Correlation between baseline regional gray matter volume and global gray matter volume decline rate

NeuroImage 54 (2011) 743–749

Contents lists available at ScienceDirect

NeuroImage j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / y n i m g

Correlation between baseline regional gray matter volume and global gray matter volume decline rate Yasuyuki Taki a,⁎, Shigeo Kinomura b, Kazunori Sato b, Ryoi Goto b, Kai Wu b, Ryuta Kawashima a,c,d, Hiroshi Fukuda b a

Division of Developmental Cognitive Neuroscience, Institute of Development, Aging and Cancer, Tohoku University, Sendai, Japan Department of Nuclear Medicine & Radiology, Institute of Development, Aging and Cancer, Tohoku University, Sendai, Japan Department of Functional Brain Imaging, Institute of Development, Aging and Cancer, Tohoku University, Sendai, Japan d Smart Ageing International Research Centre, Institute of Development, Aging and Cancer, Tohoku University, Sendai, Japan b c

a r t i c l e

i n f o

Article history: Received 29 June 2010 Revised 23 August 2010 Accepted 26 September 2010 Available online 20 October 2010 Keywords: Aging Gray matter Magnetic resonance imaging Volumetry Voxel-based morphometry Longitudinal

a b s t r a c t Evaluating whole-brain or global gray matter volume decline rate is important in distinguishing neurodegenerative diseases from normal aging and in anticipating cognitive decline over a given period in non-demented subjects. Whether a significant negative correlation exists between baseline regional gray matter volume of several regions and global gray matter volume decline in the subsequent time period in healthy subjects has not yet been clarified. Therefore, we analyzed the correlation between baseline regional gray matter volumes and the rate of global gray matter volume decline in the period following baseline using magnetic resonance images of the brains of 381 healthy subjects by applying a longitudinal design over 6 years using voxel-based morphometry. As a result, the annual percentage change in gray matter ratio (GMR, APCGMR), in which GMR represents the percentage of gray matter volume in the intracranial volume, showed a significant negative correlation with the baseline regional gray matter volumes of the right posterior cingulate cortex/precuneus and the left hippocampus. Additionally, baseline regional gray matter volume of both the right PCC/precuneus and the left hippocampus significantly distinguished whether the APCGMR was above or below the mean of APCGMR. Our results suggest that baseline regional gray matter volume predicts the rate of global gray matter volume decline in the subsequent period in healthy subjects. Our study may contribute to distinguishing neurodegenerative diseases from normal aging and to predicting cognitive decline. © 2010 Elsevier Inc. All rights reserved.

Introduction Structural neuroimaging studies applying longitudinal design show that whole-brain or global gray matter volume declines with age (Resnick et al., 2003; Scahill et al., 2003; Fotenos et al., 2005; Raz et al., 2005; Taki et al., in press) in healthy elderly humans. In Alzheimer's disease (AD), which leads to more whole-brain volume or global gray matter volume decline compared with age-matched healthy subjects (Busatto et al., 2008; Raji et al., 2009), AD patient whole-brain volume decline rate is more than double that of agematched non-demented elderly individuals (Fotenos et al., 2005). Even in non-demented participants, a high whole-brain atrophy rate is associated with an increased risk of progression to dementia (Sluimer et al., 2008; Henneman et al., 2009). Moreover, whole-brain atrophy rate is significantly greater in non-demented subjects who ⁎ Corresponding author. Division of Developmental Cognitive Neuroscience, Institute of Development, Aging & Cancer, Tohoku University, 4-1 Seiryocho, Aobaku, 980-8575 Sendai, Japan. Fax: + 81 22 717 8457. E-mail address: [email protected] (Y. Taki). 1053-8119/$ – see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.neuroimage.2010.09.071

convert to mild cognitive impairment (MCI) or AD than in those who remain stable (Jack et al., 2004). Therefore, evaluating whole-brain or global gray matter volume decline rate is important in distinguishing neurodegenerative diseases from normal aging and in anticipating cognitive decline over a given period in non-demented subjects. If the decline rate of whole-brain or gray matter volumes in the subsequent period can be anticipated using the information from baseline brain magnetic resonance (MR) images, then this rate determination might contribute to the prevention of subsequent cognitive decline and to early diagnosis of neurodegenerative diseases in several clinical settings, such as the brain disease screening system, or “brain checkup”, which is commonly performed in Japan on healthy adults to exclude asymptomatic brain diseases. In this longitudinal study, we focused on the rate of decline of the global gray matter volume instead of whole-brain volume and also focused on regional gray matter volume as information for baseline brain MR images. The reason for focusing on the rate of decline of the global gray matter volume instead of whole-brain volume, which consists of both the gray matter volume and the white matter volume, was that age-related whole-brain volume decline is mainly due to the

744

Y. Taki et al. / NeuroImage 54 (2011) 743–749

gray matter volume decline compared with white matter (Taki et al., 2004, in press), and several cognitive functions are associated with gray matter volume (Fjell et al., 2006; Zimmerman et al., 2006; Kramer et al., 2007). In addition, the reason for focusing on regional gray matter volume as baseline information for brain MR images was that the regional gray matter volume in several regions, such as medial occipito-parietal regions and posterior cingulated cortex, reflects the progression of subsequent cognitive decline in patients with AD (Kinkingnehun et al., 2008) and in MCI (Hamalainen et al., 2007). Moreover, because several cognitive functions, such as executive function and attention, are associated with gray matter volume in healthy elderly individuals (Zimmerman et al., 2006; Kramer et al., 2007), we postulated that baseline regional gray matter volume of some regions may have a significant correlation with the rate of the global gray matter volume decline in the subsequent period in healthy subjects. We focused on a significant negative correlation (i.e., no positive correlation) between baseline regional gray matter volume and the global gray matter volume decline rate in the subsequent period. A significant negative correlation between the above variables would indicate that subjects with smaller baseline regional gray matter volumes in these regions had a higher rate of global gray matter volume decline, whereas significant positive correlation would indicate that subjects with smaller baseline regional gray matter volumes in these regions had a lower rate of global gray matter volume decline or more preserved global gray matter volume in the subsequent period. Therefore, detecting baseline gray matter regions that have significant negative correlations with global gray matter volume decline rate in the subsequent period is important in clinical settings because it might contribute to the prevention of the subsequent cognitive decline and to early diagnosis of neurodegenerative diseases in several clinical settings. However, whether a significant negative correlation exists between baseline regional gray matter volume of several regions and the rate of the global gray matter volume decline in the subsequent period in healthy subjects has not yet been clarified. Therefore, the purpose of the present study was to determine whether a negative correlation exists between baseline regional gray matter volumes and the rate of global gray matter volume decline in the following period by applying a longitudinal design over 6 years using brain MR images in community-dwelling healthy subjects who encompass a large age range (mean age = 51.2 ± 11.8 years, range = 21–80 years). We applied volumetric analysis and voxelbased morphometry (VBM) (Ashburner and Friston, 2000) in MR image analysis. VBM is an established automated neuroimaging technique that enables the global analysis of brain structure without a priori identification of a region of interest. VBM is not biased toward any specific brain region, and it permits the identification of unsuspected potential brain structural differences or abnormalities. In the volumetric analysis, we calculated global gray matter volume and intracranial volume at baseline and follow-up applying a fully automated method. Next, we calculated the gray matter ratio (GMR), which represents the percentage of gray matter volume divided by intracranial volume, to normalize the differences in the head size of subjects (Taki et al., 2004). Then we obtained the rate of decline in GMR by calculating the annual percentage change of GMR (APCGMR) from these results (Taki et al., in press). Next, to obtain baseline regional gray matter volume, we applied VBM image analysis. We hypothesized that there would be a significant negative correlation between the baseline regional gray matter volume of the PCC, cuneus, precuneus, or hippocampus and APCGMR in healthy subjects due to the following: (1) baseline regional gray matter volume of medial occipito-parietal regions, such as the cuneus and the precuneus, and posterior cingulate cortex (PCC) reflects the progression of cognitive decline, which is thought to link with atrophy of gray matter volume in the subsequent period in patients with AD (Kinkingnehun et al., 2008) and in MCI (Hamalainen et al., 2007), and (2) a recent

longitudinal study showed that baseline hippocampal volume significantly predicted progression of cognitive decline in nondemented subjects (Henneman et al., 2009). Methods Subjects All subjects were Japanese individuals recruited from our previous brain-imaging project, a part of the Aoba Brain Imaging Project (Sato et al., 2003). We selected participants who had lived in Sendai City (the center of the previous project) at the time of the previous study, whose collected data had no missing values and who had no serious medical problems from an initial 1604 eligible persons. Eligible persons were contacted by mail and telephone and invited to participate in the longitudinal follow-up. Of 930 eligible persons (442 men and 488 women), 513 (55.1%) responded to the invitation, and 469 (50.4% of the total eligible pool, 91.4% of the responders) agreed to participate in the present study. All participants were screened with a mail-in health questionnaire and underwent telephone and personal interviews. Of 469 persons who agreed to participate, 442 (94.2%) completed the follow-up study. Screening criteria applied to the follow-up sample were the same as those that had been used to determine eligibility at entry. Persons who reported a history of any malignant tumor, head trauma with loss of consciousness for N5 min, cerebrovascular disease, epilepsy, any psychiatric disease, or claustrophobia were excluded from the study. Of the 442 subjects who completed the follow-up, 54 (12.2%) were excluded from the present study because they no longer met the health screening criteria. All subjects were screened for dementia using the Mini-Mental State Examination (MMSE) (Folstein et al., 1975), with a cutoff of 26. Two subjects were excluded from the study based on the MMSE (one male and one female, who each scored 25). All subjects underwent measurements of height, weight, and waist and hip circumferences in indoor clothing without shoes. Blood pressure in the right brachial artery was measured in the sitting position after a 10-min rest. An experienced neuroradiologist examined the MR scans for any tumors and cerebrovascular disease. Follow-up MRI data on an additional eight subjects (1.8%) were unsuitable for the longitudinal analysis (4 subjects showed brain tumors and 4 subjects showed cerebral infarct). Thus, the final sample consisted of 381 participants (40.1% of the eligible cohort: 158 men, 223 women). The mean ± SD follow-up interval was 7.42 ± 0.55 years (range, 6.2–9.0). The mean ± SD age of the participants at baseline was 51.2 ± 11.8 years (range, 21–80 years). Characteristics of the subjects are shown in Table 1. Of the participants, 11 subjects (mean age, 65.3 years; range, 57.7–73.4 years at follow-up; 3 men, 8 women) were scanned twice on the same day to permit an estimation of measurement reliability. Written informed consent according to the Declaration of Helsinki (1991) was obtained from each subject after a full explanation of the purpose and procedures of the study and prior to MR image scanning. Approval for these experiments was obtained from the institutional review board of Tohoku University. Image acquisition All images were collected using the same 0.5-T MR scanner (Signa contour; GE-Yokogawa Medical Systems, Tokyo, Japan), including baseline images. The scanner was routinely calibrated using the same standard GE phantom between baseline and follow-up. During this period, there no major hardware upgrade occurred. At baseline and follow-up, all subjects were scanned with identical pulse sequences: 124 contiguous, 1.5-mm-thick axial planes of three-dimensional T1weighted images (spoiled gradient recalled acquisition in steady

Y. Taki et al. / NeuroImage 54 (2011) 743–749 Table 1 Characteristics of the subjects at follow-up. Factor

Age (years) BMI (kg/m2) Systolic BP (mm Hg) Diastolic BP (mm Hg) Lifetime alcohol intake (kg) Duration of education (years) MMSE Interval between scans (years)

Men (n = 158)

Women (n = 223)

P

Mean (SD)

Range

Mean (SD)

Range

60.0 (11.6) 24.34 (2.98) 132.7 (18.2)

28–87 15.8–34.3 91–202

57.6 (12.0) 22.76 (3.27) 129.2 (19.0)

28–89 16.5–34.9 91–182

0.057 b 0.001 0.068

83.3 (11.5)

49–111

79.4 (11.8)

46–113

0.001

160.4a

0–1705

0.0a

0–1270

b 0.001 b 0.001

14.3 (3.20)

8–29

13.1 (2.24)

8–23

29.2 (1.19) 7.42 (0.56)

26–30 6.53–8.99

29.2 (1.15) 7.38 (0.53)

26–30 6.32–8.86

0.649 0.557

SD: standard deviation. BMI: body mass index (kg/m2). BP: blood pressure. MMSE: Mini-Mental State Examination. a Median instead of mean (SD) because the distribution of lifetime alcohol intake is quite different from a normal distribution.

state: repetition time (TR), 40 ms; echo time (TE), 7 ms; flip angle, 30; voxel size, 1.02 × 1.02 × 1.5 mm). Image analysis of volumetric analysis After the image acquisition, all T1-weighted MR images were analyzed using statistical parametric mapping 2 (SPM2) (Wellcome Department of Cognitive Neurology, London, UK; http://www.fil.ion. ucl.ac.uk/spm/software/spm2/) (Friston et al., 1995) in Matlab (Math Works, Natick, MA). First, the T1-weighted MR images of both baseline and follow-up sessions were transformed into the same stereotaxic space by registering each image to the same template image. The template image we used was the ICBM 152 template (Montreal Neurological Institute), which was derived from 152 normal subjects and which approximates the Talairach space (Talairach and Tournoux, 1988). This process is called linear normalization. Then tissue segmentation from the raw images to the gray matter, white matter, cerebrospinal fluid (CSF) space, and non-brain tissue segments was performed using the SPM2 default segmentation procedure. We applied these processes using the Matlab file “cg_vbm_optimized” (http://www.dbm.neuro.uni-jena. de/vbm.html). The voxel values of each segmented image consisted not of binary (i.e., 0 or 1), but of 256 grade (i.e., between 0/255 and 255/255) signal intensities according to their tissue probability. The linear-normalized, segmented images were restored to the native space to determine the volume of each segment. The actual volumes of the entire normalized, segmented, and restored segmented images were determined by summing up each voxel volume (all voxel volumes were 1 mm3) multiplied by each voxel value divided by 255. To normalize the head size of each subject, we defined the GMR as the percentage of gray matter volume divided by the intracranial volume. Next, to reveal the annualized rate of change in GMR with age, we determined the annual percentage change in GMR (APCGMR) for each subject. APCGMR was calculated as APCGMR = (GMR1 – GMR2) / T × 100, where GMR1 is the GMR at baseline, GMR2 is the GMR at follow-up, and T is interval between baseline and follow-up (in years).

745

segmentation procedure. Next, the segmented gray matter images were nonlinearly normalized to the gray matter SPM2 template using 7 × 8 × 7 nonlinear basis functions in three orthogonal directions. These normalization parameters were reapplied to the T1-weighted whole-brain structural images of each subject to perform optimal spatial normalization. The optimally normalized T1-weighted images were segmented into the gray matter, white matter, and CSF space. The normalized, segmented gray matter images were then modulated by calculating the Jacobian determinants derived from the spatial normalization step, and each voxel was multiplied by the relative change in volume as in the method of a previous study (Good et al., 2001). This modulation step was performed to correct for volume changes in nonlinear normalization. The normalized, segmented, and modulated gray matter images were smoothened by convoluting a 12-mm-FWHM isotropic Gaussian kernel. This smoothing step was applied to remove individual variations in gyral anatomy and to render data more normally distributed by the central limit theorem. A schematic of the volumetric analysis and VBM is shown in Fig. 1. Statistical analysis To investigate the correlation between baseline regional gray matter volume and APCGMR, we performed a multiple regression analysis with age, gender, intracranial volume, and APCGMR as independent valuables and baseline regional gray matter volume as a dependent valuable. We used the random field theory method to correct for the Familywise Error Rate (FWER); any resulting P-value less than 0.05 was considered significant. Next, we tested whether the gray matter regional volume that showed the significant negative correlation with APCGMR at baseline could predict whether the APCGMR was above or below the APCGMR mean by applying a standard (not stepwise) linear discriminant analysis in SPSS11.5. For the discriminant analysis, we used the mean gray matter volume over a cluster in each region and the regional gray matter volume as defined by multiple regression analysis. We set the significance level at P b 0.05. In addition, to estimate the measurement reliability of volumetric analysis, 11 randomly selected subjects were scanned twice on the same day; no significant difference in the calculated gray matter volume (t = −0.374; P = 0.716; paired t-test) or intracranial volume (t = −0.627; P = 0.545; paired t-test) was observed between the two scans. Results Volumetric analysis Gray matter volume at baseline was significantly larger in males than in females (t = 10.88, P b 0.001). Conversely, no significant difference in GMR at baseline was found between males and females (t = 1.52, P = 0.129). The intracranial volume at baseline was significantly larger in males than in females (t = 14.89, P b 0.001). APCGMR was significantly larger in males than in females (t = 5.89, P b 0.001). The distribution of APCGMR for each gender is shown in Fig. 2. The distribution of APCGMR in each gender was similar to the normal distribution. Because there was a significant difference in APCGMR between genders, we adjusted for gender effects in the next VBM analysis. In addition, because there was a significant difference in intracranial volumes between genders, we also adjusted for intracranial volume in the next VBM analysis.

Image analysis of VBM VBM First, using the linearly normalized T1-weighted MR images derived from the volumetric image analysis of the baseline scan, tissue segmentation from the transformed images to the gray matter, white matter, and CSF space was performed using the SPM2 default

Gray matter regions showing significant negative correlation with APCGMR adjusted for age, gender, and intracranial volume are shown in Fig. 3. Baseline regional gray matter volumes of the right PCC/

746

Y. Taki et al. / NeuroImage 54 (2011) 743–749

Fig. 1. A schematic representation of the volumetric analysis and voxel-based morphotmetry.

precuneus and the left hippocampus showed significant negative correlations with APCGMR after adjusting for age, gender, and intracranial volume (right PCC/precuneus, t = 5.42, P = 0.020; left

hippocampus, t = 5.29, P = 0.035). Therefore, we used the gray matter regions of the right PCC/precuneus and the left hippocampus in the next discriminant analysis. Discriminant analysis Baseline regional gray matter volume of both the right PCC/ precuneus and the left hippocampus significantly distinguished whether APCGMR was above or below the APCGMR mean. The Fvalue, P-value, and discriminant function coefficient were 13.51, b0.001, and 0.833 in the right PCC/precuneus, and 5.71, 0.017, and 0.350 in the left hippocampus, respectively. Overall, 58.4% of the APCGMR (55.8% of APCGMR below the mean of APCGMR and 60.9% of APCGMR above the mean of APCGMR) was correctly distinguished using the discriminant function. Discussion

Fig. 2. Distribution of APCGMR in (a) male and (b) female subjects.

This study provides the first longitudinal findings showing that baseline regional gray matter volumes in the right PCC/precuneus, as hypothesized, and in the left hippocampus, also as hypothesized, show a significant negative correlation with the rate of global gray matter volume decline in the following period, as represented by APCGMR, adjusting for age, gender, and intracranial volume. In addition, baseline regional gray matter volumes of both the right PCC/precuneus and the left hippocampus significantly distinguished whether the APCGMR was above or below the APCGMR mean. These results indicate that subjects who had smaller baseline regional gray matter volumes in those regions showed higher rate of global gray matter volume decline in the following period. Two interpretations of these findings occurred to us. First, subjects who have smaller baseline regional gray matter volumes in those regions, either intrinsically or due to the effects of aging, are at risk of a higher rate of global gray matter volume decline in the following period. To assess the validity of this view, consideration should be given to whether a significant correlation exists between baseline regional gray matter volume of these regions and age. Although significant negative correlations between age and baseline regional gray matter volume

Y. Taki et al. / NeuroImage 54 (2011) 743–749

Fig. 3. Gray matter regions showing significant negative correlations with annual percent change of the gray matter ratio (APCGMR) adjusted for age, gender, and intracranial volume. The results of the correlation were superimposed onto structural MR images of axial, coronal, and sagittal views. The left side of the image represents the left side of the brain. Color scales indicate the t-score. To clarify the extent of the significantly correlated regions, we show data, the significance level of which was set at p b 0.05, corrected for false discovery rate. PCC indicates posterior cingulate cortex. In the lowest line, relationship and regression line between baseline regional gray matter volume and APCGMR in each region adjusting for age, gender, and intracranial volume are shown. The 95% confidence interval of the regression line is also shown.

have been reported in areas such as the inferior parietal lobule and prefrontal regions (Raz et al., 1997; Courchesne et al., 2000; Good et al., 2001; Resnick et al., 2003; Sowell et al., 2003; Sullivan et al., 2004; Taki et al., 2004; Lemaitre et al., 2005), the baseline regional gray matter volumes such as that of the PCC/precuneus and the hippocampus show more preservation in aging compared with other gray matter regions (Sowell et al., 2003; Grieve et al., 2005; Kalpouzos et al., 2009). Actually, gray matter regions did not overlap, thus revealing a statistically significant negative correlation between age and APCGMR in this study (data not shown). These results indicate that aging affects the posterior cingulate cortex and hippocampus to a lesser degree than other gray matter regions, suggesting that subjects who have smaller baseline regional gray matter volumes of the right

747

posterior cingulate cortex or the left hippocampus intrinsically due to aging are at risk for a higher rate of gray matter volume decline in the subsequent period. The second interpretation is that the significant negative correlation between baseline regional gray matter volumes of the right PCC/ precunei and the left hippocampus and APCGMR reflects a reduction of gray matter volume indicative of a pathological process as well as a normal aging process. This interpretation is derived from findings of recent studies dealing with patients with AD and beta-amyloid deposition. The deposition of beta-amyloid in senile plaque is a characteristic feature of AD (Masters et al., 1985), and recent studies have revealed that 11C-6-OH benzothiazole (11C-PIB) uptake of cortical gray matter, especially of the PCC, shows a significant correlation with the rate of cerebral atrophy in patients with Alzheimer's disease (Archer et al., 2006). 11C-PIB is an in vivo diagnostic marker for beta-amyloid in AD (Bacskai et al., 2003; Klunk et al., 2004). In addition, the deposition of beta-amyloid is found in several gray matter regions such as the PCC and the hippocampus even in healthy subjects whose cognitive functions are well preserved (Price and Morris, 1999; Bennett et al., 2006), and the deposition of beta-amyloid is thought to spread throughout the whole neocortex (Braak and Braak, 1997). Additionally, the presence of beta-amyloid is associated with neuronal loss (Hardy and Higgins, 1992). From these studies, it is speculated that the significant negative correlation between baseline regional gray matter volume of the precuneus and APCGMR is associated with the presence of beta-amyloid in the gray matter of those regions at one point, spreading to other gray matter regions in the following period, thus accelerating the rate of global gray matter volume loss. However, several issues are difficult to explain in this way. First, a recent postmortem study has presented the controversial finding that beta-amyloid burden is not associated with rate of brain atrophy (Josephs et al., 2008). In addition, only subjects whose score of MMSE 26 or higher were included in this study, and the mean and SD of the MMSE score in a follow-up study were 29.2 ± 1.19 in men and 29.2 ± 1.15 in women, suggesting that few subjects had cognitive impairment even in the follow-up. In any case, our findings are important because they suggest that it is possible to evaluate the rate of global gray matter volume decline in healthy subjects using baseline regional gray matter volumes in the right PCC/precuneus and the left hippocampus. These findings are thought to be especially important in clinical settings such as medical check-ups of the brain, which are prevalent in Japan. We showed that baseline regional gray matter volume of both the right PCC/precuneus and the left hippocampus significantly distinguished whether the APCGMR was above or below the APCGMR mean in the discriminant analysis. This result suggests that the baseline regional gray matter volume of those regions predicts the rate of global gray matter volume decline in the subsequent time period in healthy subjects. Because a high whole-brain atrophy rate was associated with an increased risk of progression to dementia (Sluimer et al., 2008; Henneman et al., 2009), and whole-brain atrophy rate was significantly greater in non-demented subjects who converted to MCI or AD than in those who remained stable (Jack et al., 2004), predicting gray matter atrophy rate in the subsequent time period is important for evaluating cognitive decline in the following period. In our results, although regional gray matter volume at baseline predicted whether the APCGMR was above or below the APCGMR mean in the subsequent period in healthy subjects, the ratio of correct distinctions was 58.4% of the APCGMR (55.8% of APCGMR under the mean of APCGMR and 60.9% of APCGMR above the mean of APCGMR). Because several other factors such as hypertension (Raz et al., 2003, 2005), alcohol consumption (de Bruin et al., 2005; Taki et al., 2006), and obesity (Ward et al., 2005; Pannacciulli et al., 2006; Taki et al., 2008) also affect the gray matter volume decrease, the ratio may increase if we use information about those factors as well as baseline regional gray matter volume.

748

Y. Taki et al. / NeuroImage 54 (2011) 743–749

We showed that APCGMR in males was larger than that in females. In considering the reason for larger APCGMR in male, several factors are thought to affect the decline of gray matter volume. Degenerative changes in gray matter, such as shrinkage or loss of neurons (Terry et al., 1987) and loss of dendritic arborization (Jacobs et al., 1997), are all thought to be observed in gray matter volume decline. In addition, brain maturation, consisting of both regressive cellular events such as synaptic pruning and progressive cellular events such as myelination occur simultaneously in the brain during childhood, adolescence, and young adulthood, and both types of event could result in the appearance of regional gray matter volume decline or cortical thinning on MR images (Sowell et al., 2001). Thus, the decline in gray matter volume observed on MR images is thought to be composed of both degenerative and maturational changes in the gray matter. Regarding gender differences in brain maturation, the volumes of the frontal and parietal gray matter peak earlier in females (Giedd et al., 1999), and the slope of the reduction in gray matter volume in adolescence is steeper in males than in females (De Bellis et al., 2001). Additionally, estradiol delayed synaptic pruning in an animal study (Naftolin et al., 1990), and testosterone has been suggested to be associated with myelinogenesis (Martini and Melcangi, 1991). Thus, earlier brain maturation in women may lead to an estrogen-mediated delay in synaptic pruning. In addition, regarding the generational difference, previous studies have shown that hormone replacement therapy in postmenopausal women is associated with a sparing of gray matter volume, suggesting that estrogen may have neuroprotective effects, as shown in recent studies (Raz et al., 2004; Erickson et al., 2005). From our results, APCGMR in males is larger than that in females. This study had several limitations. First, regarding the MR scanner, it is possible that scanner drift may result in signal changes and altered image quality over time, although the MR scanner that we used was the same for the baseline and follow-up images, and the scanner was routinely calibrated using the same standard GE phantom at both baseline and follow-up and was appropriately maintained. Second, also related to the MR scanner, tissue contrast at 0.5 T was substantially lower than that of higher field strength magnets, so that the use of a less powerful scanner may have affected the results of the tissue segmentation process. However, it is also a strength of this study that we used the same MR scanner for the baseline and follow-up images because a different MR scanner or a different pulse sequence of MR image acquisition also could have affected the tissue segmentation process. Third, regarding the subjects, selection biases may have occurred in that subjects who were concerned about their medical or psychological condition remained in the follow-up study. Fourth, regarding image processing, the possibility of misclassification during tissue segmentation, such as the classification of white matter hyperintensities as gray matter, cannot be dismissed. Although we cannot rule out misclassification in tissue segmentation using our fully automated method, in order to reduce the possibility of tissue misclassification, we used not only voxel intensity itself but also a priori knowledge of the normal location of gray matter, white matter, and CSF to instruct the segmentation process. Considering these limitations, a fully automated method of MR image processing was a strength in dealing with a large quantity of data objectively and efficiently. In summary, using a longitudinal design over 6 years in 381 community-dwelling healthy individuals, we examined the correlation between baseline regional gray matter volume and the rate of global gray matter volume decline in the following period. We found a significant negative correlation between APCGMR and the baseline regional gray matter volumes of the right PCC/precunei and the left hippocampus after adjusting for age and gender. In addition, baseline regional gray matter volume of both the right PCC/precuneus and the left hippocampus significantly distinguished whether the APCGMR was above or below the APCGMR mean. Our results suggest that baseline

regional gray matter volume predicts the rate of global gray matter volume decline in the following period in healthy subjects. Our study may contribute to distinguishing neurodegenerative diseases from normal aging and to predicting cognitive decline. Acknowledgments We thank K. Inoue and K. Okada for insightful comments and K. Inaba, K. Saito, N. Ishibashi, and H. Masuyama for technical help in collecting data. This study was funded by the National Institute of Mental Health, the National Institute of Neurological Disorders and Stroke, the National Institute on Drug Abuse, and the US National Cancer Institute. Part of this research was supported by a grant from the Telecommunications Advancement Organisation of Japan. This study was also supported in part by the 21st Century Centre of Excellence (COE) Programme (Ministry of Education, Culture, Sports, Science and Technology; MEXT), entitled “Future Medical Engineering-based Bionanotechnology” at Tohoku University, and a grant from the JSPSCIHR Joint Health Research Programme. This work was also supported by the MEXT Grant-in-Aid for Young Scientists (B), 18790864. References Archer, H.A., Edison, P., Brooks, D.J., Barnes, J., Frost, C., Yeatman, T., Fox, N.C., Rossor, M.N., 2006. Amyloid load and cerebral atrophy in Alzheimer's disease: an 11C-PIB positron emission tomography study. Ann. Neurol. 60, 145–147. Ashburner, J., Friston, K.J., 2000. Voxel-based morphometry—the methods. Neuroimage 11, 805–821. Bacskai, B.J., Hickey, G.A., Skoch, J., Kajdasz, S.T., Wang, Y., Huang, G.F., Mathis, C.A., Klunk, W.E., Hyman, B.T., 2003. Four-dimensional multiphoton imaging of brain entry, amyloid binding, and clearance of an amyloid-beta ligand in transgenic mice. Proc. Natl Acad. Sci. USA 100, 12462–12467. Bennett, D.A., Schneider, J.A., Arvanitakis, Z., Kelly, J.F., Aggarwal, N.T., Shah, R.C., Wilson, R.S., 2006. Neuropathology of older persons without cognitive impairment from two community-based studies. Neurology 66, 1837–1844. Braak, H., Braak, E., 1997. Frequency of stages of Alzheimer-related lesions in different age categories. Neurobiol. Aging 18, 351–357. Busatto, G.F., Diniz, B.S., Zanetti, M.V., 2008. Voxel-based morphometry in Alzheimer's disease. Expert Rev. Neurother. 8, 1691–1702. Courchesne, E., Chisum, H.J., Townsend, J., Cowles, A., Covington, J., Egaas, B., Harwood, M., Hinds, S., Press, G.A., 2000. Normal brain development and aging: quantitative analysis at in vivo MR imaging in healthy volunteers. Radiology 216, 672–682. De Bellis, M.D., Keshavan, M.S., Beers, S.R., Hall, J., Frustaci, K., Masalehdan, A., Noll, J., Boring, A.M., 2001. Sex differences in brain maturation during childhood and adolescence. Cereb. Cortex 11, 552–557. de Bruin, E.A., Pol, H.E., Schnack, H.G., Janssen, J., Bijl, S., Evans, A.C., Kenemans, J.L., Kahn, R.S., Verbaten, M.N., 2005. Focal brain matter differences associated with lifetime alcohol intake and visual attention in male but not in female non-alcoholdependent drinkers. Neuroimage 26, 536–545. Erickson, K.I., Colcombe, S.J., Raz, N., Korol, D.L., Scalf, P., Webb, A., Cohen, N.J., McAuley, E., Kramer, A.F., 2005. Selective sparing of brain tissue in postmenopausal women receiving hormone replacement therapy. Neurobiol. Aging 26, 1205–1213. Fjell, A.M., Walhovd, K.B., Reinvang, I., Lundervold, A., Salat, D., Quinn, B.T., Fischl, B., Dale, A.M., 2006. Selective increase of cortical thickness in high-performing elderly —structural indices of optimal cognitive aging. Neuroimage 29, 984–994. Folstein, M.F., Folstein, S.E., McHugh, P.R., 1975. “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. J. Psychiatr. Res. 12, 189–198. Fotenos, A.F., Snyder, A.Z., Girton, L.E., Morris, J.C., Buckner, R.L., 2005. Normative estimates of cross-sectional and longitudinal brain volume decline in aging and AD. Neurology 64, 1032–1039. Friston, K.J., Holmes, A.P., Worsley, K.J., Poline, J.-P., Frith, C.D., Frackowiak, R.S.J., 1995. Statistical parametric maps in functional imaging: a general linear approach. Hum. Brain Mapp. 2, 189–210. Giedd, J.N., Blumenthal, J., Jeffries, N.O., Castellanos, F.X., Liu, H., Zijdenbos, A., Paus, T., Evans, A.C., Rapoport, J.L., 1999. Brain development during childhood and adolescence: a longitudinal MRI study. Nat. Neurosci. 2, 861–863. Good, C.D., Johnsrude, I.S., Ashburner, J., Henson, R.N.A., Friston, K.J., Frackowiak, R.S.J., 2001. A voxel-based morphometric study of ageing in 465 normal adult human brains. Neuroimage 14, 21–36. Grieve, S.M., Clark, C.R., Williams, L.M., Peduto, A.J., Gordon, E., 2005. Preservation of limbic and paralimbic structures in aging. Hum. Brain Mapp. 25, 391–401. Hamalainen, A., Tervo, S., Grau-Olivares, M., Niskanen, E., Pennanen, C., Huuskonen, J., Kivipelto, M., Hanninen, T., Tapiola, M., Vanhanen, M., Hallikainen, M., Helkala, E.L., Nissinen, A., Vanninen, R., Soininen, H., 2007. Voxel-based morphometry to detect brain atrophy in progressive mild cognitive impairment. Neuroimage 37, 1122–1131. Hardy, J.A., Higgins, G.A., 1992. Alzheimer's disease: the amyloid cascade hypothesis. Science 256, 184–185.

Y. Taki et al. / NeuroImage 54 (2011) 743–749 Henneman, W.J., Sluimer, J.D., Barnes, J., van der Flier, W.M., Sluimer, I.C., Fox, N.C., Scheltens, P., Vrenken, H., Barkhof, F., 2009. Hippocampal atrophy rates in Alzheimer disease: added value over whole brain volume measures. Neurology 72, 999–1007. Jack Jr., C.R., Shiung, M.M., Gunter, J.L., O'Brien, P.C., Weigand, S.D., Knopman, D.S., Boeve, B.F., Ivnik, R.J., Smith, G.E., Cha, R.H., Tangalos, E.G., Petersen, R.C., 2004. Comparison of different MRI brain atrophy rate measures with clinical disease progression in AD. Neurology 62, 591–600. Jacobs, B., Driscoll, L., Schall, M., 1997. Life-span dendritic and spine changes in areas 10 and 18 of human cortex: a quantitative Golgi study. J. Comp. Neurol. 386, 661–680. Josephs, K.A., Whitwell, J.L., Ahmed, Z., Shiung, M.M., Weigand, S.D., Knopman, D.S., Boeve, B.F., Parisi, J.E., Petersen, R.C., Dickson, D.W., Jack Jr., C.R., 2008. Beta-amyloid burden is not associated with rates of brain atrophy. Ann. Neurol. 63, 204–212. Kalpouzos, G., Chetelat, G., Baron, J.C., Landeau, B., Mevel, K., Godeau, C., Barre, L., Constans, J.M., Viader, F., Eustache, F., Desgranges, B., 2009. Voxel-based mapping of brain gray matter volume and glucose metabolism profiles in normal aging. Neurobiol. Aging 30, 112–124. Kinkingnehun, S., Sarazin, M., Lehericy, S., Guichart-Gomez, E., Hergueta, T., Dubois, B., 2008. VBM anticipates the rate of progression of Alzheimer disease: a 3-year longitudinal study. Neurology 70, 2201–2211. Klunk, W.E., Engler, H., Nordberg, A., Wang, Y., Blomqvist, G., Holt, D.P., Bergstrom, M., Savitcheva, I., Huang, G.F., Estrada, S., Ausen, B., Debnath, M.L., Barletta, J., Price, J.C., Sandell, J., Lopresti, B.J., Wall, A., Koivisto, P., Antoni, G., Mathis, C.A., Langstrom, B., 2004. Imaging brain amyloid in Alzheimer's disease with Pittsburgh Compound-B. Ann. Neurol. 55, 306–319. Kramer, J.H., Mungas, D., Reed, B.R., Wetzel, M.E., Burnett, M.M., Miller, B.L., Weiner, M.W., Chui, H.C., 2007. Longitudinal MRI and cognitive change in healthy elderly. Neuropsychology 21, 412–418. Lemaitre, H., Crivello, F., Grassiot, B., Alperovitch, A., Tzourio, C., Mazoyer, B., 2005. Age- and sex-related effects on the neuroanatomy of healthy elderly. Neuroimage 26, 900–911. Martini, L., Melcangi, R.C., 1991. Androgen metabolism in the brain. J. Steroid Biochem. Mol. Biol. 39, 819–828. Masters, C.L., Simms, G., Weinman, N.A., Multhaup, G., McDonald, B.L., Beyreuther, K., 1985. Amyloid plaque core protein in Alzheimer disease and Down syndrome. Proc. Natl Acad. Sci. USA 82, 4245–4249. Naftolin, F., Garcia-Segura, L.M., Keefe, D., Leranth, C., Maclusky, N.J., Brawer, J.R., 1990. Estrogen effects on the synaptology and neural membranes of the rat hypothalamic arcuate nucleus. Biol. Reprod. 42, 21–28. Pannacciulli, N., Del Parigi, A., Chen, K., Le, D.S., Reiman, E.M., Tataranni, P.A., 2006. Brain abnormalities in human obesity: a voxel-based morphometric study. Neuroimage 31, 1419–1425. Price, J.L., Morris, J.C., 1999. Tangles and plaques in nondemented aging and “preclinical” Alzheimer's disease. Ann. Neurol. 45, 358–368. Raji, C.A., Lopez, O.L., Kuller, L.H., Carmichael, O.T., Becker, J.T., 2009. Age, Alzheimer disease, and brain structure. Neurology 73, 1899–1905. Raz, N., Gunning, F.M., Head, D., Dupuis, J.H., McQuain, J., Briggs, S.D., Loken, W.J., Thornton, A.E., Acker, J.D., 1997. Selective aging of the human cerebral cortex observed in vivo: differential vulnerability of the prefrontal gray matter. Cereb. Cortex 7, 268–282. Raz, N., Rodrigue, K.M., Acker, J.D., 2003. Hypertension and the brain: vulnerability of the prefrontal regions and executive functions. Behav. Neurosci. 117, 1169–1180.

749

Raz, N., Rodrigue, K.M., Kennedy, K.M., Acker, J.D., 2004. Hormone replacement therapy and age-related brain shrinkage: regional effects. NeuroReport 15, 2531–2534. Raz, N., Lindenberger, U., Rodrigue, K.M., Kennedy, K.M., Head, D., Williamson, A., Dahle, C., Gerstorf, D., Acker, J.D., 2005. Regional brain changes in aging healthy adults: general trends, individual differences and modifiers. Cereb. Cortex 15, 1676–1689. Resnick, S.M., Pham, D.L., Kraut, M.A., Zonderman, A.B., Davatzikos, C., 2003. Longitudinal magnetic resonance imaging studies of older adults: a shrinking brain. J. Neurosci. 23, 3295–3301. Sato, K., Taki, Y., Fukuda, H., Kawashima, R., 2003. Neuroanatomical database of normal Japanese brains. Neural Netw. 16, 1301–1310. Scahill, R.I., Frost, C., Jenkins, R., Whitwell, J.L., Rossor, M.N., Fox, N.C., 2003. A longitudinal study of brain volume changes in normal aging using serial registered magnetic resonance imaging. Arch. Neurol. 60, 989–994. Sluimer, J.D., van der Flier, W.M., Karas, G.B., Fox, N.C., Scheltens, P., Barkhof, F., Vrenken, H., 2008. Whole-brain atrophy rate and cognitive decline: longitudinal MR study of memory clinic patients. Radiology 248, 590–598. Sowell, E.R., Thompson, P.M., Tessner, K.D., Toga, A.W., 2001. Mapping continued brain growth and gray matter density reduction in dorsal frontal cortex: Inverse relationships during postadolescent brain maturation. J. Neurosci. 21, 8819–8829. Sowell, E.R., Peterson, B.S., Thompson, P.M., Welcome, S.E., Henkenius, A.L., Toga, A.W., 2003. Mapping cortical change across the human life span. Nat. Neurosci. 6, 309–315. Sullivan, E.V., Rosenbloom, M., Serventi, K.L., Pfefferbaum, A., 2004. Effects of age and sex on volumes of the thalamus, pons, and cortex. Neurobiol. Aging 25, 185–192. Taki, Y., Goto, R., Evans, A., Zijdenbos, A., Neelin, P., Lerch, J., Sato, K., Ono, S., Kinomura, S., Nakagawa, M., Sugiura, M., Watanabe, J., Kawashima, R., Fukuda, H., 2004. Voxelbased morphometry of human brain with age and cerebrovascular risk factors. Neurobiol. Aging 25, 455–463. Taki, Y., Kinomura, S., Sato, K., Goto, R., Inoue, K., Okada, K., Ono, S., Kawashima, R., Fukuda, H., 2006. Both global gray matter volume and regional gray matter volume negatively correlate with lifetime alcohol intake in non-alcohol-dependent Japanese men: a volumetric analysis and a voxel-based morphometry. Alcohol. Clin. Exp. Res. 30, 1045–1050. Taki, Y., Kinomura, S., Sato, K., Inoue, K., Goto, R., Okada, K., Uchida, S., Kawashima, R., Fukuda, H., 2008. Relationship between body mass index and gray matter volume in 1, 428 healthy individuals. Obesity 16, 119–124. Taki, Y., Kinomura, S., Sato, K., Goto, R., Kawashima, R., Fukuda, H., in press. A longitudinal study of gray matter volume decline with age and modifying factors. Neurobiol Aging. Talairach, J., Tournoux, P., 1988. Co-planar Stereotaxic Atlas of the Human Brain: 3dimensional approach system: an approach to cerebral imaging. Georg Thieme Verlag, Stuttgart. Terry, R.D., DeTeresa, R., Hansen, L.A., 1987. Neocortical cell counts in normal human adult aging. Ann. Neurol. 21, 530–539. Ward, M.A., Carlsson, C.M., Trivedi, M.A., Sager, M.A., Johnson, S.C., 2005. The effect of body mass index on global brain volume in middle-aged adults: a cross sectional study. BMC Neurol. 5, 23. Zimmerman, M.E., Brickman, A.M., Paul, R.H., Grieve, S.M., Tate, D.F., Gunstad, J., Cohen, R.A., Aloia, M.S., Williams, L.M., Clark, C.R., Whitford, T.J., Gordon, E., 2006. The relationship between frontal gray matter volume and cognition varies across the healthy adult lifespan. Am. J. Geriatr. Psychiatry 14, 823–833.