Diffusion tensor metrics, T2 relaxation, and ... - Wiley Online Library

6 downloads 15173 Views 740KB Size Report
Nov 29, 2007 - tions of these discrepant results call for cautious interpre- tation of data as ... Health Science Center at Houston, Houston Texas. 2Center for ...
Magnetic Resonance in Medicine 59:7–13 (2008)

Diffusion Tensor Metrics, T2 Relaxation, and Volumetry of the Naturally Aging Human Caudate Nuclei in Healthy Young and Middle-Aged Adults: Possible Implications for the Neurobiology of Human Brain Aging and Disease Khader M. Hasan,1* Christopher Halphen,1 Michael D. Boska,2,3 and Ponnada A. Narayana1 ume depends on a number of factors that include the neuropil (neurons and glia), the extracellular space, dendrite proliferation, and connections (1,3,5). With natural aging the CNV has been reported to decrease in both males and females bilaterally in cats (6), monkeys (7), and humans (4,8 –11). The atrophy (or loss of tissue volume and density) of the CNV has been used as a marker in Alzheimer’s disease (12), Huntington’s disease (13), and other human neurological disorders (2). The caudate volume has also been reported to be affected in a host of genetic and neurodevelopmental conditions. It is also noteworthy to mention that the effectiveness of some therapeutic procedures have been marked by the CN (14). The exact neuronal mechanisms responsible for caudate volume changes have yet to be explored. There are many conflicting reports in the current noninvasive neuroimaging literature about the effect of age, sex, and lateralization on the MRI measures of the CN in both healthy children and adults. For example, using a region-of-interest (ROI) approach Agartz et al. (15) did not observe age-related T2 changes in either young and old males and females while Wansapura et al. (16) reported that the T2 in GM is significantly shorter in adult females compared to healthy age-matched males and attributed it to sex-differences in neuronal density. Bartzokis et al. (17) reported a longer T2 in females compared to males and attributed it to lower concentration of non-heme iron in females. A few diffusion tensor imaging (DTI) studies using ROI and voxel-based morphometry (VBM) analysis have provided somewhat conflicting results about the relationship between caudate diffusion anisotropy and age in normal subjects (18 –23). For example, using ROI methods, the study by Mukherjee et al. (21) predicted a small but nonsignificant increase in the fractional anisotropy (FA) of the caudate during childhood (age 4 –12 years), while Snook et al. (22) reported a steep increase in FA that extends into young adulthood. Note also that VBM-based methods may not reliably yield accurate volume and diffusion measurements of the caudate because of the lateralventricular cerebrospinal fluid (CSF) contamination and bordering white matter (WM) of the internal capsule due to excessive smoothing for spatial normalization to a standardized template (23). The measurements and implications of these discrepant results call for cautious interpretation of data as MRI intrinsic parameters are sensitive to

In this study of a cohort of 33 young and middle-age adults (19 –59 years) we report simultaneous measurements of normal age-related changes in the caudate nuclei volume, diffusion tensor metrics, and T2 relaxation time. Both the absolute caudate volume and its ratio relative to the total intracranial volume decreased rapidly with age in both men and women (r ⴝ ⴚ0.55; P < 0.001). The fractional diffusion tensor anisotropy of the caudate nuclei increased with age in both males and females (r ⴝ 0.48; P ⴝ 0.005). The corresponding age correlations of the caudate axial (r ⴝ 0.17; P ⴝ 0.35), transverse (r ⴝ ⴚ0.12; P ⴝ 0.50), mean diffusivities (r ⴝ 0.018; P ⴝ 0.92), and T2 relaxation times (r ⴝ 0.194; P ⴝ 0.28) were weaker and did not reach statistical significance (P > 0.05). Our preliminary findings warrant further studies on the older and aging adults and indicate that caudate diffusion tensor imaging-derived metrics can be used as surrogates in modeling the neuronal substrates of gray matter atrophy. Magn Reson Med 59:7–13, 2008. © 2007 WileyLiss, Inc. Key words: caudate nuclei; diffusion tensor imaging; MRI T2relaxation; normal or natural aging; young and middle-age adults

The human caudate nuclei (CN) are well-perfused gray matter (GM) structures that are part of a system that provides feedback to the prefrontal and niagrostriatal circuits and are thought to be involved in cognition and fine motor control (1,2). The human caudate nuclei volume (CNV) is estimated to be 6 –12 mL and represents about 0.5–1% of the total intracranial volume (ICV) (3,4). The caudate vol-

This article includes Supplementary Material available via the Internet at http://www.interscience.wiley.com/jpages/xxxx-xxxx/suppmat. 1Department of Diagnostic and Interventional Imaging, University of Texas Health Science Center at Houston, Houston Texas. 2Center for Neurovirology and Neurodegenerative Disorders, University of Nebraska Medical Center, Omaha, Nebraska. 3Department of Radiology, University of Nebraska Medical Center, Omaha, Nebraska. Grant sponsor: National Institute of Neurological Disorders and Stroke (NIHNINDS); Grant number: R01-NS052505-02 (to K.M.H., Diffusion Tensor Imaging of Wallerian Degeneration in Multiple Sclerosis); Grant sponsor: NIH; Grant number: S10 RR19186-01 (partial funding of the 3.0T MRI clinical scanner, to P.A.N.). *Correspondence to: Khader M. Hasan, PhD, Fannin St. MSB 2.100, Houston TX 77030. E-mail: [email protected] Received 2 May 2007; revised 6 August 2007; accepted 9 September 2007. DOI 10.1002/mrm.21434 Published online 29 November 2007 in Wiley InterScience (www.interscience. wiley.com). © 2007 Wiley-Liss, Inc.

7

8

acquisition protocols and demand high signal-to-noise (SNR) and spatial resolution. DTI studies in GM are expected to be problematic due to partial volume averaging with neighboring CSF and myelinated WM pathways (19,23) along with the fact that diffusion anisotropy in GM, which is low, tends to be overestimated at low SNR (21,24 –27). Careful attention to SNR, image quality, image registration, acquisition paradigm, and tissue demarcation are crucial in noninvasive quantitative MRI measurements of gray matter (4,8). Our central hypothesis is that the concomitant measurements of caudate volume along with DTI metrics and T2 may provide useful noninvasive baseline trajectories for understanding the natural mechanisms that lead to caudate neuronal loss. In this study we measured the T2 relaxation time and diffusion tensor metrics along with the caudate nuclei volume on a cohort of 33 healthy young and middle-age men and women to determine the agerelated changes in the MRI-derived metrics. In this preliminary study some of the problems mentioned above were overcome by utilizing double inversion sequences for delineating the caudate borders and careful attention to quality control issues in the DTI measurements (20,24 –28). MATERIALS AND METHODS Study Population This work was approved by the local Institutional Review Board (IRB). The study is also compliant with the Health Insurance Portability and Accountability Act (HIPAA). A total of 33 healthy adult controls (mean age and standard error ␮ ⫾ ␴ ⫽ 36.9 ⫾ 12.5 years) recruited from the local community were included in this study. All subjects were screened for history of trauma, surgery, chronic illness, alcohol and/or drug abuse, neurological illness, and for current pregnancy. The adult controls included did not report any neurological conditions and the MRI data were judged to be normal. The male and female adult groups (healthy men and women) were age-matched (P ⫽ 0.83) with 15 males (age ␮ ⫾ ␴ ⫽ 36.4 ⫾ 13.0 years; range ⫽ 18.7–57.6 years; two left-handed) and 18 right-handed females (age ␮ ⫾ ␴ ⫽ 37.3 ⫾ 12.5 years; range ⫽ 20.3–58.6 years). Written informed consent was obtained from all subjects. MRI Data Acquisition All MRI acquisitions were performed on a 3T Philips Intera scanner with a dual quasar gradient system with a maximum gradient amplitude of 8.0 Gauss/cm and an eight-channel SENSE-compatible head coil (Philips Medical Systems, Best, Netherlands) (28).

Hasan et al.

sensitive reconstruction (fSIR; TE/TI/TR ⫽ 8/400/4300 ms). The slice thickness for both conventional and diffusionweighted volumes was 3.0 mm with 44 contiguous axial slices covering the entire brain and a square field-of-view of 240 ⫻ 240 mm2. Diffusion Tensor Imaging Data Acquisition The DTI data were acquired using a single-shot spin-echo diffusion sensitized and fat-suppressed echo-planar imaging sequence with the balanced Icosa21 tensor encoding scheme (i.e., 21 uniformly distributed directions over the unit hemisphere) (25), b-factor ⫽ 1000 sec mm⫺2, TR/TE ⫽ 7100/65 ms. The echo-planar phase encoding used a SENSE k-space undersampling factor of two (R ⫽ 2) with an effective k-space matrix of 112 ⫻ 112 and an image matrix after zero-filling of 256 ⫻ 256. The number of b⬇0 magnitude image averages was 6; in addition, each diffusion encoding was repeated twice and magnitude-averaged to enhance the SNR. The total DTI acquisition time was ⬍7 min. The Icosa21 encoding scheme (25) used in this study provides multiple icosahedral and rotationally invariant tensor encoding subsets (Icosa6, Icosa15, and Icosa21) that can be utilized for testing the sensitivity of the estimated DTI-metrics to random thermal noise using the same dataset at fixed SNR0 (SNR at b⬇0), and hence providing some important quality control measures for DTI-metric reproducibility (25,26). DTI and Conventional MRI Data Processing Diffusion-weighted images were intra-registered to the baseline “B0” images (without diffusion weighting) to correct for the eddy-current image distortions using a Philips PRIDE workstation (Philips Medical Systems). The diffusion-weighted data were decoded using the Icosa21 and Icosa6 subsets to provide high and low SNR estimates of the DTI metrics (25,26,28). Next, the double inversion recovery and fast spin echo volumes were coaligned with the B0 reference volumes using the mutual informationbased coregistration in SPM (http://www.fil.ion.ucl.ac.uk/ spm/software/spm2/) (28). The T2 values were estimated from the early and late echoes (TE1, TE2) volumes according to standard spin-echo procedures assuming a single comTE1 partment model i.e., Si ⫽ S0 exp ⫺ ; T2 T2

冉 冊冣



冉 冊

TE2 ⫺ TE1 . The T2, FA, mean diffusivity, and eigenS共TE1 兲 ln S共TE2 兲 value maps were saved for further analysis. ⫽

Caudate Nuclei Volume Delineation Conventional MRI The MRI protocol included dual fast spin echo (FSE) with echo (TE) and repetition times of (TR) ⫽ TE1/TE2/TR ⫽ 8.2/90/6800 ms; fluid-attenuated inversion recovery (FLAIR; TE/TI/TR ⫽ 80/2500/80 ms); dual inversion recovery (DIR) sequence (29,30) for suppressing cerebrospinal fluid and white matter (TE/TI1/TI2/TR ⫽ 32/325/3400/ 15,000 ms); and inversion recovery sequence with phase-

Using MRIcro (http://www.sph.sc.edu/comd/rorden/mricro. html), the caudate was manually traced (by a trained operator with 3 years of experience in human brain anatomy and MRI) on 8 –10 consecutive slices on the axial DIR images (that has only GM contribution) to create a mask for volumetry and diffusion analysis using validated procedures as described elsewhere (28). The CN delineation followed anatomical and previously described procedures

Neurobiology of Brain Aging and Disease

9

(8,10,11,14). The most inferior slice was identified where the putamen and caudate were separated by the anterior internal capsule to the most superior slice on which the caudate was visible. Accuracy of the caudate manual tracing was verified by overlaying the FA and mean diffusivity (Dav) maps on the caudate mask (Fig. 1). All datasets were masked to remove nonbrain tissues to estimate the ICV for each subject. Caudate Volume DTI-Metric Estimation and SNR Sensitivity Using Icosa21 and Icosa6 Since the caudate volume was delineated in this study, we were able to obtain the diffusion tensor (DT) metric histograms for each subject at two subsets of the diffusion encoding scheme using the same data (25–27). The metrics associated with each subject histogram (width, mean and peak location, and median) were stored in addition to the data corresponding to all patients to extract pooled histogram statistics. The median of the caudate distribution of the corresponding DTI metrics from each subject was used in the age correlation analysis.

FIG. 2. MRI quantitative histograms of the entire caudate volume of the 15 men and 18 age-matched healthy women: (a) FA, (b) mean diffusivity, (c) principal eigenvalues, and (d) the T2 relaxation time. Notice the lack of separation between the histograms of males and females

Data Quality, Reproducibility, and Stability Our measurements were conducted at high SNR (both B0 and diffusion-weighted) on a stable system as documented by the space versus time analysis on a water phantom over a 2-year span (25). To investigate the sensitivity of DTImetrics on SNR we compared the analysis using the same caudate volume masks generated and utilizing different encoding subsets via paired Student t-tests. The caudate

manual demarcation procedure was repeated by the same rater twice and the results were compared for bias and trends. Statistical Analysis Correlations between age, caudate volume, T2 relaxation time, and DTI-derived metrics were computed using the Pearson correlation coefficient. Slopes and rates of change with age were compared using the r to z-Fisher transform as described elsewhere (31). Comparisons between group means and medians were conducted using analysis of variance (ANOVA) (t-test) and the Mann–Whitney U-test. DTI metrics versus SNR were compared using paired ttests. Rater reproducibility analysis was performed using the intraclass correlation and the Bland–Altman method (31). RESULTS Caudate Nuclei Demarcation, DTI-cMRI Fusion, Reproducibility, and Registration Accuracy An illustration of the method used for demarcating the caudate is shown in Fig. 1, in which a fusion of the coregistered GM, FA, and principal eigenvector maps provided ideal contrast between different tissue types (WM compact fiber tracks, GM, and CSF). The intraclass reproducibility coefficient on 16 randomly selected caudate volumes was ⬇0.95 with minimal bias as judged by the Bland–Altman bias analysis (P ⬎ 0.4).

FIG. 1. Illustration of the caudate nuclei volume (CNV) delineation (one axial section), the conventional MRI/DTI registration, and image fusion to maximize tissue contrast for accurate CNV localization and quantification: (a) FA, (b) dual IR gray matter only, (c) fused GM map with the FA-modulated principal vector map, and (d) a fusion of the GM, FA, and Dav showing the white matter and CSF bordering the caudate.

Caudate Nuclei T2 and DTI Metric Histograms To demonstrate the regional homogeneity of the delineated caudate volumes we plotted the histograms of various DTI metrics and T2 values of all voxels collected from the CN of all subjects included in this study as shown in Fig. 2. The axial diffusivity and T2 distributions indicate

10

Hasan et al.

Table 1 Demographic Data and Summary of Various MRI-Derived Parameters of the Entire Caudate Nuclei Volume on Both Healthy Men and Women Along with the Regression and Correlation Analysis. Men (N ⫽ 15; 2 LH)

Women (N ⫽ 18 RH)

Men and Women (N ⫽ 33)

Men vs. Women (P)

Age (years) ICV (mL) rage (p) CNV (mL) rage(p)

36.4 ⫾ 13.0 1570.7 ⫾ 127.6 0.027 (0.924) 7.57 ⫾ 1.05 ⫺0.575 (0.025)

37.3 ⫾ 12.5 1481.3 ⫾ 108.3 ⫺0.135 (0.594) 7.01 ⫾ 0.72 ⫺0.569 (0.014)

0.832 0.037 0.675 0.082 0.982

CNVp⫽100*CNV/ICV (%) rage (p)

0.482 ⫾ 0.060 ⫺0.647 (0.009)

0.474 ⫾ 0.044 ⫺0.525 (0.025)

FA r(p)

0.143 ⫾ 0.013 0.521 (0.046)

0.136 ⫾ 0.009 0.538 (0.021)

Dav (⫻10⫺3 mm2 sec⫺1) r(p) ␭⬜ (⫻10⫺3 mm2 sec⫺1) r(p) ␭// (⫻10⫺3 mm2 sec⫺1) r(p) T2.(msec) r(p)

0.742 ⫾ 0.043 0.149 (0.596) 0.676 ⫾ 0.039 0.030 (0.915) 0.884 ⫾ 0.057 0.282 (0.308) 94.93 ⫾ 6.70 0.245 (0.379)

0.749 ⫾ 0.029 0.160 (0.525) 0.686 ⫾ 0.030 ⫺0.308 (0.214) 0.888 ⫾ 0.030 0.000 (0.999) 95.52 ⫾ 4.14 0.131 (0.604)

36.9 ⫾ 12.5 1521.9 ⫾ 124.1 ⫺0.064 (0.725) 7.26 ⫾ 0.92 ⫺0.549 (0.00093) CNV (Age) ⫽ (8.75 ⫾ 0.43) ⫺(0.040 ⫾ 0.011)* Age 0.478 ⫾ 0.051 ⫺0.586 (0.00034) CNVp⫽(0.567 ⫾ 0.023) ⫺(0.0024 ⫾ 0.0006)*Age 0.139 ⫾ 0.012 0.482 (0.005) FA(CN) ⫽ (0.122 ⫾ 0.057) ⫹ (0.0045 ⫾ 0.0015)* Age 0.746 ⫾ 0.036 0.018 (0.922) 0.682 ⫾ 0.034 ⫺0.120 (0.504) 0.886 ⫾ 0.044 0.169 (0.347) 95.25 ⫾ 5.37 0.194 (0.278)

Healthy Adults

gender-independence, while the FA distribution shows a slight dependence on gender (Table 1). Caudate Volume, DTI-derived Metrics, and T2 Relaxation Versus Age The main quantitative results on both males and females (group comparisons and age correlations) are summarized in Table 1. The ICV was larger in males compared to females (P ⫽ 0.037) and exhibited no age dependence in either men and women (P ⬎ 0.59). There was no statistically significant difference in the CNV between men and women (P ⫽ 0.082). The CNV to ICV fraction was not significantly different between the age-matched healthy men and women (P ⫽ 0.65). The dependence of CNV (in units of mL ⫽ cm3) on age is described by the best linear regression fit: CNV ⫽ 8.75 – 0.040*Age (r ⫽ ⫺0.55; P ⬍ 0.001) (Fig. 3a). The corresponding age-dependence of CNV-to-ICV percentage (CNV% ⫽ CNV/ICV*100) is CNV% ⫽ 0.567 ⫺ 0.0024*Age (r ⫽ ⫺0.586; P ⫽ 0.00034; see Table 1). A significant increase in FA with age was observed in both men and women at comparable rates (r ⫽ 0.48; P ⫽ 0.005). The corresponding age correlations of the CNV axial diffusivities (r ⫽ 0.17; P ⫽ 0.35), transverse diffusivities (r ⫽ ⫺0.12; P ⫽ 0.50), mean diffusivity (r ⫽ 0.018; P ⫽ 0.92), and T2 relaxation times (r ⫽ 0.194; P ⫽ 0.28) were weaker and did not reach statistical significance (P ⬎ 0.05) (Table 1; Fig. 3). Notice that the sensitivity of caudate FA to aging resulted from the combined effects of a slightly decreasing transverse diffusivity combined with an increase in the axial diffusivity. Signal-to-Noise Sensitivity of FA of the Caudate Nuclei Figure 4a shows that the effect of diffusion-weighted image noise elevates the estimation of FA for the Icosa6

0.653 0.628

0.071 0.953

0.564 0.421 0.400 0.368 0.782 0.454 0.762 0.760

subset (low SNR) compared to the Icosa21 (high SNR), while the mean diffusivity is more immune to noise. Figure 4b shows the age-dependent changes in the median FA estimated from the Icosa6 subset of the parent Icosa21 tensor encoding scheme using the same dataset and the same SNR0. Notice the reduction in the correlation coefficient and the increase in the P value compared to the Icosa21 scheme shown in Fig. 3b. These measured trends are consistent with the theoretical predictions using the perturbation theory and Monte Carlo simulations (24,25).

FIG. 3. Analysis of the age and sex correlations and differences of the distribution median (peak location) of the MRI metrics collected from all subjects caudate nuclei voxels: (a) CNV vs. age, (b) FA vs. age, (c) mean diffusivity vs. age, and (d) T2 relaxation vs. age. Note the strong CNV vs. age and moderate FA vs. age and the poor sensitivity of T2 relaxation and mean diffusivity to age variations.

Neurobiology of Brain Aging and Disease

11

FIG. 4. Illustration of the effect of signal-to-noise ratio (⬇total number of diffusion-weighted images used in DTI metric estimation) on the estimated entire caudate nuclei DTI metrics using the Icosa6 and Icosa21 subsets of the tensor encoding scheme at equal SNR0. a: A paired ANOVA analysis of the estimated eigenvalues, mean diffusivity and FA. Note the overestimation in FA due to noise (Icosa6), which is attributable to a significant overestimation of the principal eigenvalue and the reduction in the transverse eigenvalue. The mean diffusivity is comparable at these two SNR levels obtained using the same dataset. b: A scatterplot of age vs. FA(CNV) shows the importance of SNR in the diffusion measurement in undermining the effect of thermal-random noise using the Icosa6; compare the age trends with the results using the Icosa21 (high SNR) in Fig. 3b.

The significant increase in the entire caudate FA as SNR increases (Fig. 4a) is attributable to an overestimation of the axial and an underestimation of the transverse eigenvalues (26). DISCUSSION This is the first report on simultaneous measurement of the caudate nuclei volume with the corresponding T2 relaxation times and DTI metrics at multiple SNR levels using balanced icosahedral diffusion encoding schemes from a healthy cohort of age-matched young and middle-age men and women. In this work we have focused on the caudate nuclei to provide a simple homogenous representative surrogate model of GM tissue in the CNS (1,4). Our results on the sex-independent loss of CNV and CNV/ICV ratio with age are consistent with several quantitative MRI studies on healthy controls (4,8,10,11) and some early postmortem studies on the caudate (3,9). The caudate volume decrease with age reflects a general trend of steady volume decrease of cortical and deep GM in healthy developing children, young, and older adults (4,10,11,28,32–34). Along with the decrease in caudate volume with age in both men and women (Fig. 3a; Table 1), a sex-independent trend of an increase in the diffusion tensor anisotropy as measured by FA with age was found in the entire adult control population (r ⫽ 0.48; P ⬍ 0.005; n ⫽ 33). The axial and transverse diffusivities and T2 did not exhibit significant agerelated trends in our adult population, hinting at minimal CSF contamination (Fig. 3; Table 1). A major finding of this study is the positive correlation between the caudate FA and age at two levels of SNR. The age versus FA was stronger at the higher SNR level

(Icosa21), but still observable at lower SNR (Icosa6 subset) (Figs. 3b, 4b). The measured trends of caudate FA dependence on SNR are consistent with previous DTI studies (25,27). A comprehensive biophysical interpretation of in vivo DTI measurements in the human brain is not yet available, but is an active area of research (20,24,35). A slight increase in FA of the CN in healthy young and middle-aged men and women could be the result of regressive neuronal and dendrite elimination with age resulting in reduced barriers to diffusion (35). The dendrite connection hypothesis of normal aging is also supported by histology (3,5,6,9,12,37). An analogous working hypothesis has also been invoked to explain the increase in FA in the caudate in Huntington’s disease (13), and more recently in cerebral hypertension (38). In cerebral hypertension (acute hydrocephalus) the caudate diffusion anisotropy has been reported to increase with a decrease or unchanged mean diffusivity, thus ruling out cerebral edema and contamination with cerebrospinal fluid as the sources for the increased anisotropy (38). It is noteworthy to mention that an abnormal “or paradoxical” increase in caudate and putamen diffusion tensor anisotropy along with a reduction in the mean diffusivity has also been reported in the normal-appearing basal ganglia of multiple sclerosis patients by Ciccarelli et al. (39). The authors ruled out gliosis, which would have resulted in more disorganization (e.g., reduced anisotropy and mean diffusivity) and attributed this finding to axonal degeneration due to fiber transection in remote focal multiple sclerosis lesions (39). Our findings may provide some clues to the interpretation of DTI measurements obtained in some human CNS diseases that

12

target the basal ganglia GM neurons and the corresponding WM connections (4,36). The T2 values in a largely unmyelinated and homogeneous region such as the caudate are affected generally by countering effects of increased cellular water that tends to increase T2 and the presence of free radicals, including paramagnetic iron, that would reduce T2 relaxation time (17). These two opposing effects reduce the apparent T2 sensitivity to aging in young and middle-age adults 19 –59 years and may be offset at older ages (not covered in this work ⬎60 years) where iron, for example, may accumulate due to altered metabolism (17). Our current findings of sex-independent nonsignificant age variation of T2 (between 19 –59 years), which are concordant with an older ROI study by Agartz et al. (15), may also be indicative of the poor specificity of T2 during these four decades of adult life. DTI metrics such as FA seem to be more sensitive to normal tissue aging compared to T2. Our study cannot rule out the subtle role of free radical deposits (17) as a direct cause of dendrite elimination (3,6,9,12,17,32,34)—as a working hypothesis to explain the FA increase with normal aging—and may hint toward other mechanisms that need to be explored in the future (5,6,10,34,37). Our preliminary studies may warrant further longitudinal and cross-sectional studies on normal aging in both men and women. The implication and correlation of the caudate volume reduction to specific connected WM fiber pathways (7,40) is beyond the scope of this work and will be pursued in a future study. SUMMARY AND CONCLUSIONS To the best of our knowledge, these are the first MRI studies to report simultaneous measurements of the entire human CNV along with water molecular diffusion tensor imaging metrics and T2 relaxation in a cohort of healthy age-matched young and middle-age men and women. This study demonstrates that both CNV and its volume ratio relative to the ICV decrease rapidly with natural aging in both males and females. The diffusion tensor fractional anisotropy of the caudate increased at a statistically significant and measurable rate (at two levels of SNR), while the T2 relaxation time and mean diffusivity did not exhibit significant age trends. The age-dependent changes in DTI metrics may provide important noninvasive quantitative radiological markers of the neurobiology of aging of the caudate and GM and WM connectivity degeneration in general. In conclusion, our results provide normative baseline data to help interpret the age and sex confounding effects in natural aging and in the interpretation of data collected from neurodevelopment and neurodegenerative conditions. ACKNOWLEDGMENTS We thank Vipul Kumar Patel and Ambika Sankar for helping in data acquisition and management, and Drs. Larry A. Kramer and Patrick Mukherjee for useful discussions. REFERENCES 1. Finch CE, Randall PK, Marshall JF. Aging and basal gangliar functions. Annu Rev Geront Geriatr 1981;2:49 – 87.

Hasan et al. 2. Graybiel AM. The basal ganglia. Curr Biol 2000;10:R509 –R511. 3. Eggers R, Knebel G, Haug H. Morphometric studies of biological changes in synapses of the human caudate nucleus. Z Gerontol 1991; 24:302–305. 4. Jernigan TL, Archibald SL, Berhow MT, Sowell ER, Foster DS, Hesselink JR. Cerebral structure on MRI. Part I. Localization of age-related changes. Biol Psychiatry 1991;29:55– 67. 5. Masliah E, Mallory M, Hansen L, DeTeresa R, Terry RD. Quantitative synaptic alterations in the human neocortex during normal aging. Neurology 1993;43:192–197. 6. Levine MS, Adinolfi AM, Fisher RS, Hull CD, Buchwald NA, McAllister JP. Quantitative morphology of medium-sized caudate spiny neurons in aged cats. Neurobiol Aging 1986;7:277–286. 7. Makris N, Papadimitriou GM, van der Kouwe A, Kennedy DN, Hodge SM, Dale AM, Benner T, Wald LL, Wu O, Tuch DS, Caviness VS, Moore TL, Killiany RJ, Moss MB, Rosene DL. Frontal connections and cognitive changes in normal aging rhesus monkeys: a DTI study. Neurobiol Aging 2007;28:1556 –1567. 8. Ifthikharuddin SF, Shrier DA, Numaguchi Y, Tang X, Ning R, Shibata DK, Kurlan R. MR volumetric analysis of the human basal ganglia: normative data. Acad Radiol 2000;7:627– 634. 9. Orzhekhovskaia NS. Fronto-striatal correlations in normal and pathologic aging of the human brain. Arkh Anat Gistol Embriol 1989;97:6 – 13. 10. Vernaleken I, Weibrich C, Siessmeier T, Buchholz HG, Rosch F, Heinz A, Cumming P, Stoeter P, Bartenstein P, Grunder G. Asymmetry in dopamine D(2/3) receptors of caudate nucleus is lost with age. Neuroimage 2007;34:870 – 878. 11. Walhovd KB, Fjell AM, Reinvang I, Lundervold A, Dale AM, Eilertsen DE, Quinn BT, Salat D, Makris N, Fischl B. Effects of age on volumes of cortex, white matter and subcortical structures. Neurobiol Aging 2005; 26:1261–1270. 12. Zaja-Milatovic S, Keene CD, Montine KS, Leverenz JB, Tsuang D, Montine TJ. Selective dendritic degeneration of medium spiny neurons in dementia with Lewy bodies. Neurology 2006;66:1591–1593. 13. Douaud G, Poupon C, Cointepas Y, Mangin JF, Gaura V, Golestani N, Krystkowiak P, Verny C. Damier P, Bachoud-Le´vi AC, Hantraye P, Remy P. Diffusion tensor imaging (DTI) in Huntington’s disease patients: analyses of fractional anisotropy (FA) maps and apparent diffusion coefficient (ADC) maps. In: Proc ISMRM Workshop on Methods for Quantitative Diffusion MRI of Human Brain, Lake Louise, Canada, 2005. 14. Keshavan MS, Bagwell WW, Haas GL, Sweeney JA, Schooler NR, Pettegrew JW. Changes in caudate volume with neuroleptic treatment. Lancet 1994;344:1434. 15. Agartz I, Saaf J, Wahlund LO, Wetterberg L. T1 and T2 relaxation time estimates in the normal human brain. Radiology 1991;181:537–543. 16. Wansapura JP, Holland SK, Dunn RS, Ball WS Jr. NMR relaxation times in the human brain at 3.0 Tesla. J Magn Reson Imaging 1999;9:531–538. 17. Bartzokis G, Tishler TA, Lu PH, Villablanca P, Altshuler LL, Carter M, Huang D, Edwards N, Mintz J. Brain ferritin iron may influence ageand gender-related risks of neurodegeneration. Neurobiol Aging 2007; 28:414 – 423. 18. Bhagat YA, Beaulieu C. Diffusion anisotropy in subcortical white matter and cortical gray matter: changes with aging and the role of CSFsuppression. J Magn Reson Imaging 2004;20:216 –227. 19. Camara E, Bodammer N, Rodriguez-Fornells A, Tempelmann C. Agerelated water diffusion changes in human brain: a voxel-based approach. Neuroimage 2007;34:1588 –1599. 20. Moseley M. Diffusion tensor imaging and aging — a review. NMR Biomed 2002;15:553–560. 21. Mukherjee P, Miller JH, Shimony JS, Conturo TE, Lee BC, Almli CR, McKinstry RC. Diffusion-tensor MR imaging of gray and white matter development during normal human brain maturation. AJNR Am J Neuroradiol 2002;23:1445–1456. 22. Snook L, Paulson LA, Roy D, Phillips L, Beaulieu C. Diffusion tensor imaging of neurodevelopment in children and young adults. Neuroimage 2005;26:1164 –1173. 23. Snook L, Plewes C, Beaulieu C. Voxel based versus region of interest analysis in diffusion tensor imaging of neurodevelopment. Neuroimage 2007;34:243–252. 24. Conturo TE, McKinstry RC, Aronovitz JA, Neil JJ. Diffusion MRI: precision, accuracy and flow effects [Review]. NMR Biomed 1995;8:307– 332.

Neurobiology of Brain Aging and Disease 25. Hasan KM, Narayana PA. Computation of the fractional anisotropy and mean diffusivity maps without tensor decoding and diagonalization: theoretical analysis and validation. Magn Reson Med 2003;50:589 –598. 26. Hasan KM. A framework for quality control and parameter optimization in diffusion tensor imaging: theoretical analysis and validation. Magn Reson Imaging 2007;25:1196.–1202. 27. Mamata H, Jolesz FA, Maier SE. Characterization of central nervous system structures by magnetic resonance diffusion anisotropy. Neurochem Int 2004;45:553–560. 28. Hasan KM, Halphen C, Sankar A, Eluvathingal TJ, Kramer L, Stuebing KK, Ewing-Cobbs L, Fletcher JM. Diffusion tensor imaging based tissue segmentation: validation and application to the developing child and adolescent brain. Neuroimage 2007;34:1497–1505. 29. Bedell BJ, Narayana PA. Implementation and evaluation of a new pulse sequence for rapid acquisition of double inversion recovery images for simultaneous suppression of white matter and CSF. J Magn Reson Imaging 1998;8:544 –547. 30. Redpath TW, Smith FW. Technical note: use of a double inversion recovery pulse sequence to image selectively grey or white brain matter. Br J Radiol 1994;67:1258 –1263. 31. Glantz SA. Primer of biostatistics, 5th ed. New York: McGraw-Hill; 2002. 32. Burke SN, Barnes CA. Neural plasticity in the ageing brain [Review]. Nat Rev Neurosci 2006;7:30 – 40. 33. Courchesne E, Chisum HJ, Townsend J, Cowles A, Covington J, Egaas B, Harwood M, Hinds S, Press GA. Normal brain development and aging:

13

34. 35. 36.

37.

38.

39.

40.

quantitative analysis at in vivo MR imaging in healthy volunteers. Radiology 2000;216:672– 682. Terry RD, DeTeresa R, Hansen LA. Neocortical cell counts in normal human adult aging. Ann Neurol 1987;21:530 –539. Beaulieu C. The basis of anisotropic water diffusion in the nervous system — a technical review. NMR Biomed 2002;15:435– 455. Boska MD, Hasan KM, Kibuule D, Banerjee R, McIntyre E, Nelson JA, Hahn T, Gendelman HE, Mosley RL. Quantitative diffusion tensor imaging detects dopaminergic neuronal degeneration in a murine model of Parkinson’s disease. Neurobiol Dis 2007;26:590 –596. Baquet ZC, Gorski JA, Jones KR. Early striatal dendrite deficits followed by neuron loss with advanced age in the absence of anterograde cortical brain-derived neurotrophic factor. J Neurosci 2004;24:4250 – 4258. Owler BK, Higgins JN, Pena A, Carpenter TA, Pickard JD. Diffusion tensor imaging of benign intracranial hypertension: absence of cerebral oedema. Br J Neurosurg 2006;20:79 – 81. Ciccarelli O, Werring DJ, Wheeler-Kingshott CA, Barker GJ, Parker GJ, Thompson AJ, Miller DH. Investigation of MS normal-appearing brain using diffusion tensor MRI with clinical correlations. Neurology 2001; 56:926 –933. Lehericy S, Ducros M, Van de Moortele PF, Francois C, Thivard L, Poupon C, Swindale N, Ugurbil K, Kim DS. Diffusion tensor fiber tracking shows distinct corticostriatal circuits in humans. Ann Neurol 2004;55:522–529.

Suggest Documents