Volume Decline Associated With Aging, Alzheimer’s Disease, And Socioeconomic Status: Structural Neuroimaging Across The Adult Life-span

WASHINGTON UNIVERSITY Division of Biology and Biomedical Sciences Program in Neurosciences

Dissertation Examination Committee: Randy Buckner, Chair Denise Head John Morris Joseph Price Bradley Schlaggar David Van Essen

VOLUME DECLINE ASSOCIATED WITH AGING, ALZHEIMER’S DISEASE, AND SOCIOECONOMIC STATUS: STRUCTURAL NEUROIMAGING ACROSS THE ADULT LIFE-SPAN by Anthony Frank Fotenos

A dissertation presented to the Graduate School of Arts and Sciences of Washington University in partial fulfillment of the requirements for the degree of Doctor of Philosophy

May 2008 Saint Louis, Missouri

ACKNOWLEDGEMENTS

Washington University has built its reputation on a foundation I have come to learn is rock solid: a community of inspiring people. I am fortunate to be surrounded by such people, both at the University and at home, and many have contributed to this thesis. Starting with the older adult volunteers in our neuroimaging experiments, their wisdom and service have led me to wonder about all the positive aspects of lifelong development and not just to fear growing old. From the staff of the Alzheimer Disease Research Center (ADRC) responsible for pooling so many of these generous participants, I want to thank Amy Buckley, Virginia Buckles, and Mary Coats for clinical assistance and participant recruitment; Elizabeth Grant for database assistance; and Chengjie Xiong, David Johnson, and Cathy Roe for discussion of statistical procedures. I am grateful to Bruce Fischl, Anders Dale, Xiao Han, Rudolph Pienaar, Brian Quinn, and Andre van der Kouwe from the Martinos Center for Biomedical Imaging in Boston for collaborating on the application of their innovative magnetic resonance sequence. Within the Cognitive Neuroscience and Neuroimaging Labs, I thank Daniel Marcus, Mohana Ramaratnam, and Tim Olsen for database development and support; Dana Sacco, Erin Laciny, Jamie Parker, Susan Larson, Kate O'Brien, Laura Williams, Amy Sanders, and Glenn Foster for assistance with MRI and PET data collection; and my labmates, Luigi Maccotta, Ted Satterthwaite, and Ben Shannon, for serving as friendly and stimulating neighbors. Based on my invariably positive interactions with the staff of the University’s graduate programs, I appreciate how Brian Sullivan, Andrew Richards, Christy Durbin, Sally

ii

Vogt, and Anna Cook-Linsenman (formerly of the Buckner lab) keep everyone humming. I am especially grateful that Washington University provides faculty leaders whom I have learned from, emulate, and can look up towards with admiration and gratitude. Denise Head, Daniel Goldberg, Martha Storandt, Mark Mintun, John Morris, Joel Price, Brad Schlaggar, Martha Storandt, Avi Snyder, and David Van Essen are all exemplary scientists, who have shared their expertise generously with me as mentors. No one sets a higher standard than my advisor, Randy Buckner, who has given selflessly of his resources and shown by example how to think and work productively to advance our understanding of the brain. “What would Randy do?” has already become and I am sure will continue to serve as a refrain helping to guide me throughout my career. Finally, I acknowledge my grandparents, Kal, Frank, Max, and Miriam; my parents, Jim and Carol; my beloved wife, Saori; and my children, Naomi and Noah. The love and joy we share brings life to my life.

iii

TABLE OF CONTENTS

Acknowledgements…………………………………………………………………..

ii

Table of Contents…………………………………………………………………….

iv

List of Figures………………………………………………………………………..

v

List of Tables…………………………………………………………………………

vi

Abstract…………………………………………………………………………........

vii

Chapter 1:

Introduction, Background, and Significance………………………….

Chapter 2:

Normative Estimates of Cross-Sectional And Longitudinal Brain Volume Decline In Aging and AD……………………………………

Chapter 3:

Chapter 5:

15

Summary Of Three-Month Study to Discriminate Atrophy In AD from Nondemented Aging…………………………………………….

Chapter 4:

1

42

Brain Volume Decline In Aging: Evidence For A Relationship Between SES, Preclinical AD, And Reserve………………………….

52

General Discussion……………………………………………………

82

iv

LIST OF FIGURES

Figure 1.1

Age extrapolation issues and synapse estimates…...…………………

4

Figure 1.2

Longitudinal time-courses of normal brain volume decline…………..

7

Figure 2.1

Whole-brain volume measurement and normalization procedure…….

23

Figure 2.2

Cross-sectional plots of gray and white matter, normalized for head size ……………………………………………………………………

Figure 2.3

27

Cross-sectional and longitudinal plots of whole-brain volume, normalized for head size………………………………………………

30

Figure 2.4

Summary data…………………………………………………………

32

Figure 3.1

Multi-echo FLASH (MEF) imaging from one older adult participant..

44

Figure 3.2

Post-processing and null results for three-month discrimination with multi-echo FLASH (MEF)……………………………………………

Figure 4.1

Cross-sectional plot of brain volume in nondemented adults over the adult life-span…………………………………………………………

Figure 4.2

48

69

Cross-sectional and longitudinal plots of brain volume as a function of socioeconomic status……………………………………………….

71

Figure 4.3

Adjusted whole brain volume by PIB status………………………….

75

Figure 4.4

The relationship between adjusted whole brain volume and SES is stronger in nondemented participants who subsequently develop

Figure 5.1

dementia………………………………………………………………

76

Thesis results within a multiple factor framework of brain aging……

83

v

LIST OF TABLES

Table 2.1

Sample characteristics………………………………………………...

Table 2.2

Estimated gray, white, and whole-brain volume (WBV) by age

20

decade in nondemented individuals…………………………………...

29

Table 2.3

Comparative whole-brain estimates from cross-sectional, MR studies

36

Table 3.1

Longitudinal sample with complete baseline and follow-up MEF images…………………………………………………………………

43

Table 4.1

MRI sample…………………………………………………………...

61

Table 4.2

PIB amyloid imaging sample…………………………………………

66

vi

ABSTRACT OF THE DISSERTATION

Volume Decline Associated with Aging, Alzheimer’s Disease, and Socioeconomic Status: Structural Neuroimaging across the Adult Life-span by Anthony Frank Fotenos Doctor of Philosophy in Biology and Biomedical Sciences (Neurosciences) Washington University in St. Louis, 2006 Professor Randy Buckner, Chair

This thesis concerns the neurodegeneration in the most burdensome neurodegenerative disease, Alzheimer’s (AD). A unifying aim is to clarify how neurodegeneration in AD relates to aging (a risk factor for AD) and socioeconomic status (SES, a protective factor for AD). Three quantitative structural magnetic resonance imaging (MRI) studies toward this goal are described. The first study characterized crosssectional and longitudinal rates of whole brain volume decline in nondemented adults and compared these normal rates to atrophy measured early in symptomatic AD. The results based on cross-sectional and longitudinal measures overlapped and showed that nondemented aging is accompanied by steady volume decline even in the youngest adults, with marked acceleration in the earliest stages of dementia. Standard imaging methods used in this initial study required almost two years of follow-up to discriminate between longitudinal change in demented and nondemented samples. The second study aimed to reduce this follow-up interval to a more clinically

vii

practical three months, principally through the implementation of recently developed multi-echo fast low-angle shot (MEF) MR sequences. Null results for this three-month study are summarized. The third study derived from the initial finding of brain aging in the absence of dementia. SES is one factor known to protect against dementia incidence. We found that older adults with high SES have reduced brain volume (cross-sectional result) and more rapid volume loss (longitudinal result) than less privileged peers. Additional findings based on amyloid imaging with positron emission tomography (PET) and clinical followup suggest that the capacity of individuals with high SES to cope longer with preclinical AD pathology, consistent with the reserve hypothesis, may help to explain these counterintuitive main results. Implications of this thesis research and possible future directions are discussed within the context of a multiple factor framework of brain aging.

viii

CHAPTER 1 INTRODUCTION, BACKGROUND, AND SIGNIFICANCE

Life-span brain morphometry. In 1907, Alois Alzheimer linked brain atrophy with a new disease that would come to bear his name (AD). He wrote in his case report on Auguste D., “The post-mortem showed an evenly atrophic brain without macroscopic focal degeneration…Only a tangle of fibrils indicates the place where a neuron was previously located…Many neurons, especially the ones in the upper layer, have completely disappeared” (Alzheimer et al., 1995). In the same report, Alzheimer also identified widespread “miliary foci,” since recognized as β-amyloid (Aβ) plaques (Glenner and Wong 1984, Masters et al., 1985). Today, an emerging hypothesis regarding AD pathogenesis arranges Aβ and neurofibrillary tangles (NFTs) in a cascade leading to neuronal degeneration, atrophy, and dementia (Hardy and Selkoe 2002, Walsh and Selkoe 2004). Other models start differently (for example, Lee et al., 2004, de la Torre 2004), but they end the same, with at least three related implications for structural neuroimaging, the method on which this thesis is based. First, given the link between AD pathogenesis and atrophy, neuroimaging offers a noninvasive opportunity to measure AD damage at the macroscopic structural level (Jack et al., 2002, Silbert et al., 2003, Csernansky et al., 2004). Second, curing AD pathogenesis in advanced dementia might arrest patients in a state of neurodegeneration and permanent dysfunction. Structural neuroimaging is thus increasingly oriented toward early disease detection (DeKosky and Marek 2003, Glodzik-Sobanska et al., 2005). Third, early detection of AD-dependent

1

structural change requires careful comparison to normal brain morphometry (Giedd 2004, Raz 2004). I will accordingly begin with an overview of normal brain aging, and then turn to a discussion of AD, morphometric methods, and possible modifiers as background and motivation for the thesis research discussed in Chapters 2-5. This thesis samples over an extended period of human adulthood from 18 to 97. Within this age range, older adults (≥ 65) account for a growing proportion of the developed world’s population and carry the heaviest burden of chronic disease (Goulding et al., 2003). Aging research thus tends to focus on older adults. Does the structure of their brains normally differ from that of younger adults (~18-44)? Starting at the cellular level, in terms of neuron number, the emerging consensus favors overall stability (<10% decline between age 20 and 90; Haug et al., 1984, Pakkenberg et al., 2003), with limited areas of age-related loss (for example, in the substantia nigra; reviewed in Morrison and Hof 1997, Turlejski and Djavadian 2002). Preliminary data also suggest minimal loss with age in the number of glia (Pakkenberg et al., 2003). In contrast to the numerical size stability of neurons, dendritic spine counts and spine density estimates are markedly reduced in the old (for example, over 40% less in apical dendrites of layer III pyramidal neurons around the superior temporal sulcus; Duan et al., 2003; reviewed in Uylings and de Brabander 2002). Consistent with cortical spine loss, synapse (specifically, input) elimination has been demonstrated in the parasympathetic ganglion of the aged mouse (Coggan et al., 2004). Subcortically, in white matter, autopsy studies have increasingly focused on extensive fiber loss, especially of small myelinated fibers; neuropathological estimates of adult-span white volume decline run as high as 25% (Pakkenberg et al.,

2

2003, Svennerholm et al., 1997), and these may underestimate fiber loss, which appears to decline exponentially with age (Marner et al., 2003). This pattern of late acceleration, associated ultrastructural abnormalities, and the unique ischemic vulnerability of oligodendrocytes all point to the possibility that microscopic white matter lesions represent a distinct senescent process common after middle age (Peters and Rosene 2003, Bartzokis 2004a). Microscopic autopsy research faces the difficulty of collecting representative human samples and the time demands of characterizing neurons and glia (for discussion of sampling problems, see Svennerholm et al., 1997). As a consequence, many aging studies employ the efficient extreme group design, comparing oldest to youngest adults and precluding time-course analysis (Salthouse 2000). A particularly clear example of this sampling limitation comes from the following plots of synapse density in 21 brains, 8 from adults over age 60 and 7 from infants under one (Huttenlocher 1979).

3

Synapses/Neuron x 104

Synapses/mm3 x 108

FIGURE 1.1

Figure 1.1. Age extrapolation issues and synapse estimates from layer III of human middle frontal gyrus from Huttenlocher (1979). (A) Raw synapse numbers from electron photomicrographs. Note non-uniform x-axis and skewed age distribution. (B) Synapse numbers accounting for changes in neuron number; note changed x-axis scaling. The often cited plot is on the left (Figure 1.1A); it shows that electron microscopic counts of phosphotungstic acid-stained synaptic profiles increase in the first year of life and are generally higher in early childhood than in old age. The plot on the right (Figure 1.1B) provides a moderated view by accounting for differences in neuron number with age. Subsequent investigations of pediatric samples have replicated the essential finding that synaptic density peaks within the first few years of life; they also provide evidence that a decline in synapse number begins in childhood (reviewed in Feinberg et al., 1990, Huttenlocher 2002). However critical evaluation of the age distribution in Figure 1.1, particularly accounting for the non-uniform age-axis (often a

4

sign of biased sampling), shows the 21-brain sample “ranging from newborn to 90 years” has a gaping hole through 80% of that range. The same can thus be said about the fit line through so little data. This example prompts two related observations. First, understanding of human aging tends to be weakest in the working adult range (~25-60), perhaps because this group works to survive into old age prior to autopsy. Second, it may be helpful to distinguish between aging in the familiar sense of growing old and less fit (senescent aging) versus aging in the constitutive sense of courses of change that extend over the lifetime (physiological aging, a gap in the understanding of which Figure 1.1 highlights). Estimating time-courses, even limited to adulthood, requires large sample sizes, generally favoring macroscopic measures, such as brain weight and volume. How does normal aging appear with more data at the macroscopic level? Early autopsy reports did not account for the weight (Dekaban 1978) or volume (Davis and Wright 1977) of ventricular CSF; they modeled changes with age as flat prior to around age 50, and declining thereafter (~10% total; though see Pakkenberg and Voigt 1964, Miller et al., 1980). However, more recent autopsy (Svennerholm et al., 1997) and cross-sectional MRI studies almost universally converge on a linear time-course of volume decline in young adulthood, with the largest samples showing mild acceleration after middle age (Good et al., 2001, DeCarli et al., 2004). In contrast to the whole brain pattern of early decline, white volume estimates remain stable (Pfefferbaum et al., 1994, Blatter et al., 1995, Guttmann et al., 1998) or increase slightly (Courchesne et al., 2000, Ge et al., 2002, Sowell et al., 2003, Jernigan and Fennema-Notestine 2004, Walhovd et al., 2005,

5

Kruggel 2006) until middle-age (age 40 to 60). White volume decline follows and accelerates in old age, in agreement with the recent autopsy findings on fiber loss discussed above. MRI studies that report minimal white volume decline (Van Laere and Dierckx 2001) tend to have fewer or healthier older participants, reinforcing the importance of sample health and age distribution when comparing studies. To attribute effects to age, designers of cross-sectional MRI and autopsy studies must assume that older and younger individuals differ only with respect to their ages and not their birth cohorts or other variables. Longitudinal studies are now possible to test these assumptions using in vivo neuroimaging. In the study described in Chapter 2, we directly compared longitudinal decline estimates from a subset of nondemented adults to cross-sectional estimates from the larger sample. At least seven prior studies have quantified longitudinal, whole brain decline in nondemented adults over 60. Six (Chan et al., 2001, Wang and Doddrell 2002, Liu et al., 2003, Resnick et al., 2003, Thompson et al., 2003, Jack et al., 2004) documented annualized loss of about 0.5% (0.37 to 0.88), whereas one (Tang et al., 2001) reported a rate of 2.1%. Longitudinal studies sampling from younger adults have found slower, not non-zero, rates of decline (0.2-0.3%/yr; Giedd et al., 1999, Liu et al., 2003), again consistent with an acceleration of white volume decline in older age. Figure 1.2 shows data from two independent longitudinal studies designed to provide insight on physiological aging (Liu et al., 2003, Raz et al., 2005). The left graph (Figure 1.2A) plots whole brain volume adjusted for intracranial volume, derived using automated segmentation (see legend). The right graph (Figure 1.2B) plots the volume of

6

the lateral prefrontal cortex, also adjusted for intracranial volume, but derived from manual measurements. Each line represents an individual study participant. Brain volume appears to decline continuously throughout adolescence and adulthood.

Adjusted Volume (cm3)

FIGURE 1.2 A

Age (Years)

B

Age (Years)

Figure 1.2. Longitudinal time-courses of normal brain volume decline. (A) Whole brain volume covariance adjusted (statistically matched) for intracranial volume from 90 healthy volunteers, concentrated at baseline in the 14 to 55 age range (mean 37), with repeat imaging after 3.5 years. Volume was measured via automated gray/white/csf segmentation of longitudinally registered scan pairs from each study participant (from Figure 2 in Liu et al., 2003). (B) Manually segmented lateral prefrontal cortical volume from 72 healthy volunteers, 20 to 77 at baseline (mean 52), with repeat imaging after 5 years (from Figure 6 in Raz et al., 2005). Alzheimer’s Disease. The definitive diagnosis of AD requires histopathologic confirmation (Ball et al., 1997, McKeel et al., 2004); it is uncertain when the disease process starts in any given individual (Borenstein et al., 2006). Cross-sectional study of AD progression would call for individuals to be matched on all variables other than time since disease onset. Given the uncertainty surrounding onset, longitudinal neuroimaging has a key role to play in characterizing the progression of AD, since age of onset is naturally matched within subject (Kantarci and Jack 2003, Glodzik-Sobanska et al., 7

2005). A handful of reports on serial MRI have found whole brain atrophy accelerates in AD, but differ as to the magnitude. The difference in annualized atrophy rates between nondemented individuals who converted to dementia of the Alzheimer type (DAT) during follow-up (0.8%) and those with slow-progressing (0.6%) and fast-progressing (1.4%) DAT at baseline suggests that dementia severity and atrophy rate are associated (Jack et al., 2004). Accordingly, most reports on more severely demented samples estimate whole brain atrophy rates more rapid than 2% per year (Chan et al., 2001, Wang et al., 2002, Thompson et al., 2003, Schott et al., 2005), consistent with a nonlinear (accelerating) atrophy model of DAT (Chan et al., 2003, Rusinek et al., 2004). Regarding the anatomy of atrophy in AD, a recent report from our laboratory (Buckner et al., 2005) compared the spatial distribution of maps from five different neuroimaging methods, based on different samples, and proposes a provocative hypothesis regarding the natural history of AD. Nonlinear registration (see below) was used to estimate atrophy at the voxel level in participants with DAT. Compared to nondemented participants, in which declines have been found steepest in lateral prefrontal, orbitofrontal, and inferior parietal regions (Jernigan et al., 2001, Ohnishi et al., 2001, Sowell et al., 2003, Salat et al., 2004, Raz et al., 2005), atrophy in DAT was most prominently accelerated in medial temporal regions and a distributed network of parietal cortex, including the precuneus, posterior cingulate, retrosplenium, and lateral posterior parietal regions, consistent with prior studies (Brun and Gustafson 1976, Callen et al., 2001, Ohnishi et al., 2001, Scahill et al., 2002, Yoshiura et al., 2002, Boxer et al., 2003, Miller et al., 2003, Thompson et al., 2003, Karas et al., 2004, Chetelat et al., 2005,

8

Pennanen et al., 2005). Positron emission tomography (PET) maps of default activity (areas receiving more blood flow when young healthy participants were not engaging in goal-directed tasks across a variety of experiments; reviewed in Gusnard and Raichle 2001) and MRI maps of memory retrieval (areas with increased signal when young, healthy participants correctly recognize old versus new items; reviewed in Wagner et al., 2005) showed surprising anatomical convergence, particularly in parietal cortex. This convergence between maps of functional and metabolic activity in young adults and pathology in late-onset DAT suggests functional or metabolic activity early in life may contribute to AD in older age (Borenstein et al., 2006, Selkoe 2006).

Morphometric methods. How is brain structure measured to generate the findings and hypotheses surveyed above? My interest is with individual brain change over time, but a historical review (Rushton and Ankney 1996) of controversial, between-group brain differences serves as a reminder that Morton (1849), Broca (1861), and Galton (1888), among other nineteenth-century scientists, pioneered the field of head and brain morphometry. More recent morphometric developments in the field of neuropathology include the invention of a pneumatic device for assessing intracranial volume (Davis and Wright 1977) and the random sampling, 3-D dissector method of counting neurons without bias (reviewed in Morrison and Hof 1997). In the 1970s, the developers of modern MRI variously solved the problem of 3-D NMR (Damadian 1971, Lauterbur 1973, Mansfield and Maudsley 1977). The conceptual basis for all spatial encoding schemes is the linear equation relating magnetic field

9

strength to the resonance (Larmor) frequency of the hydrogen proton. Magnetic field strength can be regularly varied using gradient coils, such that for a given gradient and radio frequency (RF) bandwidth, protons at a known fixed interval (spatial frequency) along a known slice thickness can all be made to precess at multiples of the same known Larmor frequency. Intrinsic tissue properties (T1, T2, and proton density, PD) and magnetic field susceptibility differences determine the re-transmitted radio signal received from these spatially selected protons during the readout interval. Spin-lattice relaxation time (T1) or spin-spin relaxation time (T2) can be emphasized, depending on the time between excitations (repetition time, TR) and the time between gradient switches (echo time, TE). As described in Chapter 3, we have collected longitudinal MR data using a new multi-echo fast low-angle shot (MEF) sequence, from which intrinsic tissue property estimates (T1, T2*, and PD) can be derived (Fischl et al., 2004). Several strategies exist for next transforming a high resolution (~cubic mm) brain image into morphometric data. Manual methods based on tracing the outline of known anatomical boundaries are considered the gold standard (for example, Jernigan et al., 2001, Raz et al., 2004, Head et al., 2005). However, automated methods offer the potential for greater throughput and extensibility. Within the category of automated volumetric methods, linear (Woods et al., 1992, Snyder 1996, Smith et al., 2002, Buckner et al., 2004) and nonlinear (Christensen et al., 1996, Miller 2004) registration is used in this research. In general, the goal of registration is to match for and thereby eliminate uninteresting sources of structural variance such as scanner drift, head positioning, and head size. Registration to stereotaxic atlases also enables discussion of space to occur in

10

the same language. The highest parameter linear registration is a 12-parameter affine; it represents a procedure for minimizing the difference (registration error) between a source and target image by translating, rotating, stretching, and skewing the source image. An important methodological question for measuring brain change, addressed by an earlier study from our lab (Buckner et al., 2004), is whether head size accounts for the scaling properties of linear registration. We found a very high (r = 0.93) correlation between manual (total intracranial volume, TIV) and registration-based (estimated TIV, eTIV) head size estimates, and both were independent of brain atrophy in demented samples. These findings indicate that differences in head size drive image scaling when the difference between a source MRI and sample-representative target atlas are minimized by linear registration. Head-size corrected estimates are reported either as ratios or residuals of eTIV (Mathalon et al., 1993, Sanfilipo et al., 2004, Van Petten 2004). Residual correction is more robust against group differences in head size; however, for comparisons in which these have been explicitly studied and ruled out (Buckner et al., 2004), we have used ratio correction as the more conventional (for example, (Pantel et al., 2004, DeCarli et al., 2004, Kruggel 2006) and transparent alternative (see Chapters 2 and 3). Starting where linear registration leaves off, high dimensional nonlinear registration computes transformations at the voxel level in order to reduce registration error to zero, under assumptions that brain images behave according to known material properties (for example, like a viscoelastic fluid; for a review of how fluid warping is being applied, see May et al., 2006). The extent of voxel-wise contraction or expansion

11

embedded within the nonlinear transformation provides an estimate of regional volume change, though the precise relationship between the magnitude and spatial distribution of such estimates and underlying anatomy remains to be validated. We have implemented a fluid warping algorithm in order to compare longitudinal, voxel-wise atrophy estimates in normal old aging and DAT (Buckner et al., 2005, discussed above). The widely used voxel-based morphometric (VBM) method is based on registration that ranges from linear to “global nonlinear,” the latter involving some nonlinear warping but with final registration error greater than zero (Ashburner and Friston 2000, Ashburner and Friston 2001). As a result of what is essentially a compromise between precision and computational speed, significant structural differences reported using VBM should probably be interpreted as gross regional estimates and viewed in the context of metric whole brain gray and white differences that drive them (Bookstein 2001, Tisserand et al., 2002, Mehta et al., 2003, Davatzikos 2004). Unfortunately, these metric estimates often go unreported in VBM papers and negative findings tend to be overemphasized, leading to difficulty interpreting some papers that rely exclusively on this method (for example, Maguire et al., 2000, Colcombe et al., 2003, Mechelli et al., 2004).

Modifiers. Relevant to understanding brain aging in nondemented samples (our aim in Chapter 4), there are few studies that demonstrate modification of the downward slopes illustrated in Figure 1.2. Sluming and colleagues (2002) reported that whole brain volume was less correlated with age in a group of professional musicians than matched

12

nonmusicians controls. VBM analysis with a lower statistical threshold in an a priori region of left frontal cortex confirmed that gray volume was significantly increased there in musicians. A more recent VBM study of older adults found a significant interaction between age and estimates of maximal oxygen uptake, such that higher fitness was associated with greater than age-predicted gray volume throughout association cortex and greater than age-predicted white volume, particularly in the frontal lobes (Colcombe et al., 2003). Both these studies suggest long-term motor-related activity moderates ageassociated volume decline. Five other reports on normal structural modifiers fail to find such moderation for estimates of long-term cognitive activity; if anything, the reports suggest an amplifying role. Coffey and colleagues (1999) report a very weak but significant correlation between education and cerebral spinal fluid (CSF) volume in 320 normal older adults (such that the most educated had more sulcal CSF). A meta-analysis of 33 published papers on volume decline in the hippocampus also failed to find moderation by fitness of memory, as assessed by cognitive testing (Van Petten 2004). Again, the trend pointed in the amplifying direction, especially with younger samples (such that those remembering most had the smallest hippocampi). More recently in children, vocabulary gains (Sowell et al., 2004) and superior IQ (Shaw et al., 2006) have been linked to distributed regions of increased cortical thinning. In contrast to this mixed picture of non-pathological brain aging modification, considerable evidence suggests a variety of long-term environmental or experiential factors moderate the risk of dementia. In particular, education, occupation, literacy, IQ,

13

and active lifestyle all have experimental support as protective factors against DAT (Gurland 1981, Zhang et al., 1990, Whalley et al., 2000, Fratiglioni et al., 2004, Manly et al., 2005, Valenzuela and Sachdev 2006). The evidence for modification generally falls into two categories: a negative correlation with indices of disease expression and a positive correlation with indices of pathological severity. For example, more educated individuals have been linked to a lower incidence (reduced risk) of dementia diagnosis (Launer et al., 1999, Valenzuela and Sachdev 2006). In contrast, at the cusp of dementia or in patients otherwise matched for clinical severity, cognitive performance and glucose metabolism have been found to decline more rapidly in the more educated (Stern et al., 1992, Amieva et al., 2005, Scarmeas et al., 2006). A combination of direct relationships between education and pathology and between education and physiological brain aging might conceivably contribute to these findings (Snowdon 2003, Shaw et al., 2006). However, the reserve hypothesis offers the most parsimonious explanation. The hypothesis is that education and related variables moderate the relationship between pathology and disease expression (Katzman 1993, Satz 1993, Stern 2002, Bennett et al., 2005, Scarmeas and Stern 2004, Bennett et al., 2005, Stern 2006, Roe et al., 2006). We will return to the reserve hypothesis as a potential explanation for our finding that brain volume decline is associated with educational and occupational attainment (socioeconomic status or SES) in nondemented older adults in Chapter 4.

14

CHAPTER 2 NORMATIVE ESTIMATES OF CROSS-SECTIONAL AND LONGITUDINAL BRAIN VOLUME DECLINE IN AGING AND AD

ABSTRACT

Objective: To test the hypotheses 1) that whole-brain volume decline begins in early adulthood, 2) that cross-sectional and longitudinal atrophy estimates agree in older, nondemented individuals, and 3) that longitudinal atrophy accelerates in the earliest stages of Alzheimer’s disease (AD). Methods: High-resolution, high-contrast structural magnetic resonance images (MRIs) were obtained from 370 adults (age 18 to 95). Participants over 65 (n = 192) were characterized using the Clinical Dementia Rating (CDR) as either nondemented (CDR 0, n = 94) or with very mild to mild dementia of the Alzheimer type (DAT, CDR 0.5 and 1, n = 98). Of these older participants, 79 belonged to a longitudinal cohort and were imaged again a mean 1.8 years after baseline. Estimates of gray matter (nGM), white matter (nWM), and whole-brain volume (nWBV) normalized for head sizes were generated based on atlas registration and image segmentation. Results: Hierarchical regression of nWBV estimates from nondemented individuals across the adult life-span revealed a strong linear, moderate quadratic pattern of decline beginning in early adulthood, with later onset of nWM than nGM loss. Wholebrain volume differences were detected by age 30. The cross-sectional atrophy model overlapped with the rates measured longitudinally in older, nondemented individuals

15

(mean decline of -0.45% per year). In those individuals with very mild DAT, atrophy rate more than doubled (-0.98% per year). Conclusions: Nondemented individuals exhibit a slow rate of whole-brain atrophy from early in adulthood with white-matter loss beginning in middle age; in older adults, the onset of DAT is associated with a markedly accelerated atrophy rate.

INTRODUCTION

Pathological brain processes that lead to dementia coexist with normal aging processes that also influence the brain but do not manifest as disease. To better understand the nature of normal brain development in advanced aging and how the earliest stages of dementia of the Alzheimer type (DAT) cause departure from that trajectory, we report here a large-sample study of 370 adults age 18 to 95. The goals of this study were to characterize the normal development of whole-brain volumes in the absence of dementia and determine, through a combination of cross-sectional and longitudinal estimates, to what degree the presence of early-stage DAT causes departure from normal development. An important feature of the study design is the direct contrast of cross-sectional and longitudinal estimates of brain change. Cross-sectional estimates are efficient in that a single measure can be used as the dependent measure. To the degree that different normal individuals have predictable brain sizes and changes in brain size, a single point estimate may be informative regarding their likely future course and risk of disease.

16

However, reviews of the structural aging literature highlight the need for longitudinal data because of between-subject variance (Raz 2000, Uylings and de Brabander 2002, Kantarci and Jack 2003). Longitudinal data reduce between-subject variance by using an individual as his or her own baseline and also control for differences that potentially complicate cross-sectional samples. For example, cross-sectional samples may include hidden group heterogeneity (cohort effects), such as environmental differences between when people were born (secular effects). MRI is readily able to obtain longitudinal data through repeated imaging of the same person over time (reviewed in Raz 2004). The present design, which combines cross-sectional and longitudinal approaches (Giedd et al., 1999), allows three basic questions to be addressed. First, to what extent and at what age does nondemented aging associate with cross-sectional brain volume reduction? Some volumetric reports suggest that whole-brain volume is stable in nondemented adults under 50 (Davis and Wright 1977, Miller et al., 1980, Matsumae et al., 1996, Guttmann et al., 1998, Ge et al., 2002), whereas others find volume loss in this age range (Pfefferbaum et al., 1994, Giedd et al., 1999, Courchesne et al., 2000, Rovaris et al., 2003, Sowell et al., 2003, Salat et al., 2004), a difference possibly related to differential contributions of gray- and white-matter loss to brain aging (Raz 2004). Second, does the cross-sectional rate of atrophy in nondemented older adults match the longitudinal rate? As noted above, a number of potential confounds could lead to a mismatch between cross-sectional and longitudinal findings. If the cross-sectional observations accurately predict the longitudinal atrophy rate, it is reasonable to assume that cohort and secular effects are minimal and volume loss progresses in a predictable

17

manner in the absence of dementia. Finally, to what extent does the rate of whole-brain atrophy accelerate in early-stage DAT? The available reports addressing this question have found significant acceleration, but differ as to its magnitude (Cardenas et al., 2003, Chan et al., 2003, Jack et al., 2004).

METHOD

Participants. Three hundred and seventy adults (age 18 to 95 at baseline) participated in a structural MR imaging session. Of these individuals, 79 participated on two separate occasions separated by an extended interval to allow for longitudinal data analysis (1.0 to 3.9 year interval; mean = 1.8 years). Twenty additional individuals were scanned twice at a short interval (mean = 21 days, range 1 to 64) to allow estimation of measurement reliability. Participants were paid for their participation and gave informed consent in accordance with guidelines of the Washington University Human Studies Committee. Data from subsets of the participants have been used in previous studies (Salat et al., 2004, Head et al., 2005). Young and middle-aged adults were recruited from the Washington University community. Nondemented and demented older adults were recruited exclusively from the ongoing longitudinal sample of the ADRC. The ADRC volunteers are more likely than the population of the St. Louis metropolitan area to have a high school education, and volunteers with severe comorbidities such as major depression or disabling stroke are excluded (Villareal et al., 2003). Approximately 40% of ADRC participants who met the

18

study’s clinical criteria (nondemented or dementia restricted to DAT) declined to participate in an MRI; 7% were ineligible based on MRI contraindications. There was no statistically significant difference in age, years of education, or scores on the mini-mental state exam (MMSE; Folstein et al., 1975) between ADRC participants who did and did not undergo MRI. Dementia severity was quantified using the Clinical Dementia Rating (CDR; Morris 1993) scale for all ADRC volunteers, and recruitment for MRI was independent of longitudinal clinical progression. The average duration between clinical assessment and participation in the MRI session was 101 days (range = 3 to 332 days). The 98 participants with DAT exhibited very-mild (CDR 0.5; n = 69) to mild (CDR = 1, n = 29) dementia severity. Of the 205 older adults who underwent MRI, 13 (6%; 7 CDR = 1, 3 CDR = 0.5, 3 CDR = 0) did not complete the imaging protocol; one dropped out on repeat imaging. Although several DAT participants had cognitive test scores (e.g., MMSE) that might qualify for classification as mild cognitive impairment, a CDR score of 0.5 or greater in this sample is highly predictive of Alzheimer’s disease, both in clinical progression and neuropathological diagnosis at autopsy (Berg et al., 1998, Morris et al., 2001, Galvin et al., 2005). Demographic and clinical data for participants are presented in Table 2.1.

19

TABLE 2.1. Sample Characteristics Young

N (cross-sectional)

Middle-aged

Old CDR 0

DAT (CDR 0.5)

DAT (CDR 1)

127

51

94

69

29

Female/Male

64/63

29/22

67/27

32/37

21/8

Age ± SD, yrs

23±3 (18-34)

50±8 (35-64)

78±8 (65-95)

78±6 (65-93)

79±6 (69-97)

Education ± SD, yrs

15±3 (8-23)

14±3 (7-20)

13±3 (7-20)

MMSE ± SD

29±1 (25-30)

26±3 (18-30)

22±4 (13-28)

Prescriptions, n

2.9±2.1 (0-9)

2.4±1.9 (0-8)

2.7±2.3 (0-8)

Systolic BP, mmHg

136±18 (102-192)

143±21 (104-188)

146±26 (90-192)

Diastolic BP, mmHg

73±10 (40-96)

73±10 (50-98)

77±11 (60-100)

Reported HBP, %

43.0

41.8

48.3

Diabetes, %

10.8

11.8

10.7

38

33

8

30/8

10/23

5/3

1.8±0.5 (1.1-3.9)

1.8±0.5 (1.3-3.5)

1.8±0.4 (1.0-2.4)

N with follow-up (longitudinal) Female/Male Scan interval ± SD, yrs

Notes: The sample consisted of 370 individuals (272 nondemented and 98 with DAT). DAT = dementia of the Alzheimer type; MMSE = mini-mental state examination where scores range from 30 (“best”) to 0 (“worst”); CDR = Clinical Dementia Rating, with 0, 0.5, and 1 corresponding to nondemented, very mild, and mild DAT; HBP = high blood pressure; TIA = transient ischemic attack. Mean values given ± standard deviation (SD).

20

Image acquisition. Multiple (three or four) high-resolution structural T1-weighted magnetization-prepared rapid gradient echo (MP-RAGE) images were acquired on a 1.5T Vision scanner (Siemens, Erlangen, Germany). MP-RAGE parameters were empirically optimized for gray-white contrast (repetition time (TR) = 9.7 ms, echo time (TE) = 4 ms, flip angle (FA) = 10, inversion time (TI) = 20 ms, delay time (TD) = 200 ms, 256 x 256 (1 mm x 1 mm) in-plane resolution, 128 sagittal 1.25 mm slices without gaps, time per acquisition = 6.6 min). Participants were provided cushioning, head phones, and a thermoplastic face mask for communication and to minimize head movements. Positioning was low in the head coil (toward the feet) to center the field of view on the cerebral hemispheres. The MP-RAGE images were acquired as the second part of a 70-minute CAP protocol that also included fast low angle shot (FLASH) gradient echo, turbo spin echo (TSE), and diffusion tensor imaging (DTI) acquisitions. The DTI data have been reported elsewhere (Head et al., 2004).

Image analysis. Normalized gray-matter (nGM; gray parenchyma within the entire intracranial volume down to approximately the superior arch of C1), white-matter (nWM), and whole-brain volume (nWBV; gray plus white parenchyma) were computed for each image session. The procedure was based on a validated, open-source segmentation tool (Zhang et al., 2001, Smith 2002). Prior to image segmentation, the images were pre-processed to normalize for head-size and intensity variation that might affect image segmentation.

21

Pre-processing included multiple steps. Head-size normalization used a validated method based on atlas registration (Buckner et al., 2004). The normalization is proportional to manually measured total intracranial volume (TIV, r = 0.93) and minimally biased by atrophy. Images were corrected for inter-scan head movement and spatially warped into the atlas space of Talairach and Tournoux (1988). The template atlas consisted of a combined young-and-old target previously generated from a representative sample of the young (n = 12) and nondemented old (n = 12) adults. The use of a combined template has been shown to minimize the potential bias of an atlas normalization procedure to over-expand atrophied brains (Buckner et al., 2004). For registration, a 12-parameter affine transformation was computed to minimize the variance between the first MP-RAGE image and the atlas target (Snyder 1996). The remaining MP-RAGE images were registered to the first (in-plane stretch allowed) and resampled via transform composition into a 1-mm isotropic image in atlas space. All images were visually inspected to verify appropriate atlas transformation. The result was a single, high-contrast, averaged MP-RAGE image in atlas space (see Figure 2.1). Subsequent preprocessing steps included skull removal by application of a loose-fitting atlas mask and correction for intensity inhomogeneity due to non-uniformity in the magnetic field. Intensity variation was corrected across contiguous regions, based on a quadratic inhomogeneity model. Following pre-processing, the segmentation algorithm classified each voxel of the average image as cerebral spinal fluid (CSF), gray, or white matter (Zhang et al., 2001, Smith 2002). This segmentation starts with an initial estimation step to obtain and

22

classify tissue parameters. An expectation-maximization algorithm then updates class labels and tissue parameters in order to iterate toward the maximum likelihood estimates of a hidden-Markov, random-field model. The model uses spatial proximity to constrain the probability with which voxels of a given intensity are estimated to belong to each tissue class. FIGURE 2.1

Figure 2.1. Whole-brain volume measurement and normalization procedure. (A) Single MP-RAGE image as acquired in native space. (B) Within-participant averaged MP-RAGE image (n=4); note the increased contrast-to-noise. (C) Averaged image after registration to a target atlas composed of representative young and old individuals in Talairach and Tournoux (1988) space. (D) Averaged, atlas-registered image after masking and field-inhomogeneity correction. (E) Final segmented image; normalized whole-brain volume (nWBV) is defined as the percentage of the brain mask (non-black background) occupied by voxels classified as gray and white matter. Normalized volumes were computed as the proportion of all voxels within the brain mask occupied by gray (nGM), white (nWM), or gray plus white (nWBV, equivalent to 100 – %CSF) voxels. The unit of normalized volume is percent, which represents the percentage of estimated TIV.

23

Test-retest measurement reliability. In order to estimate measurement reliability, normalized volumes for the same person were compared over two imaging sessions separated by a brief interval, during which it is reasonable to assume minimal true change. Twenty individuals contributed to the test-retest group (young, n = 16; middleaged, n = 1; older, n = 3), with a mean delay of 21 days between test and retest (range 1 to 64 days). The mean absolute percentage difference (MAPD, the absolute difference between test and retest volumes divided by the overall mean, expressed in percent) was 0.92% for nGM (CI 0.53-1.30), 0.80% for nWM (CI 0.5-1.1), and 0.49% for nWBV (CI 0.28-0.71). The coefficients of variation (CV, the standard deviation of the difference between test and retest volumes divided by the overall mean, expressed in percent) were 1.24% (nGM), 1.04% (nWM), and 0.68% (nWBV). The interclass correlations between values paired by participant, but randomly assigned to test or retest, were high (r[nGM] = 0.99, r[nWM] = 0.98, and r[nWBV] = 0.99).

Cross-sectional analysis. Normalized volumes were plotted against age for the 370 unique participants. Statistical analysis was conducted with the JMP software package (SAS Institute, Cary, North Carolina). Hierarchical polynomial regression was used to test between linear and curvilinear models of cross-sectional volume as a function of age. With the sample restricted to one volume measurement per nondemented individual, higher order terms of the subject’s age-at-scan were tested until they no longer contributed significantly to the model; the resulting models are referred to as crosssectional, nondemented aging curves. Normalized volumes in the older DAT and older

24

nondemented samples were then compared using analysis of covariance with nGM, nWM, or nWBV as the dependent measure and age, gender, and dementia status as cofactors.

Longitudinal analysis. The longitudinal analysis was restricted to the most reliable whole-brain data and sought to quantify the whole-brain atrophy rate within older nondemented and DAT individuals. Atrophy rate was computed as the slope of the line connecting nWBV measurements within each individual, divided by baseline nWBV, expressed as percent change per year. For example, in a participant with two scans, atrophy rate was computed as nWBV at scan 2 minus nWBV at scan 1, divided by the interval between measurements, divided by nWBV at scan 1, times 100. Analysis of covariance was again used to test for differences in atrophy rate based on age, gender, and dementia status.

Comparison of cross-sectional and longitudinal data. To compare cross-sectional and longitudinal data, atrophy rate was estimated from the cross-sectional, nondemented aging curve and compared with the longitudinal, nondemented atrophy rate. Atrophy rate was estimated from the cross-sectional curve by expressing its slope as a percentage of nWBV at the mean age of the older, nondemented sample (age = 78, nWBV = 74.8%). For graphical comparison, trendlines were plotted for nondemented and demented aging. The slope of the nondemented trendline, for example, was determined by the mean atrophy rate of the nondemented, longitudinal sample; the y-intercept was determined by

25

interpolating nWBV from the nondemented, cross-sectional sample (mean age = 81, nWBV = 74.1%; DAT trendline drawn equivalently, mean age 78, nWBV = 72.0%; see Figure 2.3B).

RESULTS

Cross-sectional. The cross-sectional dataset is plotted in Figures 2.2 and 2.3A and summarized in absolute volumes without head-size normalization in Table 2.2. Nondemented individuals between 18 and 95 years old exhibited an age-associated decline in normalized gray matter (nGM; r = -0.91, F[1,270] = 1311.00, p < 0.001), white matter (nWM; r = -0.25, F[1,270] = 17.71, p < 0.001), and whole-brain volume (nWBV; r = -0.88, F[1,270] = 939.59, p < 0.001). The age-by-volume correlation remained significant when considering the age range of 18 to 30 for nGM (r = -0.20, F[1,121] = 4.91, p < 0.05) and nWBV (r = -0.19, F[1,121] = 4.58, p < 0.05), but not for nWM (F < 1). Considering the full age range, adding a quadratic term significantly improved the models of nGM (F[2,269] = 681.24, p < 0.001, R2 = 0.84), nWM (F[2,269] = 18.96, p < 0.001, = R2 = 0.12), and nWBV (F[2,269] = 533.32, p < 0.001, R2 = 0.80). The addition of a cubic term failed to add a significant effect for any model (F < 1). In addition to the cross-sectional, nondemented aging curves, Figures 2.2 and 2.3A illustrate that individuals with DAT (CDR = 0.5 and 1) exhibited volume reduction disproportionate to age. A full-factorial analysis of covariance on the older sample with age, gender, and dementia status as covariates was significant for nGM (F[7, 184] =

26

20.79, p < 0.001), nWM (F[7.184] = 3.14, p < 0.01), and nWBV (F[7, 184] = 19.80, p < 0.001), with main effects for all three covariates. Post-hoc testing indicated women had more nGM (43.0% vs. 42.2%, p = 0.051), nWM (30.9% vs. 30.1%, p < 0.05), and nWBV (73.9% vs. 72.3%, p < 0.01) than men. The presence of DAT was associated with a significant decrease in nGM (43.7% vs. 41.7%, p < 0.001), nWM (31.1% vs 30.2%, p < 0.05), and nWBV (74.7% vs. 71.9%, p < 0.001; see Figure 2.4A).

Figure 2.2 (next page). Cross-sectional plots of gray and white matter, normalized for head size. (A) Cross-sectional plot of nGM across the adult life-span; each data point represents a unique participant from a single scanning session. For participants with follow-up data, the session with nearest-in-time clinical data is used (see Table 1). The best-fit polynomial regression is drawn only for the nondemented individuals (blue) and is represented by the dashed line for women (55.1 – 0.065[AGE] – 0.001[AGE2]) and the solid line for men (54.4 – 0.076[AGE] – 0.001[AGE2]). (B) Cross-sectional plot of nWM across the adult life-span. The dashed line represents the nondemented, female regression (30.4 + 0.089[AGE] – 0.001[AGE2]); the solid line represents the nondemented, male regression (30.3 + 0.100[AGE] – 0.001[AGE2]). Note the inflection in the nWM curves (around age 42 for men and 45 for women) and the greater separation between nondemented and DAT individuals (red versus blue) in the nGM plot. DAT = dementia of the Alzheimer type; nGM = normalized gray matter; nWM = normalized white matter.

27

FIGURE 2.2 A 60

55

nGM (%)

50

45

40

35

Nondemented, male Nondemented, female

DAT, male DAT, female

30 0 10

20

30

40

50

60

70

80

90

100

70

80

90

100

Age (Years) B 40

nWM (%)

35

30

25

Nondemented, male Nondemented, female

DAT, male DAT, female

20 0 10

20

30

40

50

60

Age (Years)

28

TABLE 2.2. Estimated gray, white, and whole-brain volume (WBV) by age decade in nondemented individuals n

Age ± SD, yrs

Gray ± SD, cm3

18-25

55

21.3 ± 2.0

739 ± 67

442 ± 37

1180 ± 101

26-35

9

28.2 ± 2.8

690 ± 73

420 ± 40

1110 ± 113

36-45

12

40.7 ± 3.0

684 ± 76

427 ± 45

1111 ± 117

46-55

12

50.6 ± 2.6

637 ± 51

426 ± 31

1063 ± 75

56-65

8

62.4 ± 3.0

591 ± 31

409 ± 21

1000 ± 36

66-75

26

70.8 ± 2.5

594 ± 55

416 ± 39

1010 ± 88

76-85

23

81.1 ± 2.5

581 ± 53

409 ± 41

990 ± 86

86-95

15

89.7 ± 2.2

548 ± 63

404 ± 56

952 ± 110

18-25

48

21.9 ± 1.9

794 ± 57

486 ± 34

1280 ± 85

26-35

16

29.2 ± 2.7

756 ± 49

471 ± 34

1226 ± 79

36-45

4

43.8 ± 0.5

729 ± 132

481 ± 81

1209 ± 212

46-55

9

48.8 ± 2.7

740 ± 58

485 ± 42

1224 ± 85

56-65

9

60.7 ± 3.1

706 ± 56

497 ± 53

1203 ± 102

66-75

13

72.5 ± 2.8

670 ± 55

465 ± 45

1135 ± 88

76-85

9

82.2 ± 3.4

647 ± 76

486 ± 91

1133 ± 153

86-95

4

88.7 ± 1.6

594 ± 115

420 ± 63

1015 ± 168

Age decade

White ± SD, cm3

WBV ± SD, cm3

Female

Male

Notes: Cross-sectional, nondemented sample (n = 272). Gray, white, and WBV represent native volumes (without correction for head size). Men have more gray (708.4 cm3 vs. 629.5 cm3, p < 0.001) and white (470.1 cm3 vs. 420.7 cm3, p < 0.001) volume than women.

29

Longitudinal. The longitudinal dataset, obtained in older adults, is plotted in Figure 2.3B. The whole-brain atrophy rate in nondemented older adults was -0.45% (SD = 0.53) per year. The atrophy rate in age-matched individuals with DAT was -0.98% (SD = 1.0) per year. A full-factorial analysis of covariance with age, gender, and dementia status at last scan as covariates, and atrophy rate as the dependent measure, was significant (F[7,71] = 2.16, p < 0.05), with a main effect for dementia status and a significant interaction between age and dementia status. The longitudinal, nondemented atrophy rate of -0.45% per year showed no significant correlation with age within the older sample (r = -0.17, p = 0.30; see Figure 2.4C) and closely matched the atrophy rate estimated from the crosssectional, nondemented aging curve, which varied from -0.31% to -0.46% per year over the same age range.

Figure 2.3 (next page). Cross-sectional and longitudinal plots of whole-brain volume, normalized for head size. (A) Cross-sectional plot of nWBV across the adult life-span. The line represents the best-fit polynomial regression of all nondemented individuals and is referred to as the cross-sectional, nondemented aging curve (85.3 + 0.013[AGE] – 0.002[AGE2]). (B) Longitudinal plot of nWBV in older adults (note scale change); lines connect nWBV at baseline and follow-up scans (or the best fit, for participants with multiple follow-ups), such that the slope of each line as a proportion of baseline nWBV represents an individual’s atrophy rate. The slope of the thick blue line represents the estimated longitudinal rate of change for all of the nondemented individuals and overlaps with the slope of the cross-sectional, nondemented aging curve (shown in black). The slope of the thick red line represents the longitudinal rate of change for all of the DAT individuals and suggests accelerated volume loss in DAT. Lines connected by blue triangles and red squares represent individuals who converted from a CDR of 0 to 0.5 during the inter-scan interval; they are included in the DAT mean. CDR = clinical dementia rating; DAT = dementia of the Alzheimer type; nWBV = normalized whole-brain volume.

30

FIGURE 2.3 A 90

85

nWBV (%)

80

75

70

65

S Nondemented „ DAT

60 0 10

20

30

40

50

60

70

80

90

100

Age (Years) B 85

80

nWBV (%)

75

70

65

S Nondemented „ DAT 60 0 60

65

70

75

80

Age (Years)

31

85

90

95

100

Of the 43 nondemented (CDR 0) individuals followed longitudinally from their first scan, six declined to a CDR of 0.5 at the time of their last scan. Figure 2.4B compares the rate of atrophy in individuals based on their CDR scores at first and last scan. The rate of those who started with a CDR of 0 and declined (-0.88% per year, SD = 0.60) matched the rate of those who started with CDR of 0.5 (-0.90% per year, SD = 0.74). Post-hoc testing revealed a trend toward a difference (t[41] = 3.03, p = 0.09) between the nondemented group (CDR 0 → 0) and the decliner group (CDR 0 → 0.5), though the small sample size limited statistical power.

Figure 2.4 (next page). Summary data. (A) Mean cross-sectional nWBV for individuals 65 and over separated by CDR (0, 0.5, and 1). All differences are significant. (B) Longitudinal atrophy rates, expressed in nWBV loss per year relative to baseline, are separated by CDR status at first and last session. Atrophy rate was significantly greater for the group entering the experiment with very mild dementia (CDR 0.5 → 0.5/1) than the group entering without dementia that remained stable (CDR 0 → 0) and resembles the rate for the group that manifested the earliest signs of DAT during the experiment (CDR 0 → 0.5), though we await confirmation in a larger sample. (C) Individual atrophy rates are plotted versus age, with the trendline drawn through the CDR 0 → 0 group showing minimal acceleration with age. DAT = dementia of the Alzheimer type; CDR = clinical dementia rating.

32

FIGURE 2.4 76

A

nWBV (%)

74

N=94

72

N=69

70

N=29

68

0 66

0.0

B

CDR 0

CDR 0.5

CDR 1

0→0

0 → 0.5

0.5 → 0.5/1

N=6

N=29

Atrophy (%/yr)

-0.2

N=37

-0.4 -0.6 -0.8 -1.0 -1.2

Atrophy (%/yr)

C

1

S0 → 0

‘0 → 0.5

†0.5 → 0.5/1

0

-1

-2 60

65

70

75

80

Age (Years)

33

85

90

95

DISCUSSION

In a large, cross-sectional sample of nondemented adults, significant decline in whole-brain volume was detected in early adulthood and continued into old age, with distinct patterns for gray- and white-matter loss. The cross-sectional rate of decline overlapped the longitudinal rate in the older, nondemented adults. For the longitudinal subset of older adults in the earliest stages of DAT, the rate of whole-brain atrophy (0.98% per year) was more than twice the nondemented rate (-0.45% per year), indicating marked acceleration. These main results are elaborated in terms of the three questions posed in the introduction. The cross-sectional, nondemented aging curve (see quadratic regression, Figure 2.3A) shows that normalized, whole-brain volume (nWBV) declined from 85% at age 20 to 74% at age 80, a lifespan atrophy rate of 0.23% per year, in general agreement with prior studies. Table 2.3 summarizes the results from twelve cross-sectional MR reports on healthy aging that cover the adult lifespan and report whole-brain or gray/white-matter estimates as a percentage of head size. A number other reports are qualitatively similar, but employ quantitatively distinct units that cannot be directly compared to the present results (Pfefferbaum et al., 1994, Harris et al., 1994, Christiansen et al., 1994, Murphy et al., 1996, Raz et al., 1997, Passe et al., 1997, Jernigan et al., 2001, Rovaris et al., 2003, Sullivan et al., 2004, Raz et al., 2004). As the median of all estimates shown in Table 2.3, nWBV declines from 89% at age 20 to 78% at age 80 (median 0.23% per year atrophy). Strong agreement with a recent population-based, volumetric survey of 2081 individuals,

34

age 34 to 97, argues in favor of the generaralizablilty of this sample and these findings (DeCarli et al., 2004). Comparable pathologic estimates fall below the range in Table 2.3; a quadratic regression of volumetric data from one such study suggests that across the reference age range (20-80), brain volume as a percentage of cranial cavity volume decreased from 92% to 85% (0.14% per year atrophy; Davis and Wright 1977). This early study employed a volumetric method involving fluid displacement that did not account for ventricular volume and likely overestimated brain volume and underestimated atrophy.

35

TABLE 2.3. Comparative whole brain estimates from cross-sectional, MR studies Predicted nWBV at age 80, %

Vol decline, %/yr

Age correlation, r

Best-fit regression

Reference

Sample size

Mean age

Predicted nWBV at age 20, %

Jernigan 90

58

45

93

82

-0.21

-0.62

linear

Gur 91

69

43

93

85

-0.15

-0.56

linear

Coffey 92

76

62

Blatter 95

194

39

94

86

-0.15

Matsumae 96

49

56

96

84

-0.23

linear

Guttmann 98

72

59

88

80

-0.17

linear

Courchesne 00

116

21

89

74

-0.28

Good 01

465

30

77

65

-0.27

Van Laere 01

81

44

85

77

-0.16

Ge 02

54

47

90

78

-0.24

Sowell 03

176

32

89

81

-0.16

DeCarli 04

2081

62

86

74

-0.27

89

79

-0.23

85

74

-0.23

Mean Present study

272

47

-0.27

exponential -0.68

-0.57

linear

quadratic

linear

linear -0.63

quadratic

-0.88

quadratic

Notes: nWBV = normalized whole-brain volume, employing head-size correction. Predicted nWBV at ages 20 and 80 was interpolated from regression formulas or plots of whole-brain or gray plus white volume relative to head size as a function of age. Rate of volume decline was estimated by dividing the change in nWBV per year by linearly interpolated nWBV at study age.

36

In addition to quantifying the magnitude of volume decline, the present results converge with others on a temporal sequence placing brain volume reduction at or before the start of early adulthood (Pfefferbaum et al., 1994, Giedd et al., 1999, Courchesne et al., 2000, Rovaris et al., 2003, Sowell et al., 2003, Salat et al., 2004). Whole-brain volume decline was significant within the adult sample when it was restricted to age 18 to 30, although greater volume reductions were noted in the older adults as compared to the young adults. The significant age correlation in this youngest subset argues against a sample contaminated with preclinical AD (i.e., individuals with the pathologic substrate of AD who are not yet sufficiently impaired to be recognized clinically as demented) as the only explanation for atrophy in nondemented older populations. A moderate acceleration of volume loss in nondemented aging occurred in middle age, around the inflection point of the normalized white-matter (nWM) curve at age 44 (see Figure 2.2B). Similar downward inflections (Miller et al., 1980, Harris et al., 1994) after a period of white volume stability (Pfefferbaum et al., 1994, Blatter et al., 1995, Guttmann et al., 1998, Good et al., 2001) or possibly growth (Courchesne et al., 2000, Ge et al., 2002, Liu et al., 2003, Sowell et al., 2003, Jernigan and Fennema-Notestine 2004) during the third and fourth decades have been attributed to the prolonged and heterochronologic development of brain myelination (for review and discussion, see Bartzokis 2004a, Bartzokis 2004b, Sowell et al., 2004). This delayed pattern of nWM loss (though see Van Laere and Dierckx 2001, Sullivan et al., 2004, Raz et al., 2004) contrasts with the more linear course of normalized gray-matter (nGM) decline

37

throughout adulthood, potentially suggesting separate age-related mechanisms for each (Raz 2004). Gender effects were minimal in our results. For the overall cross-sectional sample, which was not gender balanced across age, men had approximately 12% more brain volume than women prior to head-size correction and 0.3% less after head-size correction. The slightly more downward age-course in men than women did not reach significance for nWBV, nGM, or nWM, but tended in the same direction as reported ageby-gender interactions (Christiansen et al., 1994, Murphy et al., 1996, Matsumae et al., 1996, DeCarli et al., 2004, Sullivan et al., 2004). Main gender effects from studies that report no age-by-gender interactions can be compared with the present 0.3% normalized volume difference: four (Blatter et al., 1995, Courchesne et al., 2000, Van Laere and Dierckx 2001, Resnick et al., 2003) report similar gender effects (female > male), three (Guttmann et al., 1998, Ge et al., 2002, Liu et al., 2003) report no gender effect, and three (Raz et al., 1997, Good et al., 2001, Raz et al., 2004) report gender effects in the opposite direction (male > female). The small magnitude of any true difference after head-size normalization, the possibility of differential healthfulness between gender cohorts, and methodological differences likely contribute to inconsistent gender findings. The longitudinal rate of whole-brain atrophy averaged -0.45% per year in the older, nondemented sample. At least seven prior studies have quantified longitudinal, whole-brain volume change in nondemented individuals over a comparable age range. Six (Chan et al., 2001, Wang and Doddrell 2002, Liu et al., 2003, Resnick et al., 2003, Thompson et al., 2003, Jack et al., 2004) have documented annualized rates of about -

38

0.5% (-0.37 to -0.88) , comparable to the present finding, whereas one (Tang et al., 2001) reported a rate of -2.1% (reviewed in Raz 2004). Differences in inclusion/exclusion criteria, scan resolution, and MRI maintenance may explain the divergence of the latter study. Returning to the present findings, the overlap between the longitudinal atrophy rate in nondemented individuals (-0.45%) and the cross-sectional estimate (-0.31 to 0.46%) for the 65 through 95 age-range demonstrates excellent agreement. In addition, longitudinal reports covering the young adult age-range (Giedd et al., 1999, Liu et al., 2003) find slower atrophy rates than in this older, longitudinal sample, suggesting that the trend toward accelerated atrophy with age in Figure 2.4C might reach significance with wider longitudinal sampling or increased sample size. Together, our results indicate that secular effects and other confounds minimally influence cross-sectional, whole-brain volume estimates. For instance, if developmental conditions varied among sampled age cohorts, such that people born in more recent years tended to have increased brain volume in proportion to head size than people born in earlier years, we would expect the slope of the cross-sectional aging curve to exceed the longitudinal slope. A difference might similarly result if aging mechanisms were idiosyncratic and either the longitudinal rates formed a multimodal distribution or sampling characteristics differed among the longitudinal and cross-sectional cohorts. Instead, the observed agreement suggests that the brain loses volume with age according to uniform and predictable, though largely unknown, mechanisms (Resnick et al., 2003).

39

It is well established that individuals with DAT exhibit decreased brain volume relative to their nondemented peers, with the underlying pathology prominent in regions within the medial temporal lobe (Hubbard and Anderson 1981, Braak and Braak 1997, Price et al., 2001). MRI studies have detected accelerated global and regional volume change in DAT (Jernigan et al., 1991, Kantarci and Jack 2003). In addition, comparing nondemented aging with DAT in Figure 2.2 tentatively suggests that gray matter is more vulnerable than white matter to very mild to mild AD pathology (Thompson et al., 2003). Our longitudinal data indicate a -0.98% per year whole-brain atrophy rate in the earliest stages of DAT (CDR 0.5). This rate can be directly compared to recent estimates for nondemented individuals who converted to DAT during follow-up (-0.8%) and those with slow-progressing (-0.6%) and fast-progressing (-1.4%) DAT at baseline (Jack et al., 2004). Faster whole-brain atrophy rates (-5.2%; Thompson et al., 2003) and (-2.4%; Wang et al., 2002) have been reported in smaller cohorts with more advanced AD (baseline MMSEs < 20). Other estimates of longitudinal whole-brain change in DAT derive from the brain-boundary shift integral (BBSI), which models boundary changes in serially-registered scans (Fox and Freeborough 1997). With at least one exception (Cardenas et al., 2003), atrophy rate estimates based on the BBSI have exceeded -2% per year (Fox and Freeborough 1997, Chan et al., 2003), for example, a -2.37% BBSI atrophy rate in 54 DAT patients (Chan et al., 2001). Divergence in longitudinal atrophy rates may reflect differences in atrophy measurements and DAT cohorts; the DAT sample yielding the -2.37% estimate had a lower age (61 versus 79) and MMSE (20 versus 26) than in the present study and included early-onset and familial cases.

40

Providing evidence that the specific DAT sample represents an important factor, in a sample of 5 “preclinical” familial cases followed during their conversion to DAT, the atrophy rate was found to be -1.23% per year using the BBSI and -1.08% per year using TIV correction, very similar to the rates reported here (Schott et al., 2003). The estimates here and in the literature suggest that non-familial, late-onset DAT is characterized by at least a 1% per year volume loss in its earliest clinical presentation with acceleration as the disease progresses (Chan et al., 2003). This initial rate of brain volume loss represents a doubling from that of nondemented individuals. Our results tentatively suggest that accelerated loss in whole-brain volume may begin in the preclinical phase of AD (DeKosky and Marek 2003). In particular, the small sample of nondemented individuals who declined over the course of our observation period showed accelerated atrophy rates similar to the individuals with very mild DAT at baseline. The reliability, cross-sectional validity, and sensitivity to clinical progression of automated whole-brain measures such as nWBV, combined with their potential to detect preclinical AD, highlights the promise of global volumetric biomarkers (Jack et al., 2004).

41

CHAPTER 3 SUMMARY OF THREE-MONTH STUDY TO DISCRIMINATE ATROPHY IN AD FROM NONDEMENTED AGING

Fischl and colleagues (2004) recently reported a model for magnetic resonance (MR) image analysis that relates the probability of misclassifying a voxel in a segmentation procedure (Fischl et al., 2002) to the acquisition parameters (repetition time [TR], echo time [TE], and flip angle) used to sequence acquisition of fast low-angle shot (FLASH) MR images. This theoretical advance of embedding MR physics into morphometric algorithms enabled the development of a new multiecho FLASH (MEF) sequence optimized for segmentation. The motivation for developing MEF was that it might lead to practical advances in the structural characterization of neurodegenerative disorders. We tested whether MEF could advance the efficiency of longitudinal studies in Alzheimer’s disease (AD) in our three-month study, summarized in this chapter. A follow-up period of three months represents the shortest interval reported to date relative to comparables studies that have aimed to discriminate atrophy in AD from normal aging (Bradley et al., 2002, Schott et al., 2005). Baseline MR images were obtained from 30 participants enrolled in the ongoing studies of the Washington University AD Research Center (ADRC). These participants were evenly divided among a nondemented (CDR 0) group and a dementia of the Alzheimer’s type (DAT, CDR 0.5/1) group, based on their Clinical Dementia Rating (CDR, Morris 1993), as previously described (Fotenos et al., 2005). Follow-up images

42

were acquired a mean 95 (SD = 11) days after baseline. Ten additional young adults were scanned twice within a short interval (mean 11 days) to allow estimation of measurement reliability. Sample characteristics are summarized in Table 3.1. TABLE 3.1. Longitudinal sample with complete baseline and follow-up MEF images Test-retest

CDR 0

DAT (CDR 0.5) DAT (CDR 1)

Number

10

15

9

6

Female/male

6/4

10/5

4/5

3/3

Age ± SD, y

27 ± 10 (19-54)

80 ± 10 (63-95)

78 ± 5 (72-85)

75 ± 4 (70-79)

Education ± SD, y

16 ± 3 (12-23)

15 ± 2 (12-18)

15 ± 2 (12-16)

MMSE ± SD

29 ± 1 (26-30)

26 ± 3 (21-30)

19 ± 4 (15-23)

96 ± 13 (84-129)

92 ± 10 (83-115)

96 ± 7 (91-109)

Retest interval ± SD, d

12 ± 11 (3-29)

The imaging protocol ran approximately 40 minutes on a 1.5 T Siemens Sonata (van der Kouwe 2005). The protocol consisted of an initial fast FLASH sequence (TR = 2.4 ms, TE = 1.13 ms, resolution = 3.3 x 2.5 x 2.5 mm, time = 0:48 min) for online rigidbody registration to a template, thus ensuring uniform head positioning during all subsequent scans. Standard FLAIR (TR = 1000, TE = 1.13, 3.3 x 2.5 x 2.5, 3:42) and MPRAGE (TR = 2730,TE=3.41, 1.3 x 1.0 x 1.3, 8:46) runs followed. Next came the two main MEF sequences (TR = 20; TE = 1.8, 3.62, 5.44, 7.26, 9.08, 10.9, 12.72, 14.54, 1.3 x 1.0 x 1.3, 8:12), the first T1-weighted at a flip angle of 30 (MEF30) and the second proton-density weighted at a flip angle of 5 (MEF5). Finally, a magnetization transfer

43

corrected MEF5 and two more fast FLASH sequences were acquired, one with the head coil and the other with the body coil, for use in correcting B1 inhomogeneity. The multigigabyte MEF k-space data was reconstructed off-line, and the resulting 16 images from the main MEF30 and MEF5 runs, along with pre-processing output, are shown in Figure 3.1. FIGURE 3.1

B A

C

D

E

F

Figure 3.1. Multi-echo FLASH (MEF) imaging from one older adult participant. (A) The eight echoes from the T1-weighted MEF30 sequence (top row) and protondensity (PD) weighted MEF5 sequence (bottom row) show intensity decay as echotime increases to the right, reflecting dephasing of transverse magnetization (T2*). (B) Weighted average of the MEF scans. (C) Synthesized T1, (D) T2*, and (E) PD maps. (F) Standard, 8-minute structural image (MPRAGE), acquired as a control in the imaging protocol.

44

Preprocessing of the MEF images acquired at each imaging session consisted of four main steps: 3-D gradient unwarping; rigid registration of MEF5 to MEF30 images; weighted averaging; and voxel-wise T1, T2*, and PD parameter estimation. Gradient unwarping corrected for geometric distortion caused by nonlinearities in the gradient coil; the 3-D implementation improved on the 2-D algorithm installed on the Siemens Sonata. Head movement between acquisition of the MEF30 and MEF5 scans was corrected by 6paramater affine transformation (Smith et al., 2004). The transformed images were then averaged based on weightings derived from a linear discrimination analysis that maximized contrast-to-noise between manually segmented gray and white voxels (Han 2005). As an alternative method of dimensionality reduction, maps of relaxometry parameter estimates were synthesized based on the Bloch equation and changes in TE and flip angle (α) in the MEF sequence. The Bloch equation specifies the relationship between image intensity and relaxometry parameters as follows:

S (m, β ) = PD sin α

1 − e −TR / T 1 e −TE / T 2* , −TR / T 1 1 − cos α e

(0.1)

where S = image intensity, β = [T1, T2*, PD], and m = [TR, TE , α ] . Morphometric analysis started with the resulting MEF images (see Figure 3.1BE). Figure 3.2 summarizes the main analytical strategies and results. Three-month change in CDR 0 and DAT groups was compared in terms of whole brain and regional atrophy and relaxometry parameters. In preparation for these analyses, the three-month data was

45

registered to the baseline using a nine-parameter affine transform (translation plus rotation plus stretch), and this transform was algebraically composed with the baselineto-atlas transform (12-parameter affine) so that each participant’s structural data was corrected in atlas space for serial differences in head position and scanner drift (Freeborough et al., 1996, Snyder 1996). The atlas template was composed of a subsample of combined young-and-old data in the atlas space of Talairach and Tournoux (1988), with construction as described previously (Buckner et al., 2004). Normalized whole brain volume (nWBV) at baseline and follow-up was computed using the segmentation workflow described in Figure 2.1. In addition, perpendicular displacement between sessions at the edge of the segmented brain image was integrated for a direct measure of global atrophy (Smith et al., 2002, Smith et al., 2004). Maps of regional atrophy were derived from adding a nonlinear registration (fluid warping) step after the rigid body registration, as previously described (Buckner et al., 2005). Fluid warping aims to transform the follow-up image so that the difference between follow-up and baseline is minimized (Christensen et al., 1996, Miller 2004). The divergence of the intersession deformation at each voxel represents local contraction or expansion, and, when divided by the intersession interval, has units of fractional loss or gain per year. Divergence maps were compared between CDR 0 and DAT groups using voxel-wise t-tests. The maps of relaxometry estimates, and relaxometry change maps, were also compared using t-tests, as well as by visual inspection of individual intensity distributions.

46

The global atrophy (Figure 3.2A, right), regional atrophy (Figure 3.2B, right), and relaxometry analyses (Figure 3.2C) all showed no significant difference for three-month change between CDR 0 and DAT groups. Test-retest error (Figure 3.2A, left), computed as the mean absolute volume change (as a percentage of whole brain volume) between scan pairs just days apart, was 0.24% for MPRAGE pairs acquired as described in Chapter 2 (n = 23), 0.27% for MPRAGE pairs acquired as part of the present MEF protocol, and 0.25% for the MEF-average pairs (p = 0.93). Analysis of baseline whole brain volume (not shown) indicated that DAT was associated with a marginal decrease (nWBV 69% vs. 67%, p = 0.06). In contrast, change estimates over the three-month interval substantially overlapped between groups, with global atrophy rates of 1.2%/yr in the CDR 0 group and 1.4%/yr in the DAT group (p = 0.77). Similarly, three-month regional (voxel-wise) atrophy (see Figure 3.2B, right) did not show the pattern of parietal and medial temporal discrimination over years-long intervals reported elsewhere for DAT (Scahill et al., 2002, Toga and Thompson 2003). This convergent pattern is reprinted for reference from our lab (Buckner et al., 2005) along with unpublished t-maps for standard MPRAGE images acquired in 44 CDR 0 and 40 DAT participants over a mean two-year interval (Figure 3.2B, left). Finally, the distribution of T1, T2*, and PD estimates in participant’s relaxometry maps also substantially overlapped (see Figure 3.2C), and these estimates had low reliability (for example, > 2% mean absolute percentage difference for whole brain T1; see also DeMyer et al., 1988, Wang et al., 2004).

47

Figure 3.2 (next page). Post-processing and null results for three-month discrimination with multi-echo FLASH (MEF). (A) Whole brain atrophy in terms of mean absolute estimates of error from test-retest sampling (left) and annualized change from three-month sampling (right). (B) Regional discrimination of DAT and CDR 0 groups over two years of follow-up from Buckner and colleagues (left; Buckner et al., 2005) with no discrimination over three months (right). (C) Overlapping distributions of T1, T2*, and PD estimates from baseline relaxometry maps.

48

FIGURE 3.2 A 0.4

CDR 0

0.0

Atrophy (%/Yr)

Error (%)

0.3

0.2

0.1

CDR 0.5/1

-0.5

-1.0

-1.5

-2.0

0

Standard MPRAGE B

-2.5

Standard MPRAGE CDR 0 Difference

DAT

Multi-echo FLASH CDR 0 Difference

Z=-22

Z=-22

Z=38

Z=38

DAT

New Multi-echo MPRAGE FLASH

0.3

Atrophy (%/Yr)

1.0 2.0

t-stat

5.0

0.3

Atrophy (%/Yr)

1.0 2.0

5.0

Young

C

Count

t-stat

180

180

100

160

160

90

140

140

80

CDR 0 DAT

70

120

120

100

100

80

80

60

60

40

40

20

20

20

10

60 50

0 0

1000

T1 (ms)

2000

0 3000 0

40 30

0 4000 50

5000

100

T2* (ms)

0 150

3000

6000 200

9000

12000

Proton Density

49

15000

18000

Results from this study suggest that structural MRI based on the newly developed MEF sequence does not lead to ready improvement in standard volumetric estimates, such that three-month change in nondemented vs. demented samples overlapped. This null result converges with limitations encountered in related work on the application of the MEF sequence within the laboratory responsible for its innovation (Xiao Han, personal communication, July 11, 2005). In contrast, a recent report based on standard structural imaging in more advanced dementia (corresponding to DAT with a CDR of 1 to 2, n = 38) found that interval change over 6-12 months, particularly in the volume of the lateral ventricles, was sufficient for discrimination from nondemented aging (Schott et al., 2005; see also, Gunter et al., 2003). This finding suggests that sampling and processing methods, as much as or more than image acquisition, may be key to improving the efficiency of structural imaging studies. We have noted the possibility, for example, that unreliability in the processing steps prior to segmentation (such as in masking, registration, resampling, or corrections for field distortions) may be a main contributor to test-retest error, since the mean absolute percentage difference for estimated total intracranial volume (eTIV), computed prior to segmentation, and for normalized whole brain volume (nWBV), computed after, are both around 0.5% (Buckner et al., 2004, Fotenos et al., 2005). Simply put, the complexity of volumetric estimation is likely an important factor in measurement error. Based on the result with ventricular change noted above, we suggest that a promising direction for future research is to develop a method for the single-step measurement of ventricular volume from unprocessed MRIs. Such a specialized tool could take advantage of the high contrast and

50

stereotypy of ventricular CSF, and possibly also multi-spectral information, such as offered by the MEF.

51

CHAPTER 4 BRAIN VOLUME DECLINE IN AGING: EVIDENCE FOR A RELATIONSHIP BETWEEN SES, PRECLINICAL AD, AND RESERVE

ABSTRACT

Objective: To explore the influence of socioeconomic status (SES) on structural brain aging in nondemented older adults, including the role of preclinical Alzheimer’s disease (AD). Methods: Head-size adjusted whole brain volume (aWBV) was estimated from MRI in 362 individuals age 18 to 93. Of these, SES was assessed in a main cohort of 100 nondemented older adults age 65 to 93 (Clinical Dementia Rating [CDR] 0 at initial MRI) using the Hollingshead two-factor index of social position. 83 of these participants received follow-up clinical assessment subsequent to MRI, and 33 were imaged longitudinally. To test whether preclinical AD can influence brain volume, MRI data were analyzed from a sample of 58 CDR 0 participants (age 47 to 86) who had participated in amyloid imaging with [11C]Pittsburgh Compound-B (PIB). Results: aWBV was estimated to decline by 0.22%/yr between the ages of 20 and 80, with accelerated decline in advanced aging. In older adults > 65, increasing SES was associated with smaller aWBV (3.8% difference spanning the sample range from middle to high privilege) and more rapid volume loss (0.39%/yr to 0.68%/yr from middle to high privilege). Supporting an influence of preclinical AD, aWBV was reduced by 2.5% in

52

individuals positive for PIB binding (n = 9) as compared to individuals negative for PIB binding (n = 49). Follow-up clinical data revealed the volume reduction associated with SES was greater in those who developed signs of very mild dementia subsequent to MRI (preclinical dementia group, n = 19) compared to those who remained nondemented (stable CDR 0 group, n = 64). Conclusions: Brain volume loss is accelerated in privileged nondemented older adults relative to less privileged peers. The capacity of more privileged individuals to cope longer with preclinical pathology prior to disease expression, as predicted by the cognitive reserve hypothesis, may contribute to this effect.

53

INTRODUCTION

Brain volume decline is characteristic of life-long processes that begin in adolescence, or earlier, and continue into advanced aging (Giedd et al., 1999, Courchesne et al., 2000, Raz et al., 2005). Relevant to Alzheimer’s disease (AD), volume loss accelerates markedly in the earliest stages of the disease (Fox and Schott 2004, Jack et al., 2004, Fotenos et al., 2005). These and related observations lead to the conclusion that whole brain volume is determined by a constellation of factors that include normal developmental change as well as pathological processes (Buckner 2004, Raz 2004, Hedden and Gabrieli 2005). Measurement of whole brain volume differences between individuals and change within individuals provides an opportunity to better understand modifiers of normal and pathological aging. Here we sought to explore the influence of socioeconomic status (SES) on brain volume in a large cohort of nondemented older adults and to supplement this analysis with subject stratification based on amyloid binding measured with [11C]Pittsburgh Compound-B (PIB) and longitudinal clinical assessment. The primary motivation is the observation that high SES, representing both educational and occupational attainment, is a protective factor for AD (Zhang et al., 1990, Stern et al., 1995, Bennett et al., 2003, Amieva et al., 2005; reviewed in Scarmeas and Stern 2004, Valenzuela and Sachdev 2006). It is unknown to what degree the influence of SES reflects modification of disease-associated processes (Snowdon et al., 2000), antecedent capacity to buffer pathology (Katzman 1993, Satz 1993), active capability to compensate for the disease

54

(Stern 2002, Stern 2006), ascertainment bias (Tuokko et al., 2003) or any combination of these possibilities. Analysis of whole brain volume including its relation to amyloid binding (as a proxy measure for AD pathology) and longitudinal clinical evaluation (as a measure of dementia onset) may provide insight. Direct precedent for this research comes from Coffey and colleagues (Coffey et al., 1999), who observed that years of formal education may amplify volume decline in nondemented older adults. Specifically, the more educated individuals showed indications of reduced brain volume, based on a measure of peripheral cerebrospinal fluid (CSF) adjusted for age and head size. An intriguing possibility is that socioeconomically privileged individuals are better able to respond to the initial stages of neurodegenerative disease and their impairment does not reach a level that is clinically apparent. This notion of neuropsychological mediation between brain pathology and disease expression is referred to as cognitive reserve (Stern 2002, Stern 2006). Relevant to the possibility of clinically silent disease, we previously found that in vivo measurement of amyloid via PIB binding reached levels suggestive of AD pathology in some nondemented participants (Clinical Dementia Rating [CDR] 0; Buckner et al., 2005, Fagan et al., 2006, Mintun et al., 2006). Post-mortem, 34% (14/41) of participants from our center who were nondemented at expiration nevertheless met established histologic criteria for AD (Galvin et al., 2005). The discrepancy between post-mortem AD pathology and symptomatic disease expression is well replicated (Rothschild 1937, Katzman et al., 1988, Price and Morris 1999, Esiri et al., 2001, Bennett et al., 2006, Lippa

55

and Morris 2006; reviewed in Mortimer et al., 2005), and the term “preclinical AD” describes the initial period of disjunction. The preclinical period may relate to cognitive reserve, with longer preclinical periods expected to track putative markers of reserve, such as educational and occupational attainment, literacy, IQ, and active lifestyle. The literature on incident dementia offers indirect evidence for this reserve hypothesis, finding that higher reserve variables predict less frequent disease expression (reviewed in Fratiglioni et al., 2004, Valenzuela and Sachdev 2006). More direct evidence has come from recent studies that collected pathological measures at autopsy together with premorbid variables on reserve and disease expression; the reserve hypothesis predicts that increasing reserve raises the pathological threshold for impairment, and the results of these clinicopathologic studies support this (Bennett et al., 2003, Bennett et al., 2005). In the present study, we sought to address three questions concerning the possibility of SES influences on brain volume. First, in nondemented older adults carefully screened for dementia, is brain volume associated with SES? We used the Hollingshead two-factor index of social position, which includes years of formal education, but also weights for occupational attainment over the extended period between schooling and retirement (Hollingshead 1957). The dependent measure was a reliable, validated metric of whole brain volume (Buckner et al., 2004, Fotenos et al., 2005), and the sample was screened for dementia based on the sensitive, informant-based CDR scale (Morris 1993, Carr et al., 2000, Storandt et al., 2006). Extrapolating from the reported result on education (Coffey et al., 1999), we hypothesized that individuals with a life

56

experience of higher SES (more combined educational and occupational attainment) would correspond to those with less brain volume. Second, do brain volume differences related to SES reflect different rates of longitudinal brain volume change? Cross-sectional volume differences may represent baseline differences, independent of aging. Here, based on the subgroup of our nondemented sample with follow-up MRI over an extended interval, we explored the association between longitudinal brain aging and SES. Extending our cross-sectional prediction, we hypothesized that brain volume loss within individuals over time would be more rapid in the more privileged. Third, does cognitive reserve explain the hypothesized structure-privilege relationships? An explanation for smaller brain volume in privileged individuals may be that they are more likely to experience atrophy associated with preclinical AD (Coffey et al., 1999). A group with preclinical AD and other brain pathologies may be contained within the nondemented sample; in this subgroup, atrophy is expected to advance longer prior to disease progression in individuals with more reserve. We explored this possibility by examining whether a substantial number of individuals with presumptive evidence of AD pathology (PIB+ binding) were contained within our nondemented population and further whether brain volume differs in these individuals compared to PIB- peers. Based on the hypothesis that preclinical AD is a cause of reduced brain volume in the most privileged individuals, we further analyzed longitudinally measured clinical conversion (from no dementia to very mild dementia) to explore whether the interaction of SES with clinical conversion accounts for the brain volume reduction. Support for the reserve

57

hypothesis would come from finding that brain volume loss was accounted for by privileged individuals on the verge of clinically detectable dementia.

METHODS

Participants. MRI images from 362 individuals (age 18 to 93) were obtained from participants in the ongoing longitudinal studies of the Washington University Alzheimer Disease Research Center (ADRC) and our ongoing studies of normal aging and development (Berg et al., 1998, Galvin et al., 2005). More detailed attrition and selection characteristics of this population have been described previously (Fotenos et al., 2005). All participants were scanned with identical procedures. A subset of 100 nondemented older adults (age 65 to 93) comprised the main cohort of clinically screened individuals for which extensive data analysis was performed. Of these 100, 33 were followed by MRI for an extended interval to allow for longitudinal data analysis (mean = 3.1 times over a 3.1 to 6.5 year interval; mean = 4.3 years). MR participants were classified as nondemented if their CDR nearest the time of baseline MRI was 0. Clinicians determined the CDR, blind to the results of neuropsychological testing and prior clinical assessment, through examination of the participant and interview with an informant (usually a family member) who knew the participant well and could provide information regarding decline from the participant’s normal cognitive and functional abilities (Morris 1993). Designation of CDR 0.5 (very mild dementia) thus indicates early clinical impairment relative to an individual’s

58

baseline, in contrast to impaired test performance relative to group norms in mild cognitive impairment (Morris et al., 2001, Storandt et al., 2006). The orientation of the CDR to intraindividual change helps to avoid mistaking low baseline functioning for dementia or, conversely, mistaking no dementia for high baseline functioning in the presence of AD (ascertainment bias; Tuokko et al., 2003). As an example of the CDR’s sensitivity to early symptomatic AD, a recent report from our center found that over one third (276 of 728) of a CDR 0.5 sample had neuropsychological test scores that did not fall below the required cutoffs for MCI (pre-MCI group), despite the presence of individual decline and neuropathological AD (43 of 47 in the autopsied pre-MCI group; Storandt et al., 2006). In the present study, the mean duration between clinical assessment and baseline MRI was 116 days (range = 0 to 314 days). Participants were paid for their participation and gave informed consent in accordance with guidelines of the Washington University Human Studies Committee. Data from some of these participants have been reported in previous studies (Salat et al., 2004, Head et al., 2005, Fotenos et al., 2005, Buckner et al., 2005, Burns et al., 2005). SES was assessed for all older adults at a participant’s initial clinical evaluation, using the Hollingshead two-factor index of social position (Hollingshead 1957). The Hollingshead index represents a linear combination of educational and household occupational attainment, with occupation almost doubly weighted. The index ranges from 11 to 77 and can be grouped into five SES categories (I-V). In order to control for potential health confounds related to deprivation in the underprivileged (House et al.,

59

1994; reviewed in House and Williams 2000) and because the cohort from which data were drawn contains too few low-SES individuals to disentangle these effects, this study focused on variation of the Hollingshead index within the range of the high-privilege, high-middle, and middle SES groups (I-III; Fratiglioni et al., 1999). None of the health variables presented in the table of sample characteristics (Table 4.1) differed between these groups.

60

TABLE 4.1. MRI sample Young

N (cross-sectional)

Middle Aged

Nondemented Old (CDR 0) Middle SES

High-Middle SES

High SES

127

135

31

40

29

67/60

91/44

21/10

30/10

16/13

26/5

35/5

28/1

77±8 (65-90)

77±7 (65-93)

78±7 (65-92)

Hollingshead ± SD

35±4 (29-43)

22±3 (18-27)

13±2 (11-15)

MMSE ± SD

29±1 (25-30)

29±1 (26-30)

29±1 (26-30)

Prescriptions, n

2.8±1.8 (0-8)

2.6±2.1 (0-7)

2.7±1.8 (0-6)

Systolic BP, mmHg

134±17 (110-192)

139±18 (110-190)

131±15 (102-158)

Diastolic BP, mmHg

73±12 (50-94)

75±9 (56-96)

69±10 (40-90)

159±36 (100-245)

159±27 (112-220)

158±26 (121-206)

11

15

7

6/5

9/6

5/2

Duration of MRI follow-up ± SD, yrs

4.4±1.1 (3.3-6.3)

4.3±0.7 (3.1-5.8)

4.4±1.0 (3.1-6.5)

MRIs per participant ± SD, n

3.1±0.7 (2-4)

3.0±0.7 (2-4)

3.1±0.8 (2-5)

Female/Male Sum Box Score 0/0.5 Age ± SD, yrs

23±3 (18-34)

52±7 (35-64)

Weight (sex-adjusted), lbs N with follow-up (longitudinal) Female/Male

Notes: Data are shown corresponding to the earliest imaging session (baseline). SES is grouped in order of increasing privilege based on published cutoffs for the Hollingshead two-factor index of social position. The Sum Box Score represents a more quantitative form of the global CDR based on the sum of ratings in

61

each of six domains assessed by the CDR. A Sum Box Score of 0.5 indicates very mild impairment in one domain other than memory; all participants had a global CDR of 0, indicating no dementia. Mean values are given ± the standard deviation. Values in parenthesis represent the range. MMSE = Mini-Mental State Examination where scores range from 30 (best) to 0 (worst); BP = blood pressure; CDR = Clinical Dementia Rating; SES = socioeconomic status. Estimation of whole brain volume. Our method of image acquisition and brain volume estimation has been described previously (Buckner et al., 2004, Fotenos et al., 2005). Briefly, multiple (three or four) high-resolution structural T1-weighted magnetizationprepared rapid gradient echo (MP-RAGE) images were acquired on a 1.5-T Vision scanner (Siemens, Erlangen, Germany). Repetition time was 9.7 ms, echo time was 4 ms, flip angle was 10°, inversion time was 20 ms, and resolution was 1 × 1 × 1.25. Image processing involved the following fully automated steps: within-participant averaging, atlas registration, resampling to isotropic voxels (1 mm cubic), correction for intensity inhomogeneity (B1 bias-field), skull masking, and segmentation into gray, white, and CSF compartments. The registration was based on computation of a 12paramater affine transformation aimed at minimizing variance between the first MPRAGE and a combined young-and-old template in the atlas space of Talairach and Tournoux (Talairach and Tournoux 1988, Snyder 1996, Buckner et al., 2004). Estimated total intracranial volume (eTIV) was computed based on the atlas transformation. Specifically, the atlas scaling factor (ASF), which represents the determinant of the transformation matrix, is highly correlated with manually measured total intracranial volume (TIV, r = 0.93) and minimally biased by atrophy (Buckner et al., 2004). The inverse of the ASF was used as the eTIV, as in our earlier report (Fotenos et al., 2005). The segmentation used a validated algorithm for computing the maximum likelihood 62

estimates of a hidden Markov, random field model, constrained by both intensity and spatial proximity parameters (Zhang et al., 2001, Smith et al., 2004). Whole brain volume (WBV) was computed as the sum of gray and white compartments. Head size differences were corrected using a covariance procedure (as opposed to ratio normalization) in order to eliminate the possibility of shared denominator variance introducing spurious associations between corrected volume estimates and covariates of interest (Mathalon et al., 1993, Buckner et al., 2004). The term adjusted whole brain volume (aWBV) is used to denote covariance-adjusted volumes, as distinct from proportionally normalized whole brain volume (nWBV). The formula for aWBV, adjusted for head size, follows:

aWBV = WBV – b(eTIV – mean eTIV)

where WBV is the uncorrected (native) whole brain volume, b is the slope of the volume regression on eTIV, eTIV is the ASF-derived head size estimate, and mean eTIV is the sample mean. In instances where the influences of multiple variables on WBV were being explored simultaneously, eTIV was always entered as a covariate and the dependent variable is denoted as aWBV to reflect this adjustment.

Cross-sectional analysis. To explore differences in brain volume across the full life-span, aWBV was plotted cross-sectionally versus age for the entire sample of 362 individuals, including the cohort of 100 clinically screened nondemented older participants (age 65 to

63

93) and the young and middle-aged volunteers from the community (age 18 to 64), who participated in MRI under identical conditions. Statistical analysis was conducted with both JMP and SAS software packages (SAS Institute, Cary, NC). Analysis of covariance and hierarchical polynomial regression were used to model aWBV as a function of age and sex. To test for a cross-sectional relationship between SES and brain volume, analysis was restricted to the main carefully screened older adult sample of 100, and SES was entered as the predictor variable, with age and sex as covariates. We limited the analysis to whole brain estimates because they are almost twice as reliable as separated gray and white estimates and avoid potential confounds involving white-matter intensity changes (Jernigan et al., 2001, Fotenos et al., 2005).

Longitudinal analysis. To test for a longitudinal relationship between SES and brain volume, we used multilevel modeling (SAS PROC MIXED, full maximum likelihood estimation) with aWBV as the dependent measure and the time-by-SES term as the predictor; covariates were baseline age, time (expressed as years from baseline), SES, and sex. Multilevel modeling handles intrinsically correlated within-individual data of uneven number and spacing more sensitively than ordinary-least-squared (OLS) slope analysis (Singer and Willett 2003). For visualization, however, the most precise OLS regressions of aWBV against time were plotted per individual, with individuals ranked by SES (via the Hollingshead index).

64

Preclinical Alzheimer’s disease. As will be shown, SES exerts an influence on brain volume with the most privileged individuals showing reduced brain volume (crosssectional data) and accelerated volume loss (longitudinal data). The reserve hypothesis to explain this counterintuitive pattern is that preclinical AD exerts an influence on brain volume with the more privileged individuals harboring preclinical AD longer or more frequently than in the less privileged. To explore the role of preclinical AD on brain volume and its relation to SES, two sets of additional data were analyzed. The first analysis, based on amyloid plaque imaging, explored whether preclinical AD could exert an influence on brain volume. The second analysis, based on follow-up clinical data, explored whether an interaction between SES and preclinical dementia status influenced brain volume. Visualization of amyloid was enabled by PIB, a radiotracer with high affinity for amyloid in Aβ plaques (Klunk et al., 2004). As part of the larger research program at Washington University, PIB was imaged with PET in a sample of 58 nondemented ADRC participants that partially overlapped with the main sample. Characteristics of the PIB sample are described in Table 4.2; 28 of those with PIB imaging overlapped the main cohort described in Table 4.1. The rest were either younger, less privileged, or missing SES data. Those not within the main cohort nonetheless received identical MRI. The combination of PIB-PET and structural MRI allowed us to estimate the percentage of our nondemented sample that harbored a putative sign of AD pathology and to test whether CDR 0 individuals with high PIB uptake (PIB+) exhibit brain volume reduction relative

65

to peers with minimal PIB uptake (PIB-). This analysis establishes whether AD in its preclinical phase is capable of exerting an influence on brain volume. TABLE 4.2. PIB amyloid imaging sample CDR 0 PIB-

CDR 0 PIB+

49

9

Female/Male

39/10

7/2

Box score 0/0.5

46/3

9/0

Age ± SD, yrs

69±11 (47-86)

72±7 (61-81)

Education ± SD, yrs

16±3 (11-20)

14±3 (11-18)

MMSE ± SD

29±1 (26-30)

29±1 (26-30)

159±26 (114-237)

134±26 (118-176)

N (cross-sectional)

Weight (sex-adjusted), lbs

Notes: Preclinical AD, as suggested by imaging of [11C]Pittsburgh Compound-B (PIB) in a separate sample, was explored for potential contributions to structural MRI findings. Positive/negative groupings (PIB-/+) were based on mean regional PIB uptake, as described in the text. Abbreviations and format are the same as Table 4.1. The PIB+ group weighed less than the PIB- group after adjusting for sex (t[53] = 2.63, p < 0.05). There were no other significant group differences, including for additional clinical variables in Table 4.1 (not shown). PIB-PET image acquisition and analysis are detailed elsewhere (Buckner et al., 2005, Fagan et al., 2006, Mintun et al., 2006). Briefly, 10 mCi PIB, synthesized according to published methods (Mathis et al., 2003), was injected into the antecubital vein of participants resting eyes-closed in a 961 HR ECAT PET scanner (CTI, Knoxville, TN). Images were acquired in 3-D and reconstructed into 5-min frames (septa withdrawn;

66

with scatter correction and a ramp filter; ~5.5-6 mm full width half maximum) over a 60min scanning interval. Frames were motion corrected and atlas registered via composition of affine transforms of PET to MRI to atlas (Snyder 1996). PIB uptake in four brain regions (prefrontal, lateral temporal, precuneus, and gyrus rectus) was obtained by manual drawing of ROIs on the coregistered MRI and application to the dynamic PET data. Binding potential (BP) was calculated using Logan graphical analysis with a cerebellar reference ROI (Logan et al., 1996; for ROI descriptions, see Fagan et al., 2006). A mean BP for these four regions greater than 0.2 was used to classify individuals with higher relative cortical binding as PIB+, based on the demonstrated association between CDR, CSF amyloid-β42, and quantitative PIB uptake (Fagan et al., 2006). Baseline MRIs from individuals classified as PIB+ were then compared against PIBMRIs, in a separate analysis of the PIB sample using aWBV as the dependent measure. Age and sex were covariates. For the second analysis, preclinical dementia was assessed by examining the longitudinal history of clinical examinations. A large overlapping clinical sample (in contrast to the limited overlapping PIB sample) allowed us to explore interactions with SES. Specifically, 83 of the 100 participants characterized in Table 4.1 as CDR 0 around the time of baseline MRI received at least one subsequent clinical evaluation (mean 2.9, range 1 to 6; mean follow-up interval of 3.0 years, range 0.5 to 6.4 years). Participants were grouped relative to their initial MRI as preclinically demented if they received a CDR of 0.5 at any subsequent clinical examination. Group status (preclinical versus

67

stable CDR 0) was added as an additional term in the cross-sectional analysis of SES described above.

RESULTS

Brain volume is reduced in nondemented aging. Cross-sectional brain volumes in nondemented individuals, age 18 to 93, are illustrated in Figure 4.1 (using covarianceadjusted whole brain volume; aWBV). Parameter estimates for age, age2, sex, and ageby-sex were all significant in the model (F[5,356] = 1394.14, p < 0.001, R2 = 0.95). For a sense of effect size, aWBV can be compared at age 20 and age 80, with estimated declines from 1199 cm3 to 1025 cm3 in men and from 1195 cm3 to 1050 cm3 in women (decline in annualized percentage terms, 0.24%/yr and 0.20%/yr respectively). Note that initial aWBV is almost identical between men and women, reflecting the adjustment’s ability to accommodate head size differences (Buckner et al., 2004). The significance of the quadratic age term reflects the acceleration of volume decline in advanced aging. Considering the age range between 65 and 80, the estimated declines were 0.40%/yr for men and 0.35%/yr for women.

68

FIGURE 4.1 1250

Adjusted Whole Brain Volume (cm3)

1200

1150

1100

1050

1000

950

900 0 10

20

30

40

50

60

70

80

90

100

Age (Years)

Figure 4.1. Cross-sectional plot of brain volume in nondemented adults over the adult life-span. Adjusted whole brain volume (aWBV, shown adjusted for head size) declines 13% (0.22%/year) between age 20 and age 80; deviation from a linear timecourse is mild, though significant. Privileged older adults have reduced brain volume. Figure 4.2A focuses on the role of SES as a potential modifier of whole brain volume among nondemented older adults (age > 65 yr). After accounting for effects of age, sex, and age-by-sex on aWBV (model F[5,94] = 218.74, p < 0.001, R2 = 0.92), more privileged individuals were associated with lower volume estimates (β = 1.3 cm3 per Hollingshead unit, p < 0.01). For example, spanning the sample range from middle privilege (Hollingshead = 43) to highest privilege

69

(Hollingshead = 11), aWBV was estimated to decrease from 1066 to 1026 cm3 (3.8% difference).

Privileged older adults show accelerated longitudinal volume loss. To determine whether cross-sectional differences associated with SES relate to aging, volume change was estimated within participants using longitudinal MRI (Figure 4.2B). Consistent with the cross-sectional analysis, more privileged individuals exhibited accelerated loss of aWBV (time-by-SES β = 0.11 cm3 per year per Hollingshead, p < 0.05), controlling for sex-bytime and main effects of SES and time within the multilevel model (χ2 = 191.96, d.f. = 3, p < 0.001; adding baseline age did not contribute). For example, spanning the longitudinal sample range from middle privilege (Hollingshead = 40) to highest privilege (Hollingshead = 11), model estimates of aWBV loss nearly doubled from 4.3 cm3/yr to 7.4 cm3/yr (0.39%/yr to 0.68%/yr, relative to model intercept). Similar parameter estimates for the time-by-SES interaction (β = 0.11 cm3 per year per Hollingshead unit, p < 0.05) were obtained with a model of aWBV controlling for baseline aWBV (χ2 = 20.48, d.f. = 2, p < 0.001).

70

Figure 4.2 (next page). Cross-sectional and longitudinal plots of brain volume as a function of socioeconomic status. (A) Cross-sectional adjusted whole brain volume (aWBV, shown adjusted for effects of head size, age, and sex) is reduced in more privileged individuals (1.3 cm3 per Hollingshead unit). Each data point represents a nondemented older adult from the main sample of 100. (B) Longitudinal aWBV from 33 of the above 100 who participated in follow-up MRI; here each data point represents an MRI, with best-fit lines connecting each participants’s data. Lines are positioned according to participants’ Hollingshead, and time is nested with 5 years scaled as shown (Hollingshead does not vary per individual). Baseline reduction in aWBV with privilege is readily apparent in the longitudinal subgroup. The downward tilting in slope with privilege (time-by-SES interaction) is more subtle, though visible, and of primary interest. Modeling predicts an increase in volume loss from 4.3 cm3/yr to 7.4 cm3/yr across the sample range, after accounting for effects of baseline age, SES, and sex (aWBV is shown adjusted for head size). SES = socioeconomic status, with privilege shown increasing from left to right, based on the Hollingshead twofactor index of social position

71

FIGURE 4.2

Adjusted Whole Brain Volume (cm3)

A 1150

1100

1050

1000

950

Middle

900 0 -47

-43

-39

-35

High-Middle -31

-27

-23

High -19

-15

-11

-7

SES (Hollingshead)

Adjusted Whole Brain Volume (cm3)

B

1150

5 yrs

1100

1050

1000

950

Middle

High-Middle

High

900 0 -47

-43

-39

-35

-31

-27

-23

SES (Hollingshead)

72

-19

-15

-11

-7

TABLE 4.3. Summary Data Adjusted Whole Brain Volume ± SD (n), cm3 SES

Change ± SD (n), cm3/y

All (100)

Stable CDR 0 (64)

Preclinical CDR 0 (19)

Longitudinal (33)

Middle

1058 ± 59 (31)

1057 ± 54 (20)

1095 ± 20 (5)

-4.8 ± 3.2 (7)

High-Middle

1038 ± 42 (40)

1041 ± 44 (26)

1022 ± 35 (8)

-6.6 ± 3.3 (15)

High

1029 ± 40 (29)

1025 ± 42 (18)

1017 ± 37 (6)

-6.8 ± 2.5 (11)

Notes: Mean values ± standard deviation for cross-sectional volume and longitudinal change estimates. Adjusted whole brain volume (aWBV) is shown for the main nondemented sample of 100, as well by group for the sub-sample with clinical follow-up. For summary, WBV was adjusted only for eTIV on the n = 100 sample (see text for estimates after correcting for additional covariates). Summary change estimates were based on the slopes of ordinary-least-square regressions through each participant’s sample of aWBV measurements over time. SES = socioeconomic status.

Evidence that reserve may be an important factor in AD. Figures 4.3 and 4.4 display results that explore aWBV in relation to amyloid imaging with PIB and follow-up clinical assessments. The results, in composite, suggest brain volume differences linked to SES are associated with preclinical dementia. First, 9 of 58 individuals (16%) within the separate CDR 0 PIB sample (ages 47 to 86) summarized in Table 4.2 were positive for PIB binding, suggesting many may harbor preclinical AD. Second, Figure 4.3 shows that there was a main effect (p < 0.05) of positive PIB binding on brain volume: aWBV was estimated to decline 27 cm3 (2.5%, from 1066 to 1039 cm3) in the CDR 0 PIB+ group, after adjusting for effects of age (model F[3,54] = 151.62, p < 0.001, R2 = 0.89). This result indicates that a marker of AD pathology (PIB binding to amyloid) is associated with brain volume differences in a nondemented sample.

73

Third, Figure 4.4 offers tentative support for a contribution of preclinical dementia to the effect of SES depicted in Figure 4.2A. Participants were grouped as preclinical if subsequent clinical evaluation using the CDR indicated the onset of very mild dementia (CDR 0.5). Adding group status to the cross-sectional model (F[7,92] = 164.10.02, p < 0.001, R2 = 0.93) revealed a trend for a group-by-privilege interaction (β = 1.9 cm3 per Hollinshead per clinical conversion, p = 0.08; the interaction term became significant [β = 2.4, p < 0.05] when the model was run on the full nondemented sample, which assumes those CDR 0 individuals with no clinical follow-up remain nondemented.) The magnitude of the interaction predicts that the cross-sectional decline in aWBV with privilege (β = 1.3 cm3 less aWBV per Hollingshead unit overall) will increase by 1.9 cm3 per Hollingshead unit in individuals with preclinical dementia.

74

FIGURE 4.3 1080

Adjusted Whole Brain Volume (cm 3)

1070

n = 49 1060

* 1050

n=9

1040

1030

10200

CDR 0 PIB-

CDR 0 PIB+

Figure 4.3. Adjusted whole brain volume (aWBV, shown adjusted for effects of head size and age) is 27 cm3 (2.5%) lower in nondemented participants with high uptake of PIB, which binds to amyloid-β in plaques. PIB = [11C]Pittsburgh Compound-B (PIB); CDR = Clinical Dementia Rating, with 0 indicating no dementia.

75

FIGURE 4.4 5 yrs

1150

Adjusted Whole Brain Volume (cm3)

CDR 0.5 1100

1050

1000

950

CDR 0 (n = 19) y{ Preclinical Stable CDR 0 (n = 64) Middle

900 0 -47

-43

-39

-35

High-Middle -31

-27

-23

High -19

-15

-11

-7

SES (Hollingshead)

Figure 4.4. The relationship between adjusted whole brain volume (aWBV, as shown in Figure 2A) and SES is stronger in nondemented participants who subsequently develop dementia. Each data point represents the same nondemented older adult as shown in Figure 2A for the 83 participants who received clinical follow-up subsequent to MRI. Lines extending from each point represent the duration of clinical follow-up, with 5 years nested and scaled as shown. A vertical tick marks when certain participants received a CDR of 0.5, indicating very mild dementia. These participants are classified as preclinically demented with respect to MRI, acquired when all were nondemented (CDR 0). The magnitude of the decline in aWBV with privilege is tentatively greater in the preclinical group. As shown, the best-fit regressions decline by 1.4 cm3 per Hollingshead in the stable CDR group and 3.4 cm3 per Hollingshead in the preclinical group. CDR = Clinical Dementia Rating; SES = socioeconomic status, with privilege shown increasing from left to right, based on the Hollingshead two-factor index of social position.

76

DISCUSSION

Nondemented participants with high SES (the most privileged individuals) were found to have reduced whole brain volume (cross-sectional analysis) and accelerated volume loss (longitudinal analysis). The capacity for more privileged individuals to cope longer with brain pathology before manifesting signs of dementia, consistent with the reserve hypothesis, may explain this counterintuitive association as elaborated below.

SES associated with brain volume reduction in nondemented aging. The main result in the present paper is the strong evidence that high SES is associated with lower adjusted whole brain volume in nondemented older adults (see Figure 4.2). It is worth emphasizing that, by design, this study concerns individual differences in long-term structural change (illustrated in Figure 4.1), not early-established differences such as in head size (reviewed in Rushton and Ankney 1996, Wickett et al., 2000). This focus on change is clear in the longitudinal result, which shows accelerated volume loss in more privileged individuals, but even the cross-sectional analysis indirectly uses individuals as their own control, primarily in the adjustment for head size. Thus, baseline differences in head size across SES levels do not account for the present results. Moreover, in the present sample, we did not find significant head size differences attributable to SES. Our main cross-sectional finding (Figure 4.2A) replicates and strengthens the most comparable study from Coffey and colleagues (Coffey et al., 1999). Spanning a representative range of their sample from 8 to 17 years of education, peripheral CSF was

77

reported to increase by 15.9 cm3 (1.7% of whole brain mean; for a similar result with analysis restricted to patients with probable AD, see Kidron et al., 1997). Peripheral CSF was defined from traced structures as cranial minus brain minus ventricular volume, and its increase was interpreted as indirect evidence of greater age-related brain volume decline in the more educated. The negative finding for the direct measure of whole brain volume was attributed to measurement error. Here, we restricted our cross-sectional analysis to a direct measure of whole brain volume using estimates from multiple high contrast MRI images and the effect was found to be robust. The longitudinal finding illustrated in Figure 4.2B confirms the direction of the association between volume and privilege and provides novel evidence that this association is related to aging and present in older age. The overall rate of decline with age (6.36 cm3/yr or 0.58%/yr) is within the range (0.37%/yr to 0.88%/yr) reported by comparable studies with longitudinal MRI of the whole brain in nondemented older adults (reviewed in Fotenos et al., 2005). Considering the accelerated atrophy rate associated with SES in this study (ranging from of 0.39%/yr to 0.68%/yr between middle and high privilege), it is possible that some of the variance between prior studies arose due to sample differences in SES.

The role of preclinical AD and cognitive reserve. What explains these results? Based on similar cross-sectional findings in advanced aging (Coffey et al., 1999) and AD (Kidron et al., 1997), others have offered theoretical accounts invoking the cognitive reserve hypothesis (see also Kramer et al., 2004, Scarmeas and Stern 2004). In addition to clear

78

documentation of the influence of SES on brain volume, we present here three novel observations to support the possibility that cognitive reserve contributes to these observations. First, 16% of our nondemented PIB sample showed high levels of binding indicative of amyloid plaque presence, suggesting a number of individuals harbor preclinical AD. Second, high PIB binding was associated with reduced aWBV (Figure 4.3). Third, in the larger sample with follow-up clinical data, a trend for a group-byprivilege interaction was observed with reduced aWBV associated with more privileged individuals who soon showed signs of very mild dementia (Figure 4.4). This interaction was significant when considering all nondemented individuals, including those without follow-up clinical assessment. Together, these results suggest that AD-dependent atrophy is detectable prior to the earliest currently recognizable clinical expression of dementia (Gosche et al., 2002, den Heijer et al., 2006, Jagust et al., 2006). The interaction clarifies the association with SES and suggests that the most privileged individuals are able to harbor neurodegenerative disease further into its course without clinical detection. While indirect, evidence for accelerated atrophy linked to disease also suggests that preclinical pathology may have advanced to the point of influencing cellular integrity or cell loss (Terry et al., 1991, Price et al., 2001), and this underscores the potential of volumetric biomarkers in the assessment and tracking of early AD (Mortimer et al., 2005). A reserve explanation for our results is consistent with recent evidence that more educated individuals declined more rapidly on neuropsychological tests several years prior to AD diagnosis (Amieva et al., 2005, Scarmeas et al., 2006). Other studies have controlled for AD pathology and shown that the same plaque burden lead to less global

79

cognitive decline in more educated groups (Bennett et al., 2003, Bennett et al., 2005). Comparable education-related differences in dementia expression have been demonstrated in a large, multi-center sample restricted to individuals meeting neuropathologic criteria for AD (Roe et al., submitted). Preclinical individuals with greater reserve would be expected to have more advanced pathology in order to explain why they show more severe structural or functional decline. In this study, assuming pathology develops independent of reserve variables (Del Ser et al., 1999, Bennett et al., 2003, Mortimer et al., 2005; though see Jankowsky et al., 2005, Lazarov et al., 2005), such explanatory differences in pathology might arise from within-individual differences in preclinical duration or between-individual differences in the frequency of clinical exclusion related to SES. The present data thus are consistent with SES influencing the ability to detect cognitive impairment in the presence of pathology and possibly cell loss. It is unclear whether there is any modification of the underlying disease process by life experiences associated with SES. Education and occupational attainment may protect against AD through a “use it and hide it” mechanism in comparison to the more traditionally assumed “use it or lose it” explanation (Swaab et al., 2002).

Limitations and caveats. Limitations of this study highlight open questions and may help to guide future research. The present study explored SES between the middle and high range. Thus, our results cannot speak to how low SES affects brain aging and AD, and future structural research in the underprivileged is needed (Del Ser et al., 1999, Mortimer et al., 2003).

80

Not all reports exploring reserve and brain volume measures find an association in nondemented older adults (Coffey et al., 1992, Passe et al., 1997, Raz et al., 2005). The convergence of cross-sectional and longitudinal evidence here strengthens our confidence that the association holds for this sample. Sample and measurement differences may account for null findings, and we are aware of no contradictory studies that suggest a moderating role (opposite to the amplifying role reported here) of educational or occupational attainment on brain aging (for review of modifiers, see Kramer et al., 2004, Van Petten 2004). Replication in additional large-sample studies will be important for generalizing these results. A final point to raise is that, while the present results support a cognitive reserve model, further data will be required to test the model and to explore whether it can fully account for all of the data or whether additional factors are at work. Specifically, it remains difficult to account fully for the magnitude of the SES-related volume difference unless the reserve differential is longer and/or CDR 0 pathology more burdensome than recent research suggests (Amieva et al., 2005, Bennett et al., 2005, Galvin et al., 2005). We thus conclude that cognitive reserve likely explains some, but perhaps not all, of the relationship between SES and structural brain aging.

81

CHAPTER 5 GENERAL DISCUSSION

Centenarians are the most rapidly growing age group in the United States (Wan et al., 2005). One centenarian who serves as an example of successful aging is Sister Marcella. When Sister Marcella died, age 101, at the Counsel Hill convent of the School Sisters of Notre Dame, she had been followed as part of the Nun Study for four years (Snowdon 2003). Up until her death, she rated her ability to care for herself as “excellent.” She was known for her remarkable memory and storytelling, and cognitive testing confirmed a high and maintained level of function. At autopsy, her brain was largely free of pathology: it weighed 1280 grams, was rated Braak stage 0 (Braak and Braak 1991), and showed no evidence of stroke. As Sister Marcella’s example makes clear, advanced age is not synonymous with AD and dementia. This thesis has compared structural brain change between demented and nondemented adults. Underlying our experimental strategy has been the hypothesis that multiple factors may contribute independently to demented and nondemented brain aging (Berg 1985, Morris 1999, Della-Maggiore et al., 2002, Buckner 2004, Hedden and Gabrieli 2005). This multiple factor framework compares to a unitary framework that places AD and dementia at the tail-end of a single aging continuum (Brayne and Calloway 1988, Whalley 2002, Bartzokis 2004, de la Torre 2004a). Taken together, the results of this thesis strengthen the multiple factor framework. We thus review our results

82

in terms of AD and non-AD factors that may separately contribute to age-related structural change (see Figure 5.1).

Figure 5.1 (next page). Thesis results within a multiple factor framework of brain aging. This figure summarizes the main results in the left column, with the middle and right columns showing which parts on the left are hypothesized to reflect AD (middle column) and non-AD (right column) factors. Possible pathophysiological processes underlying such factors are discussed within the text. (A) From Figure 2.3; note the possibility that that separate non-AD processes (right column) may contribute to early-onset (black) and older-onset (blue) components of the adult-span curve; (B) from Figure 3.2; and (C) from Figure 4.3. See individual figure legends for details.

83

FIGURE 5.1 Result

AD Factors

Non-AD Factors

Age

Age

Brain Volume

Brain Volume

Brain Volume

Brain Volume

A

Age

Age B DAT

CDR 0

Atrophy

1.0

CDR 0

Z=-7

Z=38 Z=-7

Z=-22

Z=-22

Z=-7 Z=-22 0.3

DAT Z=38

CDR 0

Z=38

DAT

0.3

Atrophy

1.0

0.3

Atrophy

SES

Brain Volume

Brain Volume

Brain Volume

C

SES

84

SES

1.0

Before elaborating on this decomposition, what might AD and non-AD factors represent? At least two increasingly well characterized pathophysiological cascades emerge as leading candidates from the literature: AD as a process of amyloid-beta (Aβ) misfolding and toxicity (Selkoe 2004, Walsh and Selkoe 2004) and sporadic small vessel disease (SVD) as a process of microvasculature fibrosis and ischemia (Pugh and Lipsitz 2002, Ringelstein and Nabavi 2005, Farkas et al., 2006). We begin with a brief overview of this pathophysiology. In the case of AD, alterations in four genes have been identified that cause early-onset or accelerated late-onset familial disease; strikingly, all have been linked to the metabolism of Aβ (Tanzi and Bertram 2005). Extracellular soluble Aβ oligomers (Dodart et al., 2002) and diffuse and fibrillar plaques (Urbanc et al., 2002) have been found to interfere with synaptic function and associate with the activation of microglia and astrocytes, though gaps remain in the mechanism linking Aβ to tauopathy. In DAT, loss of synapses strongly correlates with cognitive decline (Terry et al., 1991, Coleman and Yao 2003), prominently involving impairment in declarative memory (Spaan et al., 2003, Galvin et al., 2005). Tauopathy and atrophy progress sequentially (Braak et al., 1999, Delacourte et al., 1999), with early involvement of medial temporal regions and associated neocortex (see Figure 3.2B; also Buckner et al., 2005); Aβ aggregation is spatiotemporally heterogeneous, but correlated with disease staging (Delacourte et al., 2002, Klunk et al., 2004). Turning to sporadic SVD, an accelerated form (analogous to early-onset familial AD) may be represented by cerebtal autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL; reviewed in Kalimo et al., 2002,

85

Ringelstein and Nabavi 2005). The implicated mutation weakens the contact between vascular smooth muscle cells (Shawber and Kitajewski 2004), a possible model for how risk factors, including hypertension, hyperinsulinism, hyperhomocysteinemia, and hyperlipidemia (Longstreth et al., 1996, Laloux et al., 2004, Ringelstein and Nabavi 2005), may function in sporadic (wild-type) SVD. MRI findings in SVD include lacunar infarcts, T2-weighted hyperintensities, perivascular (Virchow-Robin) spaces, and atrophy, most prominently in frontal white matter (Weis et al., 1991, Double et al., 1996, Head et al., 2004). Though sporadic SVD may rarely progress to frank vascular dementia in isolation (Nolan et al., 1998), it is associated with the exaggeration of age-related declines on measures of executive function, cognitive speed, affect, gait, and bladder control (Sakakibara et al., 1999, de Groot et al., 2000, Schmidtke and Hull 2002, Wolfson et al., 2005, Charlton et al., 2006). In summary, sporadic AD and SVD are both widely prevalent, adverse, age-related processes, but SVD has a distinct pathological signature, spatiotemporal distribution, genetic risk profile, and neuropsychological sequela. Thus SVD represents a plausible non-AD factor in the cause of common and consequential age-related brain change, providing context for the following decomposition of these thesis results into AD and non-AD factors of structural brain change. For whole brain volume, we found that nondemented aging is accompanied by steady volume decline even in the youngest adults, with atrophy rate more than doubling in the earliest stages of DAT. By itself, the acceleration of structural loss in DAT does not distinguish between unitary or multiple factor interpretations; however, closer inspection raises at least three problems for the unitary account. First, as Figure 2.3A

86

makes clear, whole brain volume in the oldest old nondemented individuals falls below the range for the younger old with dementia. This deviation from a directly proportional relationship between volume decline and DAT suggests that certain factors that do not cause DAT nevertheless contribute to volume decline. Second, it is difficult to account for age-related volume decline in young adults (<30) in terms of known pathological cascades, particularly as studies of childhood development show that a downward linear trajectory may start in adolescence (Buckner et al., 2005, Courchesne et al., 2000, Liu et al., 2003). Third, the decomposition of whole brain volume into gray and white compartments in Figure 2.2 shows a mostly linear course for gray decline and nonlinear course for white matter, again consistent with potentially independent factors. In addition to these whole brain findings, regional analysis provides evidence of an anatomical dissociation. Although the statistical maps in Figure 3.2B showed minimal difference between DAT and CDR 0 groups for three-month change, there were significant differences in the reference study with longer (mean two-year) follow-up (from Buckner et al., 2005). Atrophy in DAT was most prominently accelerated in medial temporal regions and a distributed network of parietal cortex, including the precuneus, posterior cingulate, retrosplenium, and lateral posterior parietal regions. The basis for a relationship between medial temporal atrophy and parietal atrophy in the natural history of AD is a target of active investigation (for example, Vincent et al., 2006). Relevant to the anatomical dissociation of AD and non-AD factors, prefrontal atrophy was prominent in the nondemented sample, but there was no acceleration in these prefrontal regions in DAT. This observation of differential vulnerability suggests that contributions to age-

87

related prefrontal atrophy may be independent of AD (Hubbard and Anderson 1981, Double et al., 1996, Ohnishi et al., 2001, Head et al., 2005). Future research should continue toward understanding these non-AD factors. A key unresolved question is how non-AD factors contribute to age-related change in white matter versus gray matter (Raz 2004). Neuroimaging methods for measuring white matter aging focus on change in relaxometry (white matter hyperintensities and infarcts), water diffusion along and dependent on the integrity of fiber tracts (diffusion tensor anisotropy), and structure (white matter volume loss). It remains unclear whether vascular disease linked to these white matter changes, another pathological process, or a normal physiological process contributes to age-related, AD-independent gray matter loss. Modifier studies will be key to resolving questions of cause and effect in brain aging and motivated our interest in a candidate modifier, socioeconomic status (SES), in our final study. Our SES results draw attention to uncertainties and interactions within the multiple factor framework. We found a positive association between SES and age-related volume loss in nondemented older adults. Yet we also found evidence that preclinical AD may account for this association. Preclinical AD represents a source of uncertainty in the multiple factor framework because it complicates experimental isolation of non-AD factors. Relating preclinical AD to brain structure, preliminary results from amyloid imaging (see Figure 4.3) suggest that atrophy may accelerate in association with high plaque binding prior to the onset of dementia (see also Kaye et al., 1997, Fox et al., 1999, Gosche et al., 2002, Marquis et al., 2002, Archer et al., 2006, den Heijer et al., 2006,

88

Jagust et al., 2006). A methodological implication is that observational studies may be insufficient to isolate factor-pure experimental variables for studies of physiological aging. For example, to determine whether SES associates with accelerated brain aging, not just in the absence of dementia, but in the absence of potential pathological factors, interventional designs would offer the best control. Neuroimaging studies along these lines might assign participants to high- and low-intensity education courses, or (a quasi intervention) track students during the school-year versus summer months (see also Draganski et al., 2004, May et al., 2006). Lastly, we have argued that the reserve hypothesis may explain why SES relates most strongly to brain volume in nondemented participants on the cusp of dementia (see Figure 4.4). Reserve represents the accumulated capacity to buffer and compensate for pathology; as a reserve variable, SES is a putative proxy for this lifetime process of accumulation (Stern 2006). The simplifying assumption has generally been that accumulation of reserve has no effect on AD or brain aging. However, our results leave open this possibility. Further challenging an isolated reserve model, evidence is beginning to build for a direct connection between cognitive stimulation and AD pathology (Jankowsky et al., 2003, Kamenetz et al., 2003, Cirrito et al., 2005, Lazarov et al., 2005; reviewed in Selkoe 2006). As more preclinical candidates are identified with PIB, it will be interesting to test whether amyloid pathology, and not just dementia risk, varies as a function of reserve variables, such as SES, in support of a more direct interaction (Snowdon et al., 1996, Snowdon et al., 2000). For in the final analysis, the decomposition of aging into multiple factors as discussed above is just a first step toward

89

understanding how AD, vascular disease, reserve, and other physiological and pathophysiological processes interact to create the constellation of changes observed in the brain throughout the life-span.

90

REFERENCES

Alzheimer A, Stelzmann RA, Schnitzlein HN, Murtagh FR (1995) An English translation of Alzheimer's 1907 paper, "Uber eine eigenartige Erkankung der Hirnrinde". Clin Anat 8:429-31. Amieva H, Jacqmin-Gadda H, Orgogozo JM, Le Carret N, Helmer C, Letenneur L, Barberger-Gateau P, Fabrigoule C, Dartigues JF (2005) The 9 year cognitive decline before dementia of the Alzheimer type: a prospective population-based study. Brain 128:1093-1101. Archer HA, Edison P, Brooks DJ, Barnes J, Frost C, Yeatman T, Fox NC, Rossor MN (2006) Amyloid load and cerebral atrophy in Alzheimer's disease: An C-11-PIB positron emission tomography study. Ann Neurol 60:145-147. Ashburner J, Friston KJ (2000) Voxel-based morphometry - The methods. Neuroimage 11:805-821. Ashburner J, Friston KJ (2001) Why voxel-based morphometry should be used. Neuroimage 14:1238-1243. Ball M, Braak H, Coleman P, Dickson D, Duyckaerts C, Gambetti P, Hansen L, Hyman B, Jellinger K, Markesbery W, Perl D, Powers J, Price J, Trojanowski JQ, Wisniewski H, Phelps C, Khachaturian Z (1997) Consensus recommendations for the postmortem diagnosis of Alzheimer's disease. Neurobiol Aging 18:S1-S2. Bartzokis G (2004a) Age-related myelin breakdown: a developmental model of cognitive decline and Alzheimer's disease. Neurobiol Aging 25:5-18.

91

Bartzokis G (2004b) Quadratic trajectories of brain myelin content: unifying construct for neuropsychiatric disorders. Neurobiol Aging 25:49-62. Bennett DA, Schneider JA, Wilson RS, Bienias JL, Arnold SE (2005) Education modifies the association of amyloid but not tangles with cognitive function. Neurology 65:953-955. Bennett DA, Wilson RS, Schneider JA, Evans DA, de Leon C, Arnold SE, Barnes LL, Bienias JL (2003) Education modifies the relation of AD pathology to level of cognitive function in older persons. Neurology 60:1909-1915. Bennett DA, Schneider JA, Arvanitakis Z, Kelly JF, Aggarwal NT, Shah RC, Wilson (2006) Neuropathology of older persons without cognitive impairment from two community-based studies. Neurology 66:1837-1844. Berg L, McKeel DW, Miller JP, Storandt M, Rubin EH, Morris JC, Baty J, Coats M, Norton J, Goate AM, Price JL, Gearing M, Mirra SS, Saunders AM (1998) Clinicopathologic studies in cognitively healthy aging and Alzheimer disease relation of histologic markers to dementia severity, age, sex, and apolipoprotein E genotype. Arch Neurol 55:326-335. Berg L (1985) Does Alzheimer's disease represent an exaggeration of normal aging? Arch Neurol 42:737-739. Blatter DD, Bigler ED, Gale SD, Johnson SC, Anderson CV, Burnett BM, Parker N, Kurth S, Horn SD (1995) Quantitative volumetric analysis of brain MR: normative database spanning 5 decades of life. Am J Neuroradiol 16:241-251.

92

Bookstein FL (2001) "Voxel-based morphometry" should not be used with imperfectly registered images. Neuroimage 14:1454-1462. Borenstein AR, Copenhaver CI, Mortimer JA (2006) Early-life risk factors for Alzheimer disease. Alzheimer Dis. Assoc. Dis. 20:63-72. Boxer AL, Rankin KP, Miller BL, Schuff N, Weiner M, Gorno-Tempini ML, Rosen HJ (2003) Cinguloparietal atrophy distinguishes Alzheimer disease from semantic dementia. Arch Neurol 60:949-956. Braak E, Griffing K, Arai K, Bohl J, Bratzke H, Braak H (1999) Neuropathology of Alzheimer's disease: what is new since A. Alzheimer? Eur Arch Psych Clin Neurosci 249:14-22. Braak H, Braak E (1991) Neuropathological stageing of Alzheimer-related changes. Acta Neuropathol 82:239-259. Braak H, Braak E (1997) Frequency of stages of Alzheimer-related lesions in different age categories. Neurobiol Aging 18:351-357. Bradley KM, Bydder GM, Budge MM, Hajnal JV, White SJ, Ripley BD, Smith AD (2002) Serial brain MRI at 3-6 month intervals as a surrogate marker for Alzheimer's disease. Br J Radiol 75:506-513. Brayne C, Calloway P (1988) Normal ageing, impaired cognitive function, and senile dementia of the Alzheimer's type: a continuum? Lancet 1:1265-1267. Broca P (1861) Sur le volume et la forme du cerveau suivant les individus et suivant les races. Paris: Hennuyer.

93

Brun A, Gustafson L (1976) Distribution of cerebral degeneration in Alzheimer's disease. A clinico-pathological study. Arch Psychiatr Nervenkr 223:15-33. Buckner RL, Head D, Parker J, Fotenos AF, Marcus D, Morris JC, Snyder AZ (2004) A unified approach for morphometric and functional data analysis in young, old, and demented adults using automated atlas-based head size normalization: reliability and validation against manual measurement of total intracranial volume. Neuroimage 23:724-738. Buckner RL, Snyder AZ, Shannon BJ, LaRossa G, Sachs R, Fotenos AF, Sheline YI, Klunk WE, Mathis CA, Morris JC, Mintun MA (2005) Molecular, structural, and functional characterization of Alzheimer's disease: evidence for a relationship between default activity, amyloid, and memory. J Neurosci 25:7709-7717. Buckner RL (2004) Memory and executive function in aging and AD: multiple factors that cause decline and reserve factors that compensate. Neuron 44:195-208. Burns JM, Church JA, Johnson DK, Xiong CJ, Marcus D, Fotenos AF, Snyder AZ, Morris JC, Buckner RL (2005) White matter lesions are prevalent but differentially related with cognition in aging and early Alzheimer disease. Arch Neurol 62:18701876. Callen D, Black SE, Gao F, Caldwell CB, Szalai JP (2001) Beyond the hippocampus MRI volumetry confirms widespread limbic atrophy in AD. Neurology 57:16691674.

94

Cardenas VA, Du AT, Hardin D, Ezekiel F, Weber P, Jagust WJ, Chui HC, Schuff N, Weiner MW (2003) Comparison of methods for measuring longitudinal brain change in cognitive impairment and dementia. Neurobiol Aging 24:537-544. Carr DB, Gray S, Baty J, Morris JC (2000) The value of informant versus individual's complaints of memory impairment in early dementia. Neurology 55:1724-1726. Chan D, Fox NC, Jenkins R, Scahill RI, Crum WR, Rossor MN (2001) Rates of global and regional cerebral atrophy in AD and frontotemporal dementia. Neurology 57:1756-1763. Chan D, Janssen JC, Whitwell JL, Watt HC, Jenkins R, Frost C, Rossor MN, Fox NC (2003) Change in rates of cerebral atrophy over time in early-onset Alzheimer's disease: longitudinal MRI study. Lancet 362:1121-1122. Charlton RA, Morris RG, Nitkunan A, Markus HS (2006) The cognitive profiles of CADASIL and sporadic small vessel disease. Neurology 66:1523-1526. Chetelat G, Landeau B, Eustache F, Mezenge F, Viader F, de la Sayette V, Desgranges B, Baron JC (2005) Using voxel-based morphometry to map the structural changes associated with rapid conversion in MCI: A longitudinal MRI study. Neuroimage 27:934-946. Christensen GE, Rabbitt RD, Miller MI (1996) Deformable templates using large deformation kinematics. IEEE Trans Image Process 5:1435-1447. Christiansen P, Larsson HB, Thomsen C, Wieslander SB, Henriksen O (1994) Age dependent white matter lesions and brain volume changes in healthy volunteers. Acta Radiol 35:117-122.

95

Cirrito JR, Yamada KA, Finn MB, Sloviter RS, Bales KR, May PC, Schoepp DD, Paul SM, Mennerick S, Holtzman DM (2005) Synaptic activity regulates interstitial fluid amyloid-beta levels in vivo. Neuron 48:913-922. Coffey CE, Saxton JA, Ratcliff G, Bryan RN, Lucke JF (1999) Relation of education to brain size in normal aging - implications for the reserve hypothesis. Neurology 53:189-196. Coffey CE, Wilkinson WE, Parashos IA, Soady SAR, Sullivan RJ, Patterson LJ, Figiel GS, Webb MC, Spritzer CE, Djang WT (1992) Quantitative cerebral anatomy of the aging human brain - a cross-sectional study using magnetic resonance imaging. Neurology 42:527-536. Coggan JS, Grutzendler J, Bishop DL, Cook MR, Gan WB, Heym J, Lichtman JW (2004) Age-associated synapse elimination in mouse parasympathetic ganglia. J Neurobiol 60:214-226. Colcombe SJ, Erickson KI, Raz N, Webb AG, Cohen NJ, McAuley E, Kramer AF (2003) Aerobic fitness reduces brain tissue loss in aging humans. J Gerontol Ser A-Biol Sci Med Sci 58:176-180. Coleman PD, Yao PJ (2003) Synaptic slaughter in Alzheimer's disease. Neurobiol Aging 24:1023-1027. Courchesne E, Chisum HJ, Townsend J, Cowles A, Covington J, Egaas B, Harwood M, Hinds S, Press GA (2000) Normal brain development and aging: quantitative analysis at in vivo MR imaging in healthy volunteers. Radiology 216:672-682.

96

Csernansky JG, Hamstra J, Wang L, McKeel D, Price JL, Gado M, Morris JC (2004) Correlations between antemortem hippocampal volume and postmortem neuropathology in AD subjects. Alzheimer Dis Assoc Dis 18:190-195. Damadian R (1971) Tumor detection by nuclear magnetic resonance. Science 171:11511153. Davatzikos C (2004) Why voxel-based morphometric analysis should be used with great caution when characterizing group differences. Neuroimage 23:17-20. Davis P, Wright EA (1977) A new method for measuring cranial cavity volume and it application to the assessment of cerebral atrophy at autopsy. Neuropathol Appl Neurobiol 3:341-358. de Groot JC, de Leeuw FE, Oudkerk M, Hofman A, Jolles J, Breteler M (2000) Cerebral white matter lesions and depressive symptoms in elderly adults. Arch Gen Psychiatry 57:1071-1076. de la Torre JC (2004) Alzheimer's disease is a vasocognopathy: a new term to describe its nature. Neurol Res 26:517-524. DeCarli C, Massaro J, Danielle H, Hald J, Tullberg M, R. A, Beiser A, D'Agostino R, Wold P (2004) Measures of brain morphology and infarction in the framingham heart study: establishing what is normal. Neurobiol Aging. Dekaban AS (1978) Changes in brain weights during the span of human life: relation of brain weights to body heights and body weights. Ann Neurol 4:345-56.. DeKosky ST, Marek K (2003) Looking backward to move forward: early detection of neurodegenerative disorders. Science 302:830-834.

97

Del Ser T, Hachinski V, Merskey H, Munoz DG (1999) An autopsy-verified study of the effect of education on degenerative dementia. Brain 122:2309-2319. Delacourte A, David JP, Sergeant N, Buee L, Wattez A, Vermersch P, Ghozali F, FalletBianco C, Pasquier F, Lebert F, Petit H, Di Menza C (1999) The biochemical pathway of neurofibrillary degeneration in aging and Alzheimer's disease. Neurology 52:1158-1165. Delacourte A, Sergeant N, Champain D, Wattez A, Maurage CA, Lebert F, Pasquier F, David JP (2002) Nonoverlapping but synergetic tau and APP pathologies in sporadic Alzheimer's disease. Neurology 59:398-407. Della-Maggiore V, Grady CL, McIntosh AR (2002) Dissecting the effect of aging on the neural substrates of memory: Deterioration, preservation or functional reorganization? Rev Neurosci 13:167-181. DeMyer MK, Gilmor RL, Hendrie HC, DeMyer WE, Augustyn GT, Jackson RK (1988) Magnetic resonance brain images in schizophrenic and normal subjects: influence of diagnosis and education. Schizophr Bull 14:21-37. den Heijer T, Geerlings MI, Hoebeek FE, Hofman A, Koudstaal PJ, Breteler M (2006) Use of hippocamal and amygdalar volumes on magnetic resonance imaging to predict dementia in cognitively intact elderly people. Arch Gen Psychiatry 63:5762. Dodart JC, Bales KR, Gannon KS, Greene SJ, DeMattos RB, Mathis C, DeLong CA, Wu S, Wu X, Holtzman DM, Paul SM (2002) Immunization reverses memory deficits

98

without reducing brain A beta burden in Alzheimer's disease model. Nat Neurosci 5:452-457. Double KL, Halliday GM, Kril JJ, Harasty JA, Cullen K, Brooks WS, Creasey H, Broe GA (1996) Topography of brain atrophy during normal aging and Alzheimer's disease. Neurobiol. Aging 17:513-521. Draganski B, Gaser C, Busch V, Schuierer G, Bogdahn U, May A (2004) Neuroplasticity: Changes in grey matter induced by training - Newly honed juggling skills show up as a transient feature on a brain-imaging scan. Nature 427:311-312. Duan HL, Wearne SL, Rocher AB, Macedo A, Morrison JH, Hof PR (2003) Age-related dendritic and spine changes in corticocortically projecting neurons in macaque monkeys. Cereb Cortex 13:950-961. Esiri MM, Matthews F, Brayne C, Ince PG, Matthews FE, Xuereb JH, Broome JC, McKenzie J, Rossi M, McKeith IG, Lowe J, Morris JH, CA Neuropathology Grp Med Res Council (2001) Pathological correlates of late-onset dementia in a multicentre, community-based population in England and Wales. Lancet 357:169175. Fagan AM, Mintun MA, Mach RH, Lee SY, Dence CS, Shah AR, LaRossa GN, Spinner ML, Klunk WE, Mathis CA, DeKosky ST, Morris JC, Holtzman DM (2006) Inverse relation between in vivo amyloid imaging load and cerebrospinal fluid a beta(42) in humans. Ann Neurol 59:512-519.

99

Farkas E, de Vos R, Donka G, Steur E, Mihaly A, Luiten P (2006) Age-related microvascular degeneration in the human cerebral periventricular white matter. Acta Neuropathol 111:150-157. Feinberg I, Thode HC Jr, Chugani HT, March JD (1990) Gamma distribution model describes maturational curves for delta wave amplitude, cortical metabolic rate and synaptic density. J Theor Biol 142:149-161. Fischl B, Salat DH, Busa E, Albert M, Dieterich M, Haselgrove C, van der Kouwe A, Killiany R, Kennedy D, Klaveness S, Montillo A, Makris N, Rosen B, Dale AM (2002) Whole brain segmentation: Automated labeling of neuroanatomical structures in the human brain. Neuron 33:341-355. Fischl B, Salat DH, van der Kouwe A, Makris N, Segonne F, Quinn B, Dale AM (2004) Sequence-independent segmentation of magnetic resonance images. Neuroimage 23:S69-S84. Folstein MF, Folstein SE, McHugh PR (1975) "Mini-mental state". A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 12:189198. Fotenos AF, Snyder AZ, Girton LE, Morris JC, Buckner RL (2005) Normative estimates of cross-sectional and longitudinal brain volume decline in aging and AD. Neurology 64:1032-1039. Fox NC, Freeborough PA (1997) Brain atrophy progression measured from registered serial MRI: Validation and application to Alzheimer's disease. J Magn Reson Imaging 7:1069-1075.

100

Fox NC, Schott JM (2004) Imaging cerebral atrophy: Normal ageing to Alzheimer's disease. Lancet 363:392-394. Fox NC, Warrington EK, Rossor MN (1999) Serial magnetic resonance imaging of cerebral atrophy in preclinical Alzheimer's disease. Lancet 353:2125-2125. Fratiglioni L, De Ronchi D, Aguero-Torres H (1999) Worldwide prevalence and incidence of dementia. Drugs Aging 15:365-375. Fratiglioni L, Paillard-Borg S, Winblad B (2004) An active and socially integrated lifestyle in late life might protect against dementia. Lancet Neurol. 3:343-353. Freeborough PA, Woods RP, Fox NC (1996) Accurate registration of serial 3D MR brain images and its application to visualizing change in neurodegenerative disorders. Journal of Computer Assisted Tomography 20:1012-1022. Galton F (1888) Head growth in students at the University of Cambridge. Nature 38:1415. Galvin JE, Powlishta KK, Wilkins K, McKeel DW Jr, Xiong C, Grant E, Storandt M, Morris JC (2005) Predictors of preclinical Alzheimer disease and dementia: a clinicopathologic study. Arch Neurol 62:758-765. Ge YL, Grossman RI, Babb JS, Rabin ML, Mannon LJ, Kolson DL (2002) Age-related total gray matter and white matter changes in normal adult brain. Part I: Volumetric MR imaging analysis. Am J Neuroradiol 23:1327-1333. Giedd JN, Blumenthal J, Jeffries NO, Castellanos FX, Liu H, Zijdenbos A, Paus T, Evans AC, Rapoport JL (1999) Brain development during childhood and adolescence: a longitudinal MRI study. Nat Neurosci 2:861-863.

101

Giedd JN (2004) Structural magnetic resonance imaging of the adolescent brain. Ann NY Acad Sci 1021:77-85. Glenner GG, Wong CW (1984) Alzheimer's disease: initial report of the purification and characterization of a novel cerebrovascular amyloid protein. Biochem Biophys Res Commun 120:885-890. Glodzik-Sobanska L, Rusinek H, Mosconi L, Li Y, Zhan J, de Santi S, Convit A, Rich K, Brys M, de Leon MJ (2005) The role of quantitative structural imaging in the early diagnosis of Alzheimer's disease. Neuroimaging Clin N Am 15:803-826. Good CD, Johnsrude IS, Ashburner J, Henson R, Friston KJ, Frackowiak R (2001) A voxel-based morphometric study of ageing in 465 normal adult human brains. Neuroimage 14:21-36. Gosche KM, Mortimer JA, Smith CD, Markesbery WR, Snowdon DA (2002) Hippocampal volume as an index of Alzheimer neuropathology - findings from the Nun study. Neurology 58:1476-1482. Goulding MR, Rogers ME, Smith SM (2003) Public health and aging: Trends in aging United States and worldwide JAMA 289:1371-1373. Gunter JL, Shiung MM, Manduca A, Jack CR (2003) Methodological considerations for measuring rates of brain atrophy. J Magn Reson Imaging 18:16-24. Gur RC, Mozley PD, Resnick SM, Gottlieb GL, Kohn M, Zimmerman R, Herman G, Atlas S, Grossman R, Berretta D, et al (1991) Gender differences in age effect on brain atrophy measured by magnetic resonance imaging. Proc Natl Acad Sci USA 88:2845-2849.

102

Gurland B (1981) The borderlands of dementia: the influence of sociocultural characteristics on rates of dementia occurring in the senium. In: Clinical aspects of Alzheimer's disease and senile dementia (Miller N, Cohene G, eds), Vol 15. pp 6184. New York: Raven Press. Gusnard DA, Raichle ME (2001) Searching for a baseline: Functional imaging and the resting human brain. Nat Rev Neurosci 2:685-694. Guttmann C, Jolesz FA, Kikinis R, Killiany RJ, Moss MB, Sandor T, Albert MS (1998) White matter changes with normal aging. Neurology 50:972-978. Han X (2005) Segmentation of multi-echo multi-spectral brain images. Grant material. Hardy J, Selkoe DJ (2002) The amyloid hypothesis of Alzheimer's disease: Progress and problems on the road to therapeutics. Science 297:353-356. Harris GJ, Schlaepfer TE, Peng LW, Lee S, Federman EB, Pearlson GD (1994) Magnetic resonance imaging evaluation of the effects of ageing on grey-white ratio in the human brain. Neuropathol Appl Neurobiol 20:290-3. Haug H, Kuhl S, Mecke E, Sass NL, Wasner K (1984) The significance of morphometric procedures in the investigation of age changes in cytoarchitectonic structures of human brain. J Hirnforsch 25:353-374. Head D, Buckner RL, Shimony JS, Williams LE, Akbudak E, Conturo TE, McAvoy M, Morris JC, Snyder AZ (2004) Differential vulnerability of anterior white matter in nondemented aging with minimal acceleration in dementia of the Alzheimer type: Evidence from diffusion tensor imaging. Cereb Cortex 14:410-423.

103

Head D, Snyder AZ, Girton LE, Morris JC, Buckner RL (2005) Frontal-hippocampal double dissociation between normal aging and Alzheimer's disease. Cereb Cortex 15:732-739. Hedden T, Gabrieli J (2005) Healthy and pathological processes in adult development: new evidence from neuroimaging of the aging brain. Curr Opin Neurol 18:740-747. Hollingshead AB (1957) Two factor index of social position. Technical report 1-11. House JS, Lepkowski JM, Kinney AM, Mero RP, Kessler RC, Herzog AR (1994) The social stratification of aging and health. J Health Soc Behav 35:213-234. House J, Williams D (2000) Understanding and reducing socioeconomic and racial/ethnic disparities in health. In: Promoting health: intervention strategies from social and behavioral research (Smedly B, Syme S, eds), pp 81-104. Washington D.C.: National Academy Press. Hubbard BM, Anderson JM (1981) A quantitative study of cerebral atrophy in old age and senile dementia. J Neurol Sci 50:135-45.. Huttenlocher PR (1979) Synaptic density in human frontal cortex - developmental changes and effects of aging. Brain Res 163:195-205. Huttenlocher PR (2002) Neural plasticity : the effects of environment on the development of the cerebral cortex. Cambridge, MA: Harvard University Press. Jack CR, Dickson DW, Parisi JE, Xu YC, Cha RH, O'Brien PC, Edland SD, Smith GE, Boeve BF, Tangalos EG, Kokmen E, Petersen RC (2002) Antemortem MRI findings correlate with hippocampal neuropathology in typical aging and dementia. Neurology 58:750-757.

104

Jack CR, Shiung MM, Gunter JL, O'Brien PC, Weigand SD, Knopman DS, Boeve BF, Ivnik RJ, Smith GE, Cha RH, Tangalos EG, Petersen RC (2004) Comparison of different MRI brain atrophy rate measures with clinical disease progression in AD. Neurology 62:591-600. Jagust W, Gitcho A, Sun F, Kuczynski B, Mungas D, Haan M (2006) Brain imaging evidence of preclinical Alzheimer's disease in normal aging. Ann Neurol 59:673681. Jankowsky JL, Melnikova T, Fadale DJ, Xu GM, Slunt HH, Gonzales V, Younkin LH, Younkin SG, Borchelt DR, Savonenko AV (2005) Environmental enrichment mitigates cognitive deficits in a mouse model of Alzheimer's disease. J Neurosci 25:5217-5224. Jankowsky JL, Xu G, Fromholt D, Gonzales V, Borchelt DR (2003) Environmental enrichment exacerbates amyloid plaque formation in a transgenic mouse model of Alzheimer disease. J Neuropathol Exp Neurol 62:1220-1227. Jernigan TL, Archibald SL, Fennema-Notestine C, Gamst AC, Stout JC, Bonner J, Hesselink JR (2001) Effects of age on tissues and regions of the cerebrum and cerebellum. Neurobiol Aging 22:581-594. Jernigan TL, Fennema-Notestine C (2004) White matter mapping is needed. Neurobiol Aging 25:37-39. Jernigan TL, Press GA, Hesselink JR (1990) Methods for measuring brain morphologic features on magnetic resonance images. Validation and normal aging. Arch Neurol 47:27-32.

105

Jernigan TL, Salmon DP, Butters N, Hesselink JR (1991) Cerebral structure on MRI, Part II: specific changes in Alzheimer's and Huntington's diseases. Biol Psychiatry 29:68-81. Kalimo H, Ruchoux MM, Viitanen M, Kalaria RN (2002) CADASIL: a common form of hereditary arteriopathy causing brain infarcts and dementia. Brain Pathol. 12:371384. Kamenetz F, Tomita T, Hsieh H, Seabrook G, Borchelt D, Iwatsubo T, Sisodia S, Malinow R (2003) APP processing and synaptic function. Neuron 37:925-937. Kantarci K, Jack CR (2003) Neuroimaging in Alzheimer disease: An evidence-based review. Neuroimaging Clin N Am 13:197-209. Karas GB, Scheltens P, Rombouts S, Visser PJ, van Schijndel RA, Fox NC, Barkhof F (2004) Global and local gray matter loss in mild cognitive impairment and Alzheimer's disease. Neuroimage 23:708-716. Katzman R, Terry R, DeTeresa R, Brown T, Davies P, Fuld P, Renbing X, Peck A (1988) Clinical, pathological, and neurochemical changes in dementia: a subgroup with preserved mental status and numerous neocortical plaques. Ann Neurol 23:138-144. Katzman R (1993) Education and the prevalence of dementia and Alzheimer's disease. Neurology 43:13-20. Kaye JA, Swihart T, Howieson D, Dame A, Moore MM, Karnos T, Camicioli R, Ball M, Oken B, Sexton G (1997) Volume loss of the hippocampus and temporal lobe in healthy elderly persons destined to develop dementia. Neurology 48:1297-1304.

106

Kidron D, Black SE, Stanchev P, Buck B, Szalai JP, Parker J, Szekely C, Bronskill MJ (1997) Quantitative MR volumetry in Alzheimer's disease: topographic markers and the effects of sex and education. Neurology 49:1504-1512. Klunk WE, Engler H, Nordberg A, Wang YM, Blomqvist G, Holt DP, Bergstrom M, Savitcheva I, Huang GF, Estrada S, Ausen B, Debnath ML, Barletta J, Price JC, Sandell J, Lopresti BJ, Wall A, Koivisto P, An (2004) Imaging brain amyloid in Alzheimer's disease with Pittsburgh Compound-B. Ann Neurol 55:306-319. Kramer AF, Bherer L, Colcombe SJ, Dong W, Greenough WT (2004) Environmental influences on cognitive and brain plasticity during aging. J Gerontol Ser A-Biol Sci Med Sci 59:940-957. Kruggel F (2006) MRI-based volumetry of head compartments: Normative values of healthy adults. Neuroimage 30:1-11. Laloux P, Galanti L, Jamart J (2004) Lipids in ischemic stroke subtypes. Acta Neurol. Belg 104:13-19. Launer LJ, Andersen K, Dewey ME, Letenneur L, Ott A, Amaducci LA, Brayne C, Copeland J, Dartigues JF, Kragh-Sorensen P, Lobo A, Martinez-Lage JM, Stijnen T, Hofman A, CA EURODEM Incidence Res Grp Work Grp (1999) Rates and risk factors for dementia and Alzheimer's disease - Results from EURODEM pooled analyses. Neurology 52:78-84. Lauterbur P (1973) Image formation by induced local interactions: examples employing nuclear magnetic resonance. Nature 242:190-191.

107

Lazarov O, Robinson J, Tang YP, Hairston IS, Korade-Mirnics Z, Lee V, Hersh LB, Sapolsky RM, Mirnics K, Sisodia SS (2005) Environmental enrichment reduces Abeta levels and amyloid deposition in transgenic mice. Cell 120:701-713. Lee HG, Casadesus G, Zhu XW, Takeda A, Perry G, Smith MA (2004) Challenging the amyloid cascade hypothesis - Senile plaques and amyloid-beta as protective adaptations to Alzheimer disease. Ann NY Acad Sci 1019:1-4. Lippa CF, Morris JC (2006) Alzheimer neuropathology in nondemented aging: keeping mind over matter. Neurology 66:1801-1802. Liu R, Lemieux L, Bell GS, Sisodiya SM, Shorvon SD, Sander J, Duncan JS (2003) A longitudinal study of brain morphometrics using quantitative magnetic resonance imaging and difference image analysis. Neuroimage 20:22-33. Logan J, Fowler JS, Volkow ND, Wang GJ, Ding YS, Alexoff DL (1996) Distribution volume ratios without blood sampling from graphical analysis of PET data. J. Cereb. Blood Flow Metab 16:834-840. Longstreth WT, Manolio TA, Arnold A, Burke GL, Bryan N, Jungreis CA, Enright PL, OLeary D, Fried L (1996) Clinical correlates of white matter findings on cranial magnetic resonance imaging of 3301 elderly people - The cardiovascular health study. Stroke 27:1274-1282. Maguire EA, Gadian DG, Johnsrude IS, Good CD, Ashburner J, Frackowiak R, Frith CD (2000) Navigation-related structural change in the hippocampi of taxi drivers. Proc Natl Acad Sci USA 97:4398-4403.

108

Manly JJ, Schupf N, Tang MX, Stern Y (2005) Cognitive decline and literacy among ethnically diverse elders. J Geriatr Psychiatry Neurol 18:213-217. Mansfield P, Maudsley AA (1977) Medical imaging by NMR. Br J Radiol 50:188-194. Marner L, Nyengaard JR, Tang Y, Pakkenberg B (2003) Marked loss of myelinated nerve fibers in the human brain with age. J Comp Neurol 462:144-152. Marquis S, Moore MM, Howieson DB, Sexton G, Payami H, Kaye JA, Camicioli R (2002) Independent predictors of cognitive decline in healthy elderly persons. Arch. Neurol 59:601-606. Masters CL, Simms G, Weinman NA, Multhaup G, McDonald BL, Beyreuther K (1985) Amyloid plaque core protein in Alzheimer disease and Down syndrome. Proc Natl Acad Sci USA 82:4245-4249. Mathalon DH, Sullivan EV, Rawles JM, Pfefferbaum A (1993) Correction for head size in brain-imaging measurements. Psychiatry Res 50:121-139. Mathis CA, Wang YM, Holt DP, Huang GF, Debnath ML, Klunk WE (2003) Synthesis and evaluation of C-11-labeled 6-substituted 2-arylbenzothiazoles as amyloid imaging agents. J Med Chem 46:2740-2754. Matsumae M, Kikinis R, Morocz IA, Lorenzo AV, Sandor T, Albert MS, Black PL, Jolesz FA (1996) Age-related changes in intracranial compartment volumes in normal adults assessed by magnetic resonance imaging. J Neurosurg 84:982-991. May A, Hajak G, Ganssbauer S, Steffens T, Langguth B, Kleinjung T, Eichhammer P (2006) Structural brain alterations following 5 days of intervention: dynamic aspects of neuroplasticity. Cereb Cortex, in press.

109

McKeel DW, Price JL, Miller JP, Grant EA, Xiong CJ, Berg L, Morris JC (2004) Neuropathologic criteria for diagnosing Alzheimer disease in persons with pure dementia of Alzheimer type. J Neuropathol Exp Neurol 63:1028-1037. Mechelli A, Crinion JT, Noppeney U, O'Doherty J, Ashburner J, Frackowiak RS, Price CJ (2004) Structural plasticity in the bilingual brain - Proficiency in a second language and age at acquisition affect grey-matter density. Nature 431:757-757. Mehta S, Grabowski TJ, Trivedi Y, Damasio H (2003) Evaluation of voxel-based morphometry for focal lesion detection in individuals. Neuroimage 20:1438-1454. Miller AK, Alston RL, Corsellis JA (1980) Variation with age in the volumes of grey and white matter in the cerebral hemispheres of man: measurements with an image analyser. Neuropathol Appl Neurobiol 6:119-32. Miller MI, Hosakere M, Barker AR, Priebe CE, Lee N, Ratnanather JT, Wang L, Gado M, Morris JC, Csernansky JG (2003) Labeled cortical mantle distance maps of the cingulate quantify differences between dementia of the Alzheimer type and healthy aging. Proc Natl Acad Sci USA 100:15172-15177. Miller MI (2004) Computational anatomy: shape, growth, and atrophy comparison via diffeomorphisms. Neuroimage 23 Suppl 1:S19-33. Mintun M, LaRossa G, Sheline Y, Dence C, Yoon Lee S, Mach R, Klunk W, Mathis C, DeKosky S, Morris J (2006) [11C]PIB in a nondemented population: potential antecedent marker of Alzheimer disease. Neurology 67:446-452.

110

Morris JC, Storandt M, Miller JP, McKeel DW, Price JL, Rubin EH, Berg L (2001) Mild cognitive impairment represents early-stage Alzheimer disease. Arch Neurol 58:397-405. Morris JC (1993) The Clinical Dementia Rating (CDR): current version and scoring rules. Neurology 43:2412-2414. Morris JC (1999) Is Alzheimer's disease inevitable with age? Lessons from clinicopathologic studies of healthy aging and very mild Alzheimer's disease. J Clin Invest 104:1171-1173. Morrison JH, Hof PR (1997) Life and death of neurons in the aging brain. Science 278:412-419. Mortimer JA, Borenstein AR, Gosche KM, Snowdon DA (2005) Very early detection of Alzheimer neuropathology and the role of brain reserve in modifying its clinical expression. J Geriatr Psychiatry Neurol 18:218-223. Mortimer JA, Snowdon DA, Markesbery WR (2003) Head circumference, education and risk of dementia: findings from the Nun study. J Clin Exp Neuropsychol 25:671679. Morton S (1849) Observations on the size of the brain in various races and families of man. Proc Natl Acad Sci USA 4:221-224. Murphy DG, DeCarli C, McIntosh AR, Daly E, Mentis MJ, Pietrini P, Szczepanik J, Schapiro MB, Grady CL, Horwitz B, Rapoport SI (1996) Sex differences in human brain morphometry and metabolism: an in vivo quantitative magnetic resonance

111

imaging and positron emission tomography study on the effect of aging. Arch Gen Psychiatry 53:585-594. Nolan KA, Lino MM, Seligmann AW, Blass JP (1998) Absence of vascular dementia in an autopsy series from a dementia clinic. J Am Geriatr Soc 46:597-604. Ohnishi T, Matsuda H, Tabira T, Asada T, Uno M (2001) Changes in brain morphology in Alzheimer disease and normal aging: Is Alzheimer disease an exaggerated aging process? Am J Neuroradiol 22:1680-1685. Pakkenberg B, Pelvig D, Marner L, Bundgaard MJ, Gundersen HJ, Nyengaard JR, Regeur L (2003) Aging and the human neocortex. Exp Gerontol 38:95-99. Pakkenberg H, Voigt J (1964) Brain weight of the Danes. Acta Anatomica 56:297-307. Pantel J, Schonknecht P, Essig M, Schroder J (2004) Distribution of cerebral atrophy assessed by magnetic resonance imaging reflects patterns of neuropsychological deficits in Alzheimer's dementia. Neurosci Lett 361:17-20. Passe TJ, Rajagopalan P, Tupler LA, Byrum CE, MacFall JR, Krishnan K (1997) Age and sex effects on brain morphology. Prog Neuro-Psychopharmacol Biol Psychiatry 21:1231-1237. Pennanen C, Testa C, Laakso MP, Hallikainen M, Helkala EL, Hanninen T, Kivipelto M, Kononen M, Nissinen A, Tervo S, Vanhanen M, Vanninen R, Frisoni GB, Soininen H (2005) A voxel based morphometry study on mild cognitive impairment. J Neurol Neurosurg Psychiatry 76:11-14. Peters A, Rosene DL (2003) In aging, is it gray or white? J Comp Neurol 462:139-143.

112

Pfefferbaum A, Mathalon DH, Sullivan EV, Rawles JM, Zipursky RB, Lim KO (1994) A quantitative magnetic resonance imaging study of changes in brain morphology from infancy to late adulthood. Arch Neurol 51:874-887. Price JL, Ko AI, Wade MJ, Tsou SK, McKeel DW, Morris JC (2001) Neuron number in the entorhinal cortex and CA1 in preclinical Alzheimer disease. Arch Neurol 58:1395-1402. Price JL, Morris JC (1999) Tangles and plaques in nondemented aging and "preclinical" Alzheimer's disease. Ann Neurol 45:358-368. Pugh KG, Lipsitz LA (2002) The microvascular frontal-subcortical syndrome of aging. Neurobiol. Aging 23:421-431. Raz N, Gunning FM, Head D, Dupuis JH, McQuain J, Briggs SD, Loken WJ, Thornton AE, Acker JD (1997) Selective aging of the human cerebral cortex observed in vivo: Differential vulnerability of the prefrontal gray matter. Cereb Cortex 7:268282. Raz N, Gunning-Dixon F, Head D, Rodrigue KM, Williamson A, Acker JD (2004) Aging, sexual dimorphism, and hemispheric asymmetry of the cerebral cortex: replicability of regional differences in volume. Neurobiol Aging 25:377-396. Raz N, Lindenberger U, Rodrigue KM, Kennedy KM, Head D, Williamson A, Dahle C, Gerstorf D, Acker JD (2005) Regional brain changes in aging healthy adults: general trends, individual differences and modifiers. Cereb Cortex 15:1676-1689.

113

Raz N (2004) The aging brain observed in vivo: differential changes and their modifiers. In: Cognitive neuroscience of aging: linking cognitive and cerebral aging (Cabeza R, Nyberg L et al., eds), pp 1-39. London, UK: Oxford University Press. Raz N (2000) Aging of the brain and its impact on cognitive performance: integration of structural and functional findings. In: Handbook of Aging and Cognition (Craick F, Salthouse TA, eds), Vol 2. pp 1-90. Mahwah, NJ: Erlbaum. Resnick SM, Pham DL, Kraut MA, Zonderman AB, Davatzikos C (2003) Longitudinal magnetic resonance imaging studies of older adults: a shrinking brain. J Neurosci 23:3295-3301. Ringelstein EB, Nabavi DG (2005) Cerebral small vessel diseases: cerebral microangiopathies. Curr Opin Neurol 18:179-188. Roe CM, Xiong C, Miller JP, Morris JC (2006) Education and Alzheimer’s disease without dementia. Submitted. Rothschild D (1937) Pathologic changes in senile psychoses and their psychobiologic significance. Am J Psychiatry 93:757-788. Rovaris M, Iannucci G, Cercignani M, Sormani MP, De Stefano N, Gerevini S, Comi G, Filippi M (2003) Age-related changes in conventional, magnetization transfer, and diffusion-tensor MR imaging findings: Study with whole-brain tissue histogram analysis. Radiology 227:731-738. Rushton JP, Ankney CD (1996) Brain size and cognitive ability: correlations with age, sex, social class, and race. Psychon Bull Rev 3:21-36.

114

Rusinek H, Endo Y, De Santi S, Frid D, Tsui WH, Segal S, Convit A, de Leon MJ (2004) Atrophy rate in medial temporal lobe during progression of Alzheimer disease. Neurology 63:2354-2359. Sakakibara R, Hattori T, Uchiyama T, Yamanishi T (1999) Urinary function in elderly people with and without leukoaraiosis: relation to cognitive and gait function. J Neurol Neurosurg Psychiatry 67:658-660. Salat DH, Buckner RL, Snyder AZ, Greve DN, Desikan R, Busa E, Morris JC, Dale AM, Fischl B (2004) Thinning of the cerebral cortex in aging. Cereb Cortex 14:721-730. Salthouse TA (2000) Methodological assumptions in cognitive aging research. (Craik FIM, Salthouse TA, eds), pp 467-498. Mahwah, NJ: USum Associates. Sanfilipo MP, Benedict R, Zivadinova R, Bakshi R (2004) Correction for intracranial volume in analysis of whole brain atrophy in multipls Sclerosis: the proportion vs. residual method. Neuroimage 22:1732-1743. Satz P (1993) Brain reserve capacity on symptom onset after brain injury: a formulation and review of evidence for threshold theory. Neuropsychology 7:273-295. Scahill RI, Schott JM, Stevens JM, Rossor MN, Fox NC (2002) Mapping the evolution of regional atrophy in Alzheimer's disease: Unbiased analysis of fluid-registered serial MRI Proc Natl Acad Sci USA. 99:4703-4707. Scarmeas N, Albert SM, Manly JJ, Stern Y (2006) Education and rates of cognitive decline in incident Alzheimer's disease. J Neurol Neurosurg Psychiatry 77:308-316. Scarmeas N, Stern Y (2004) Cognitive reserve: implications for diagnosis and prevention of Alzheimer's disease. Curr Neurol Neurosci Rep 4:374-380.

115

Schmidtke K, Hull M (2002) Neuropsychological differentiation of small vessel disease, Alzheimer's disease and mixed dementia. J Neurol Sci 203:17-22. Schott JM, Fox NC, Frost C, Scahill RI, Janssen JC, Chan D, Jenkins R, Rossor MN (2003) Assessing the onset of structural change in familial Alzheimer's disease. Ann Neurol 53:181-188. Schott JM, Price SL, Frost C, Whitwell JL, Rossor MN, Fox NC (2005) Measuring atrophy in Alzheimer disease - A serial MRI study over 6 and 12 months. Neurology 65:119-124. Selkoe DJ (2004) Cell biology of protein misfolding: The examples of Alzheimer's and Parkinson's diseases. Nat Cell Biol 6:1054-1061. Selkoe DJ (2006) The ups and downs of A beta. Nat Med 12:758-759. Shaw P, Greenstein D, Lerch J, Clasen L, Lenroot R, Gogtay N, Evans A, Rapoport J, Giedd J (2006) Intellectual ability and cortical development in children and adolescents. Nature 440:676-679. Shawber CJ, Kitajewski J (2004) Notch function in the vasculature: insights from zebrafish, mouse and man. Bioessays 26:225-234. Silbert LC, Quinn JF, Moore MM, Corbridge E, Ball MJ, Murdoch G, Sexton G, Kaye JA (2003) Changes in premorbid brain volume predict Alzheimer's disease pathology. Neurology 61:487-492. Singer JD, Willett JB (2003) Applied longitudinal data analysis: modeling change and event occurrence. Oxford, UK: Oxford University Press.

116

Sluming V, Barrick T, Howard M, Cezayirli E, Mayes A, Roberts N (2002) Voxel-based morphometry reveals increased gray matter density in Broca's area in male symphony orchestra musicians. Neuroimage 17:1613-1622. Smith SM, Jenkinson M, Woolrich MW, Beckmann CF, Behrens T, Johansen-Berg H, Bannister PR, De Luca M, Drobnjak I, Flitney DE, Niazy RK, Saunders J, Vickers J, Zhang YY, De Stefano N, Brady JM, Matthews (2004) Advances in functional and structural MR image analysis and implementation as FSL. Neuroimage 23 Suppl 1:S208-S219. Smith SM, Zhang YY, Jenkinson M, Chen J, Matthews PM, Federico A, De Stefano N (2002) Accurate, robust, and automated longitudinal and cross-sectional brain change analysis. Neuroimage 17:479-489. Smith SM (2002) Fast robust automated brain extraction. Hum Brain Mapp 17:143-155. Snowdon DA, Greiner LH, Markesbery WR (2000) Linguistic ability in early life and the neuropathology of Alzheimer's disease and cerebrovascular disease - Findings from the Nun study. Ann NY Acad Sci 903:34-38. Snowdon DA, Kemper SJ, Mortimer JA, Greiner LH, Wekstein DR, Markesbery WR (1996) Linguistic ability in early life and cognitive function and Alzheimer's disease in late life - Findings from the Nun Study. JAMA 275:528-532. Snowdon DA (2003) Healthy aging and dementia - Findings from the nun study. Ann Intern Med 139:450-454.

117

Snyder AZ (1996) Difference image vs ratio image error function forms in PET-PET realignment. In: Quantification of brain function using PET (Bailey D, Jones T, eds), pp 131-137. San Diego: Academic Press. Sowell ER, Peterson BS, Thompson PM, Welcome SE, Henkenius AL, Toga AW (2003) Mapping cortical change across the human life span. Nat Neurosci 6:309-315. Sowell ER, Thompson PM, Leonard CM, Welcome SE, Kan E, Toga AW (2004) Longitudinal mapping of cortical thickness and brain growth in normal children. J Neurosci 24:8223-8231. Sowell ER, Thompson PM, Toga AW (2004) Mapping changes in the human cortex throughout the span of life. Neuroscientist 10:372-392. Spaan P, Raaijmakers J, Jonker C (2003) Alzheimer's disease versus normal ageing: A review of the efficiency of clinical and experimental memory measures. J Clin Exp Neuropsychol 25:216-233. Stern Y, Alexander GE, Prohovnik I, Mayeux R (1992) Inverse relationship between education and parietotemporal perfusion deficit in Alzheimer's disease. Ann Neurol 32:371-375. Stern Y, Alexander GE, Prohovnik I, Stricks L, Link B, Lennon MC, Mayeux R (1995) Relationship between lifetime occupation and parietal flow: implications for a reserve against Alzheimer's disease pathology. Neurology 45:55-60. Stern Y (2002) What is cognitive reserve? Theory and research application of the reserve concept. J Int Neuropsychol Soc 8:448-460.

118

Stern Y (2006) Cognitive reserve and Alzheimer disease. Alzheimer Dis Assoc Dis 20:112-117. Storandt M, Grant E, Miller J, Morris J (2006) Longitudinal course and neuropathologic outcomes in original vs revised MCI and in pre-MCI. Neurology 67:467-473. Sullivan EV, Rosenbloom M, Serventi KL, Pfefferbaum A (2004) Effects of age and sex on volumes of the thalamus, pons, and cortex. Neurobiol Aging 25:185-192. Svennerholm L, Bostrom K, Jungbjer B (1997) Changes in weight and compositions of major membrane components of human brain during the span of adult human life of Swedes. Acta Neuropathol 94:345-352. Swaab DF, Dubelaar E, Hofman MA, Scherder E, van Someren E, Verwer R (2002) Brain aging and Alzheimer's disease; use it or lose it. Prog Brain Res 138:343-373. Talairach J, Tournoux P (1988) Co-planar stereotaxic atlas of the human brain: an approach to medical cerebral imaging. New York: Thieme Medical Publishers. Tang Y, Whitman GT, Lopez I, Baloh RW (2001) Brain volume changes on longitudinal magnetic resonance imaging in normal older people. J Neuroimaging 11:393-400. Tanzi RE, Bertram L (2005) Twenty years of the Alzheimer's disease amyloid hypothesis: A genetic perspective. Cell 120:545-555. Terry RD, Masliah E, Salmon DP, Butters N, DeTeresa R, Hill R, Hansen LA, Katzman R (1991) Physical basis of cognitive alterations in Alzheimer's disease: synapse loss is the major correlate of cognitive impairment. Ann Neurol 30:572-580.

119

Thompson PM, Hayashi KM, de Zubicaray G, Janke AL, Rose SE, Semple J, Herman D, Hong MS, Dittmer SS, Doddrell DM, Toga AW (2003) Dynamics of gray matter loss in Alzheimer's disease. J Neurosci 23:994-1005. Tisserand DJ, Pruessner JC, Arigita E, van Boxtel M, Evans AC, Jolles J, Uylings H (2002) Regional frontal cortical volumes decrease differentially in aging: An MRI study to compare volumetric approaches and voxel-based morphometry. Neuroimage 17:657-669. Toga AW, Thompson PM (2003) Temporal dynamics of brain anatomy. Annu Rev Biomed Eng 5:119-145. Tuokko H, Garrett DD, McDowell I, Silverberg N, Kristjansson B (2003) Cognitive decline in high-functioning older adults: reserve or ascertainment bias? Aging Ment Health 7:259-270. Turlejski K, Djavadian R (2002) Life-long stability of neurons: a century of research on neurogenesis, neuronal death and neuron quantification in adult CNS. Prog Brain Res 136:39-65. Urbanc B, Cruz L, Le R, Sanders J, Ashe KH, Duff K, Stanley HE, Irizarry MC, Hyman BT (2002) Neurotoxic effects of thioflavin S-positive amyloid deposits in transgenic mice and Alzheimer's disease. Proc Natl Acad Sci USA. 99:1399013995. Uylings H, de Brabander JM (2002) Neuronal changes in normal human aging and Alzheimer's disease. Brain Cogn 49:268-276.

120

Valenzuela MJ, Sachdev P (2006) Brain reserve and dementia: a systematic review. Psychol Med 36:441-454. Van der Kouwe A (2005) MGH neuroanatomical imaging package. Technical report. Van Laere KJ, Dierckx RA (2001) Brain perfusion SPECT: Age-and sex-related effects correlated with voxel-based morphometric findings in healthy adults. Radiology 221:810-817. Van Petten C (2004) Relationship between hippocampal volume and memory ability in healthy individuals across the lifespan: review and meta-analysis. Neuropsychologia 42:1394-1413. Villareal DT, Grant E, Miller JP, Storandt M, McKeel DW, Morris JC (2003) Clinical outcomes of possible versus probable Alzheimer's disease. Neurology 61:661-667. Vincent JL, Snyder AZ, Fox MD, Shannon BJ, Andrews JR, Raichle ME, Buckner RL (2006) Coherent Spontaneous Activity Identifies a Hippocampal-Parietal Mnemonic Network. J Neurophysiol, in press. Wagner AD, Shannon BJ, Kahn I, Buckner RL (2005) Parietal lobe contributions to episodic memory retrieval. Trends Cogn Sci 9:445-453. Walhovd KB, Fjell AM, Reinvang I, Lundervold A, Fischl B, Salat D, Quinn BT, Makris N, Dale AM (2005) Cortical volume and speed-of-processing are complementary in prediction of performance intelligence. Neuropsychologia 43:704-713. Walsh DM, Selkoe DJ (2004) Deciphering the molecular basis of memory failure in Alzheimer's disease. Neuron 44:181-193.

121

Wan H, Sengupta M, Velkofff VA, DeBarros AK (2005) 65+ in the United States: 2005. Washington, DC: U.S. Census Bureau. Wang DM, Chalk JB, Rose SE, de Zubicaray G, Cowin G, Galloway GJ, Barnes D, Spooner D, Doddrell DM, Semple J (2002) MR image-based measurement of rates of change in volumes of brain structures. Part II: Application to a study of Alzheimer's disease and normal aging. Magn Reson Imaging 20:41-48. Wang DM, Doddrell DM (2002) MR image-based measurement of rates of change in volumes of brain structures. Part 1: method and validation. Magn Reson Imaging 20:27-40. Wang HL, Yuan HS, Shu L, Xie JX, Zhang D (2004) Prolongation of T-2 relaxation times of hippocampus and amygdala in Alzheimer's disease. Neurosci Lett 363:150153. Weis S, Jellinger K, Wenger E (1991) Morphometry of the corpus callosum in normal aging and Alzheimer's disease. J Neural Transm Suppl 33:35-38. Whalley LJ, Starr JM, Athawes R, Hunter D, Pattie A, Deary IJ (2000) Childhood mental ability and dementia. Neurology 55:1455-1459. Whalley LJ (2002) Brain ageing and dementia: what makes the difference? Br J Psychiatry 181:369-371. Wickett JC, Vernon PA, Lee DH (2000) Relationships between factors of intelligence and brain volume. Pers Individ Differ 29:1095-1122.

122

Wolfson L, Wei XC, Hall CB, Panzer V, Wakefield D, Benson RR, Schmidt JA, Warfield SK, Guttmann C (2005) Accrual of MRI white matter abnormalities in elderly with normal and impaired mobility. J Neurol Sci 232:23-27. Woods RP, Cherry SR, Mazziotta JC (1992) Rapid automated algorithm for aligning and reslicing PET images. J Comput Assist Tomogr 16:620-633. Yoshiura T, Mihara F, Ogomori K, Tanaka A, Kaneko K, Masuda K (2002) Diffusion tensor in posterior cingulate gyrus: correlation with cognitive decline in Alzheimer's disease. Neuroreport 13:2299-2302. Zhang MY, Katzman R, Salmon D, Jin H, Cai GJ, Wang ZY, Qu GY, Grant I, Yu E, Levy P, et al (1990) The prevalence of dementia and Alzheimer's disease in Shanghai, China: impact of age, gender, and education. Ann Neurol 27:428-437. Zhang YY, Brady M, Smith S (2001) Segmentation of brain MR images through a hidden Markov random field model and the expectation-maximization algorithm. IEEE Trans Med Imaging 20:45-57.

123