENCES nosis of Alzheimers disease at the mild cognitive impair-
1. Dubois B, Feldman HH, Jacova C, et al. Advancing ment stage. Brain 2015;138:13271338.
research diagnostic criteria for Alzheimers disease: the 20. Schneider JA, Arvanitakis Z, Leurgans SE, Bennett DA. The
IWG-2 criteria. Lancet Neurol 2014;13:614629. neuropathology of probable Alzheimer disease and mild
2. Dubois B, Feldman HH, Jacova C, et al. Research criteria for cognitive impairment. Ann Neurol 2009;66:200208.
the diagnosis of Alzheimers disease: revising the NINCDS- 21. Jack CR Jr, Knopman DS, Weigand SD, et al. An opera-
ADRDA criteria. Lancet Neurol 2007;6:734746. tional approach to NIA-AA criteria for preclinical Alzheimers
3. Sperling RA, Aisen PS, Beckett LA, et al. Toward defining disease. Ann Neurol 2012;71:765775.
the preclinical stages of Alzheimers disease: recommenda- 22. Klunk WE, Engler H, Nordberg A, et al. Imaging brain
tions from the National Institute on AgingAlzheimers Asso- amyloid in Alzheimers disease with Pittsburgh com-
ciation workgroups on diagnostic guidelines for Alzheimers pound-B. Ann Neurol 2004;55:306319.
disease. Alzheimers Dement 2011;7:280292. 23. Fagan AM, Roe CM, Xiong C, Mintun MA, Morris JC,
4. Albert MS, DeKosky ST, Dickson D, et al. The diagnosis of Holtzman DM. Cerebrospinal fluid tau/beta-amyloid(42)
mild cognitive impairment due to Alzheimers disease: rec- ratio as a prediction of cognitive decline in nondemented
ommendations from the National Institute on Aging and older adults. Arch Neurol 2007;64:343349.
Alzheimers Association Workgroup. Alzheimers Dement 24. Mattsson N, Zetterberg H, Hansson O, et al. CSF bio-
2011;7:270279. markers and incipient Alzheimer disease in patients with
5. McKhann GM, Knopman DS, Chertkow H, et al. The mild cognitive impairment. JAMA 2009;302:385393.
diagnosis of dementia due to Alzheimers disease: recom- 25. Visser PJ, Verhey F, Knol DL, et al. Prevalence and prog-
mendations from the National Institute on Aging and the nostic value of CSF markers of Alzheimers disease pathol-
Alzheimers Association Workgroup. Alzheimers Dement ogy in patients with subjective cognitive impairment or mild
2011;7:263269. cognitive impairment in the DESCRIPA Study: a prospec-
6. Jack CR Jr, Albert MS, Knopman DS, et al. Introduction tive cohort study. Lancet Neurol 2009;8:619627.
to the recommendations from the National Institute on 26. Buerger K, Ewers M, Pirttila T, et al. CSF phosphorylated
AgingAlzheimers Association workgroups on diagnostic tau protein correlates with neocortical neurofibrillary
guidelines for Alzheimers disease. Alzheimers Dement pathology in Alzheimers disease. Brain 2006;129:3035
2011;7:257262. 3041.
7. Hampel H, Frank R, Broich K, et al. Biomarkers for 27. Besson FL, La Joie R, Doeuvre L, et al. Cognitive and brain
Alzheimers disease: academic, industry and regulatory per- profiles associated with current neuroimaging biomarkers of
spectives. Nat Rev Drug Discov 2010;9:560574. preclinical Alzheimers disease. J Neurosci 2015;35:10402
8. Villemagne VL, Fodero-Tavoletti MT, Masters CL, 10411.
Rowe CC. Tau imaging: early progress and future directions. 28. Jagust WJ, Landau SM, Shaw LM, et al. Relationships
Lancet Neurol 2015;14:114124. between biomarkers in aging and dementia. Neurology
9. Marquie M, Normandin MD, Vanderburg CR, et al. Vali- 2009;73:11931199.
dating novel tau positron emission tomography tracer [F-18]- 29. Strozyk D, Blennow K, White LR, Launer LJ. CSF Abeta
AV-1451 (T807) on postmortem brain tissue. Ann Neurol 42 levels correlate with amyloid-neuropathology in a pop-
2015;78:787800. ulation-based autopsy study. Neurology 2003;60:652
10. Johnson KA, Shultz A, Betensky RA, et al. Tau positron 656.
emission tomographic imaging in aging and early Alzheimers 30. Tapiola T, Alafuzoff I, Herukka SK, et al. Cerebrospinal
disease. Ann Neurol 2016;79:110119. fluid {beta}-amyloid 42 and tau proteins as biomarkers of
11. Scholl M, Lockhart SN, Schonhaut DR, et al. PET Imag- Alzheimer-type pathologic changes in the brain. Arch
ing of tau deposition in the aging human brain. Neuron Neurol 2009;66:382389.
2016;89:971982. 31. Ikonomovic MD, Klunk WE, Abrahamson EE, et al. Post-
12. Braak H, Thal DR, Ghebremedhin E, Del Tredici K. mortem correlates of in vivo PiB-PET amyloid imaging in
Stages of the pathologic process in Alzheimer disease: age a typical case of Alzheimers disease. Brain 2008;131:
categories from 1 to 100 years. J Neuropathol Exp Neurol 16301645.
2011;70:960969. 32. Hong YT, Veenith T, Dewar D, et al. Amyloid imaging
13. Duyckaerts C, Braak H, Brion JP, et al. PART is part of with carbon 11-labeled Pittsburgh compound B for trau-
Alzheimer disease. Acta Neuropathol 2015;129:749756. matic brain injury. JAMA Neurol 2014;71:2331.
14. Duyckaerts C. Tau pathology in children and young adults: 33. Krut JJ, Zetterberg H, Blennow K, et al. Cerebrospinal
can you still be unconditionally baptist?. Acta Neuropathol fluid Alzheimers biomarker profiles in CNS infections.
2011;121:145147. J Neurol 2013;260:620626.

546 Neurology 87 August 2, 2016

2016 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 34. Holmberg B, Johnels B, Blennow K, Rosengren L. Cerebro- 47. Rabinovici GD, Jagust WJ, Furst AJ, et al. Abeta amyloid
spinal fluid Abeta42 is reduced in multiple system atrophy and glucose metabolism in three variants of primary pro-
but normal in Parkinsons disease and progressive supranu- gressive aphasia. Ann Neurol 2008;64:388401.
clear palsy. Mov Disord 2003;18:186190. 48. Alexopoulos P, Kriett L, Haller B, et al. Limited agreement
35. Blennow K, Wallin A, Agren H, Spenger C, Siegfried J, between biomarkers of neuronal injury at different stages of
Vanmechelen E. Tau protein in cerebrospinal fluid: a bio- Alzheimers disease. Alzheimers Dement 2014;10:684689.
chemical marker for axonal degeneration in Alzheimer dis- 49. De Meyer G, Shapiro F, Vanderstichele H, et al. Diagno-
ease? Mol Chem Neuropathol 1995;26:231245. sis-independent Alzheimer disease biomarker signature in
36. Olsson A, Vanderstichele H, Andreasen N, et al. Simulta- cognitively normal elderly people. Arch Neurol 2010;67:
neous measurement of beta-amyloid(1-42), total tau, and 949956.
phosphorylated tau (Thr181) in cerebrospinal fluid by the 50. Whitworth JA. 2003 World Health Organization (WHO)/
xMAP technology. Clin Chem 2005;51:336345. International Society of hypertension (ISH) statement on man-
37. Hall S, Ohrfelt A, Constantinescu R, et al. Accuracy of agement of hypertension. J Hypertens 2003;21:19831992.
a panel of 5 cerebrospinal fluid biomarkers in the differ- 51. Murray ME, Lowe VJ, Graff-Radford NR, et al. Clinico-
ential diagnosis of patients with dementia and/or parkin- pathologic and 11C-Pittsburgh compound B implications
sonian disorders. Arch Neurol 2012;69:14451452. of Thal amyloid phase across the Alzheimers disease spec-
38. Hesse C, Rosengren L, Andreasen N, et al. Transient increase trum. Brain 2015;138:13701381.
in total tau but not phospho-tau in human cerebrospinal 52. Thal DR, Beach TG, Zanette M, et al. [(18)F]flutemeta-
fluid after acute stroke. Neurosci Lett 2001;297:187190. mol amyloid positron emission tomography in preclinical
39. Ost M, Nylen K, Csajbok L, et al. Initial CSF total tau and symptomatic Alzheimers disease: specific detection of
correlates with 1-year outcome in patients with traumatic advanced phases of amyloid-beta pathology. Alzheimers
brain injury. Neurology 2006;67:16001604. Dement 2015;11:975985.
40. Skillback T, Rosen C, Asztely F, Mattsson N, Blennow K, 53. Wright JT Jr, Williamson JD, Whelton PK, et al. A ran-
Zetterberg H. Diagnostic performance of cerebrospinal domized trial of intensive versus standard blood-pressure
fluid total tau and phosphorylated tau in Creutzfeldt- control. N Engl J Med 2015;373:21032116.
Jakob disease: results from the Swedish Mortality Registry. 54. Grasbeck R. The evolution of the reference value concept.
JAMA Neurol 2014;71:476483. Clin Chem Lab Med 2004;42:692697.
41. Buerger K, Otto M, Teipel SJ, et al. Dissociation between 55. Knopman DS, Jack CR Jr, Wiste HJ, et al. Short-term
CSF total tau and tau protein phosphorylated at threonine clinical outcomes for stages of NIA-AA preclinical Alz-
231 in Creutzfeldt-Jakob disease. Neurobiol Aging 2006; heimer disease. Neurology 2012;78:15761582.
27:1015. 56. Klunk WE, Koeppe RA, Price JC, et al. The Centiloid
42. Blennow K, Hampel H. CSF markers for incipient Alz- Project: standardizing quantitative amyloid plaque estima-
heimers disease. Lancet Neurol 2003;2:605613. tion by PET. Alzheimers Dement 2015;11:115.
43. Fotuhi M, Do D, Jack C. Modifiable factors that alter the 57. Whitwell JL, Josephs KA, Murray ME, et al. MRI corre-
size of the hippocampus with ageing. Nat Rev Neurol lates of neurofibrillary tangle pathology at autopsy: a voxel-
2012;8:189202. based morphometry study. Neurology 2008;71:743749.
44. Crary JF, Trojanowski JQ, Schneider JA, et al. Primary age- 58. Hampel H, Schneider LS, Giacobini E, et al. Advances in
related tauopathy (PART): a common pathology associated the therapy of Alzheimers disease: targeting amyloid beta
with human aging. Acta Neuropathol 2014;128:755766. and tau and perspectives for the future. Expert Rev Neuro-
45. Josephs KA, Duffy JR, Fossett TR, et al. Fluorodeoxyglu- ther 2015;15:83105.
cose F18 positron emission tomography in progressive 59. Portelius E, Zetterberg H, Skillback T, et al. Cerebrospinal
apraxia of speech and primary progressive aphasia variants. fluid neurogranin: relation to cognition and neurodegen-
Arch Neurol 2010;67:596605. eration in Alzheimers disease. Brain 2015;138:3373
46. Wirth M, Villeneuve S, Haase CM, et al. Associations 3385.
between Alzheimer disease biomarkers, neurodegenera- 60. Villeneuve S, Reed BR, Madison CM, et al. Vascular risk
tion, and cognition in cognitively normal older people. and Abeta interact to reduce cortical thickness in AD vul-
JAMA Neurol 2013;70:15121519. nerable brain regions. Neurology 2014;83:4047.

Neurology 87 August 2, 2016 547

2016 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

A/T/N: An unbiased descriptive classification scheme for Alzheimer disease biomarkers
Clifford R. Jack, Jr, David A. Bennett, Kaj Blennow, et al.
Neurology 2016;87;539-547 Published Online before print July 1, 2016
DOI 10.1212/WNL.0000000000002923

This information is current as of July 1, 2016

Neurology is the official journal of the American Academy of Neurology. Published continuously since
1951, it is now a weekly with 48 issues per year. Copyright 2016 American Academy of Neurology. All
rights reserved. Print ISSN: 0028-3878. Online ISSN: 1526-632X.
 Updated Information & including high resolution figures, can be found at:
Services http://www.neurology.org/content/87/5/539.full.html

Supplementary Material Supplementary material can be found at:

http://www.neurology.org/content/suppl/2016/07/01/WNL.0000000000
002923.DC1
References This article cites 60 articles, 8 of which you can access for free at:
http://www.neurology.org/content/87/5/539.full.html##ref-list-1
Citations This article has been cited by 5 HighWire-hosted articles:
http://www.neurology.org/content/87/5/539.full.html##otherarticles
Subspecialty Collections This article, along with others on similar topics, appears in the
following collection(s):
All Imaging
http://www.neurology.org//cgi/collection/all_imaging
Alzheimer's disease
http://www.neurology.org//cgi/collection/alzheimers_disease
Cognitive aging
http://www.neurology.org//cgi/collection/cognitive_aging
MRI
http://www.neurology.org//cgi/collection/mri
PET
http://www.neurology.org//cgi/collection/pet
Permissions & Licensing Information about reproducing this article in parts (figures,tables) or in
its entirety can be found online at:
http://www.neurology.org/misc/about.xhtml#permissions
Reprints Information about ordering reprints can be found online:
http://www.neurology.org/misc/addir.xhtml#reprintsus

Neurology is the official journal of the American Academy of Neurology. Published continuously since
1951, it is now a weekly with 48 issues per year. Copyright 2016 American Academy of Neurology. All
rights reserved. Print ISSN: 0028-3878. Online ISSN: 1526-632X.