Reproducibility of repeated measures of deuterium substituted [11C]L-deprenyl ([11C]L-deprenyl-D2) binding in the human brain

Nuclear Medicine & Biology, Vol. 27, pp. 43– 49, 2000 Copyright © 2000 Elsevier Science Inc. All rights reserved.

ISSN 0969-8051/00/$–see front matter PII S0969-8051(00)00088-8

Reproducibility of Repeated Measures of Deuterium Substituted [11C]L-Deprenyl ([11C]L-deprenyl-D2) Binding in the Human Brain Jean Logan,1 Joanna S. Fowler,1 Nora D. Volkow,2 Gene-Jack Wang,2 Robert R. MacGregor1 and Colleen Shea1 1

CHEMISTRY AND 2MEDICAL DEPARTMENTS, BROOKHAVEN NATIONAL LABORATORY, UPTON, NEW YORK, USA

ABSTRACT .The purpose of this study was to assess the reproducibility of repeated positron emission tomography (PET) measures of brain monoamine oxidase B (MAO B) using deuterium-substituted [11C]L-deprenyl ([11C]L-deprenyl-D2) in normal subjects and to validate the method used for estimating the kinetic constants from the irreversible 3-compartment model applied to the tracer binding. Five normal healthy subjects (age range 23–73 years) each received two PET scans with [11C]L-deprenyl-D2. The time interval between scans was 7–27 days. Time-activity data from eight regions of interest and an arterial plasma input function was used to calculate ␭k3, a model term proportional to MAO B, and K1, the plasma to brain transfer constant that is related to blood flow. Linear (LIN) and nonlinear least-squares (NLLSQ) estimation methods were used to calculate the optimum model constants. A comparison of time-activity curves for scan 1 and scan 2 showed that the percent of change for peak uptake varied from ⴚ18.5 to 15.0% and that increases and decreases in uptake on scan 2 were associated with increases and decreases in the value of the arterial input of the tracer. Calculation of ␭k3 showed a difference between scan 1 and scan 2 in the global value ranging between ⴚ6.97 and 4.5% (average ⴚ2.1 ⴞ 4.7%). The average percent change for eight brain regions for the five subjects was ⴚ2.84 ⴞ 7.07%. Values of ␭k3 for scan 1 and scan 2 were highly correlated (r2 ⴝ 0.98; p < 0.0001; slope 0.955). Similarly, values of K1 showed a significant correlation between scan 1 and scan 2 (r2 ⴝ 0.61; p < 0.0001; slope 0.638) though the values for scan 2 were generally lower than those of scan 1. There was essentially no difference between the values of model constants calculated using the NLLSQ or LIN methods. Regional brain uptake of [11C]Ldeprenyl-D2 varied between scan 1 and scan 2, driven by the differences in arterial tracer input. Application of a 3-compartment model to regional time-activity data and arterial input function yielded ␭k3 values for scan 1 and scan 2 with an average difference of ⴚ2.84 ⴞ 7.07%. Linear regression applied to values of ␭k3 from the LIN and NLLSQ methods validated the use of the linear method for calculating ␭k3. NUCL MED BIOL 27;1:43– 49, 2000. © 2000 Elsevier Science Inc. All rights reserved. KEY WORDS. PET, Monoamine oxidase, Kinetic model, [11C]L-deprenyl-D2

INTRODUCTION Monoamine oxidase (MAO) is an enzyme that breaks down amines from endogenous and exogenous sources (14). It occurs in two subtypes, MAO A, which oxidizes norepinephrine and serotonin, and MAO B, which oxidizes benzylamine and phenethylamine. MAO inhibitor drugs are used to treat depression and Parkinson’s disease and investigations have linked brain and platelet MAO B levels to neurologic and psychiatric disorders (15). We have developed deuterium-substituted [11C]L-deprenyl ([11C]L-deprenyl-D2) for measuring MAO B in the human brain (7). In the rate-limiting (or rate-contributing) step of the enzymatic catalysis, MAO B cleaves a C–H bond in the methylene group of the propargyl group of [11C]Ldeprenyl. This generates a reactive intermediate that irreversibly binds to and labels the enzyme. Substitution of the hydrogen atoms, which are cleaved by the enzyme with deuterium atoms, reduces the rate of cleavage because a C–D bond is more difficult to break than a C–H bond. The reduced rate of trapping of [11C]L-deprenyl-D2 Address correspondence to: Jean Logan, Ph.D., Chemistry Department, Brookhaven National Laboratory, Upton, NY 11973, USA; e-mail: [email protected]. Received 27 September 1999. Accepted 21 October 1999.

relative to [11C]L-deprenyl improves the sensitivity of the tracer to changes in MAO B (7, 9). Many of the medical issues related to brain MAO B center on whether drugs or other factors change the concentration of the enzyme. Such issues are most effectively addressed using a within-subject design wherein a subject serves as his/her own control and the effect of the drug or other intervention is assessed by observing a change in radiotracer binding relative to the baseline condition. The purpose of the present study was to assess the reproducibility of repeated measures of [11C]L-deprenyl-D2 binding in the human brain with no intervention and to assess the reliability of model parameters estimated using a linear least-squares algorithm based on Blomqvist’s linearized equations of the (irreversible) 3-compartment model (2). The standard for comparison was the classical nonlinear least squares method for estimating the kinetic constants, which requires considerably more computation time. The model terms compared were ␭k3 and K1, which are functions of MAO B activity and brain blood flow, respectively (8). MATERIALS AND METHODS

Subjects Five healthy normal subjects (age range 23–73 years; 4 females, 1 male) were recruited for the study. Exclusion criteria in-

44

J. Logan et al.

FIG. 1. Time-activity curves for scan 1 (circles) and scan 2 (squares) for the global region for the five subjects.

cluded history of neurologic or psychiatric disorder, head trauma with loss of consciousness, alcohol or substance abuse, and medical conditions that may alter cerebral functioning. Prescan tests ensured the absence of any psychoactive drug use.

The studies followed the guidelines of the Institutional Review Board at Brookhaven National Laboratory and subjects gave informed consent after the procedure had been fully explained.

Reproducibility of [11C]L-Deprenyl-D2 Binding

45

TABLE 1. Values for Injected Dose of [11C]L-deprenyl-D2 and the Plasma Integral at 60 Minutes for Total Carbon-11 and of [11C]L-deprenyl-D2 Corrected for the Presence of Labeled Metabolites for Scan 1 and Scan 2

Subject

Scan

␮g Injected

Plasma integral (Carbon-11)

Plasma integral ([11C]L-dep-D2)

1 2

5.2 14.4

1 2

15.6 2.8

1 2

12.0 5.3

1 2

11.1 31.4

1 2

3.0 3.4

0.046 0.054 1.17 0.041 0.035 0.84 0.057 0.064 1.13 0.0341 0.041 1.20 0.0584 0.0534 0.92

0.025 0.0296 1.18 0.031 0.026 0.83 0.035 0.039 1.13 0.020 0.025 1.29 0.042 0.037 0.89

1 Ratio (2:1) 2 Ratio (2:1) 3 Ratio (2:1) 4 Ratio (2:1) 5 Ratio (2:1) Units are %/injected dose/cc min.

Positron Emission Tomography Methods Positron emission tomography (PET) scans were performed on a whole-body, high-resolution positron emission tomograph (CTI 931; 6.5 ⫻ 6.5 ⫻ 6.5 mm, full width half maximum, 15 slices; Computer Technologies Inc, Knoxville, TN USA). To ensure accurate repositioning of subjects in the PET scanner for the repeated scans, an individually molded headholder was made for each subject. The head of the subject was then positioned in the gantry with the aid of three orthogonal laser lines, two of which were placed parallel to the left and right canthomeatal lines and one parallel to the sagittal plane. A transmission scan was obtained with a 68Ge ring source before the emission scan to correct for attenuation. Catheters were placed in an antecubital vein for radiotracer injection and in the radial artery for blood sampling. Each subject had two PET scans with [11C]L-deprenyl-D2 (5–9 mCi, injected dose ranged between 2.8 ␮g and 31.4 ␮g) (7) performed within 1 month of each other (range 7–22 days; average 18 ⫾ 8 days). Sequential PET scans were obtained immediately after injection for a total of 60 min with the following timing: 10 ⫻ 1 min frames; 4 ⫻ 5 min frames; 3 ⫻ 10 min frames. Arterial blood samples were taken over the course of the experiment and plasma was counted for total carbon-11 concentration. The percent of [11C]L-deprenyl-D2 in six plasma samples (1, 5, 10, 30, 45, and 60 min) was determined using

a solid phase extraction method performed by a programmed laboratory robot as determined previously (1). This analysis was performed for scans 1 and 2 and the values for each of the time points were averaged for each subject. The total carbon-11 concentration in plasma for each subject for each scan was corrected for the presence of labeled metabolites using each subject’s averaged values for the percent of [11C]L-deprenyl-D2.

Data Analysis Regions of interest (ROI) were drawn directly on the PET scans (summed image from 30 – 60 min) as described previously (7). The following regions were analyzed: frontal, temporal, parietal and occipital cortices, thalamus, basal ganglia, and cerebellum. A global region was obtained by summing six central planes. For structures occurring bilaterally, right and left regions were summed. ROIs from the baseline study were projected onto the second study and manually adjusted to account for positioning errors. ROIs obtained in this way were projected onto the dynamic scans to obtain time-activity curves. The reproducibility of [11C]L-deprenyl-D2 binding for scan 1 and scan 2 was assessed using the following measures:

TABLE 2. Means and Coefficients of Variation (COV) for K1 for Scan 1 and Scan 2 Along with the Average Percent Change and the Range of Percent Changes

ROI Global Basal ganglia Thalamus Cerebellum Frontal cortex Occipital cortex Parietal cortex Temporal cortex

K1 scan 1

COV %

K1 scan 2

COV %

% Change (average)

% Change (range)

0.511 0.678 0.845 0.654 0.664 0.727 0.634 0.670

18.6 16.0 15.2 14.9 23.5 33.2 22.4 22.8

0.476 0.675 0.772 0.567 0.608 0.658 0.547 0.600

17.1 12.9 12.9 18.2 21.1 14.6 30.8 24.4

⫺6.1 0.07 ⫺7.3 ⫺12.5 ⫺7.5 ⫺3.64 ⫺14.4 ⫺10.4

⫺20.6 to 3.0 ⫺4.6 to 7.6 ⫺11.6 to 6.6 ⫺36 to 0.82 ⫺26 to ⫺0.82 ⫺39.6 to 16.8 ⫺24.7 to 9.6 ⫺24.4 to 3.7

% change ⫽ 100 (scan 2 ⫺ scan 1)/scan 1.

J. Logan et al.

46

TABLE 3. Means and Coefficients of Variation (COV) for ␭k3 for Scan 1 and Scan 2 Along with the Average Percent Change and the Range of Percent Changes

ROI Global Basal ganglia Thalamus Cerebellum Frontal cortex Occipital cortex Parietal cortex Temporal cortex

␭k3 scan 1

COV %

␭k3 scan 2

COV %

% Change (average)

0.171 0.398 0.331 0.161 0.193 0.172 0.176 0.190

23.1 28.3 25.6 24.6 23.4 20.3 29.4 27.9

0.167 0.381 0.326 0.159 0.187 0.165 0.167 0.185

24.3 27.2 25.5 30.4 22.6 20.0 27.2 27.0

2.1 ⫺4.1 ⫺1.3 ⫺1.4 ⫺3.2 ⫺3.7 ⫺4.5 ⫺2.5

% Change (range) ⫺6.7 to 4.5 ⫺12 to 1.31 ⫺6.5 to 5.6 ⫺19 to 14 ⫺5.6 to 2.7 ⫺19 to 13.4 ⫺10.5 to 2.6 ⫺6.7 to 6.3

% change ⫽ 100 (scan 2 ⫺ scan 1)/scan 1.

1. Brain region time-activity curves expressed as percent injected dose per cc. Percent change is calculated by comparing the peak value in the time-activity curve from scan 1 to that in scan 2. This was done for the global value for each of the subjects. 2. Arterial plasma input function expressed as the value of the integral between 0 and 60 min for total carbon-11 activity and for [11C]L-deprenyl-D2. 3. Values of ␭k3, which is a function of MAO B activity, and K1, which is related to brain blood flow from an irreversible 3-compartment model.

Calculation of Model Terms LINEAR METHOD. The linearized 3-compartment model equation (LIN) (8), which relates PET time-activity data from different brain regions (ROI), and time-activity data in arterial plasma (Cp) can be expressed in terms of the influx constant Ki (12) and kinetic constants K1, k2, and k3, giving

ROI共T兲 ⫽ K 1

冕

T

冉 冕冕 T

t

Cp共t兲dt ⫹ 共k 2 ⫹ k 3兲 Ki

0

0

⫺

Cp共t⬘兲dt⬘ dt

0

冕

T

0

ROI共t兲dt

冊

more robust parameter than k3 (8), where ␭ is defined as K1/k2, which is independent of blood flow (11). ␭k3 can be calculated from equation 3 if Ki and K1 are known. Ki was calculated using the average of slopes between 6 and 45 min and 7 and 55 min. K1 was then calculated from equation 2, which is a bilinear regression with two parameters, K1 and k2 ⫹ k3, which can be computed using standard methods. Equation 2 was solved for several values of K1 by successively increasing the maximum time T from 5 to 18 min, because K1 is more sensitive to data at these earlier time points. An average K1 was calculated from these values. NLLSQ METHOD. Optimal estimates of the model parameters K1, k2, and k3 were determined iteratively using the downhill simplex method (13). The differential equations for this system and the method of solution have been described previously (9).

Statistical Analysis Values for ␭k3 for scan 1 and scan 2 and for K1 for scan 1 and scan 2 for each brain region (global, basal ganglia, thalamus, cerebellum, frontal cortex, occipital cortex, parietal cortex, and temporal cortex) were compared for each individual subject using a paired two-tailed Student’s t-test. A correlation analysis was performed

(1)

written as

ROI共T兲 ⫽ K 1

冕

T

Cp共t兲 ⫹ 共k 2 ⫹ k 3兲 Z共T兲

(2)

0

where Z(T), the expression within parentheses in equation 1, can be calculated separately once Ki is determined using the method of Patlak et al. (12). T is the upper integration time limit and t is the variable time of integration. K1 and k2 describe uptake (plasma to tissue) and efflux (tissue to plasma), respectively. The trapping of [11C]L-deprenyl-D2 by MAO B is described by k3. ␭k3 is a function of MAO B concentration and is related to the influx constant Ki (11) as follows: Ki ⫽ K 1k 3/共k 2 ⫹ k 3兲 ⫽ K 1␭k 3/共K 1 ⫹ ␭k 3兲

(3)

Because Ki is a function of blood flow through K1 and k2 (11), it is preferable to express MAO B activity in terms of ␭k3, which is a

FIG. 2. Plot of individual values of percent change between scan 1 and scan 2 for each brain region for the five subjects. gl ⴝ global; bg ⴝ basal ganglia; th ⴝ thalamus; cb ⴝ cerebellum; fr ⴝ frontal cortex; occ ⴝ occipital cortex; par ⴝ parietal cortex; tmp ⴝ temporal cortex.

Reproducibility of [11C]L-Deprenyl-D2 Binding

47

FIG. 3. Correlation plot between individual values of (A) ␥k3 and (B) K1 for scan 1 and scan 2 for all brain regions for all subjects. between the values of ␭k3 for all brain regions for scan 1 and scan 2, and for K1 for all brain regions for scan 1 and scan 2. Values of ␭k3 and K1 for both the LIN and NLLSQ methods were compared by linear regression. RESULTS

Assessment of Reproducibility Time-activity curves were generated for different brain regions and for the global value. For the global regions the percent injected dose

per cc varied from 0.004 %ID/cc to 0.010 %ID/cc. The percent change from scan 1 to scan 2 ranged from ⫺18.5% to 15.0%. Figure 1 shows time-activity curves for the five subjects for the global region. The values of the integrated arterial plasma input function showed changes ranging from ⫺26% to 26.9% between scan 1 and scan 2 (Table 1). The injected mass of [11C]L-deprenyl-D2 (Table 1) was generally not equal for the two scans because of variations in tracer specific activity. However, there was not a correlation between injected mass and the value of the arterial input function (p ⬎ 0.9). The model terms K1 and ␭k3 showed no significant

FIG. 4. Correlation plots of (A) ␥k3 and (B) K1 for the nonlinear (NLLSQ; x axis) and linear (LIN; y axis) methods.

J. Logan et al.

48

differences for any brain region between scan 1 and scan 2 (paired t-test, two-tailed). Tables 2 and 3 give the averaged values of K1 and ␭k3 for five brain regions for each subject along with the coefficient of variation and the average percent change between scans. Individual percent changes in ␭k3 for all brain regions examined is presented in Figure 2 for the five subjects. The average percent change for all five subjects for eight brain regions for ␭k3 was ⫺2.84 ⫾ 7.07. A correlation plot for all brain regions for all individuals for ␭k3 showed a highly significant relationship between scan 1 and scan 2 (r2 ⫽ 0.96; F ⫽ 1164; p ⬍ 0.0001; slope 0.955; Fig. 3A). A significant correlation was also obtained for K1 for scan 1 and scan 2 (r2 ⫽ 0.61; F ⫽ 81; p ⬍ 0.0001; slope 0.638; Fig. 3B). The average percent change for K1 for all brain regions for all five subjects was ⫺7.71 ⫾ 13.2%. From this and from inspection of Figure 3B, it is apparent that the value of K1 was generally lower on the second scan.

Validation of Model Calculations Correlation plots of LIN and NLLSQ for K1 and ␭k3 are shown in Figure 4. From Figure 4A, K1 (NLLSQ; x axis) and K1 (LIN; y axis), the slope was 1.009 ⫾ 0.017 with intercept 0.0022 ⫾ 0.011 (r ⫽ 0.99). For ␭k3 (Fig. 4B), the slope was 0.995 ⫾ 0.006 with intercept ⫺0.0118 ⫾ 0.117 (r ⫽ 0.998) indicating only a small bias. For Ki (not shown) the slope was 1.029 ⫾ 0.008 with intercept ⫺0.0117 ⫾ 0.0015 (r ⫽ 0.997). The values from both methods are in very good agreement. Figure 5 is a plot of k3 versus ␭k3, both calculated by the NLLSQ method. The relationship does not appear to be linear; furthermore, it was generally found that the smaller values of ␭ were associated with larger values of k3. The average value of ␭ over all regions was 4.10 ⫾ 0.90. This confirms our earlier results that ␭k3 is more reliably estimated than k3 (8). DISCUSSION MAO B levels in the human brain are highly variable. Although we do not know all of the factors that contribute to this variability, we do know that MAO B increases with increasing age (5) and that tobacco smoke exposure reduces MAO B in the human brain (6). We previously measured the test/retest reproducibility of repeated measures of the first generation MAO B tracer [11C]L-deprenyl binding in the human brain as part of a study to measure the efficacy of different doses of the MAO B inhibitor drug lazabemide to inhibit MAO B (8). The test/retest reproducibility for the model term ␭k3 for [11C]L-deprenyl averaged ⫺0.8 ⫾ 6.3% (range ⫺17 to 15%) for studies repeated 1 week apart with no intervention. In the present study we determined the reproducibility of repeated measure of [11C]L-deprenyl-D2, a second generation MAO B tracer with improved sensitivity resulting from its reduced rate of binding to MAO B induced by the deuterium isotope effect. We observed considerable differences in brain uptake between scans performed 1– 4 weeks apart in the same individual. Both increases and decreases in uptake are observed. These are driven by differences in the arterial plasma input function. An increased uptake in scan 2 is accompanied by an increased plasma input function and vice versa. Basing the comparison on the model term ␭k3 (a function of MAO B), which is derived from a 3-compartment model that incorporates the plasma input function, the changes between scan 1 and scan 2 decrease. The average percent change between scan 1 and scan 2 is ⫺2.84 ⫾ 7.07% (Fig. 2) for all regions for the five subjects and the values of scan 1 and scan 2 are highly correlated with a regression slope of 0.955 (Fig. 3A).

FIG. 5. Correlation plot of ␥k3 versus K1, both of which were calculated using the nonlinear (NLLSQ) method. We also calculated K1, the plasma to brain transfer constant, which is a function of brain blood flow. We previously reported that the K1 calculated for time-activity data from [11C]L-deprenyl-D2 shows the expected decreases with age in healthy normal subjects, indicating that it is sensitive to changes in blood flow (5). Figure 3B shows that most of the values for K1 for the second scan are lower than the values from the first scan. This is also reflected in the average percent change for K1, which is ⫺7.7 ⫾ 13.2%. Lower values of K1 on the second scan (which may be related to lower blood flow) may reflect familiarization with the PET procedure and decreases in anxiety for some of the subjects. We previously found that regional metabolism is consistently reduced in the second PET scan regardless of condition (16). In addition, prior studies have reported that cortical blood flow is affected by anxiety (10). Because [11C]L-deprenyl-D2 is an irreversibly bound tracer, its binding is sensitive to radiotracer delivery to the brain. The irreversible 3-compartment model allows the calculation of separate model terms ␭k3 and K1. The 3-compartment model was implemented in two different ways. The LIN method based on Ki is computationally rapid but can produce biased estimates due to correlated noise (4). However, for [11C]L-deprenyl-D2 using ROI analysis, the LIN method performs as well as the NLLSQ method. There is less variability in ␭k3 than in k3 for which average percent changes between scan 1 and scan 2 were ⫺11 and 7 compared with ⫺4 and ⫺3 in thalamus and basal ganglia, respectively. Similar difficulties in estimating the individual model parameters k2 and k3 have been reported with other tracers (3). Generally, combinations of model parameters, for example, the distribution volume in the case of reversible ligands (3), and in this case, ␭k3, are more reliably estimated. CONCLUSION The regional brain uptake of [11C]L-deprenyl-D2 varied between scan 1 and scan 2, driven by the differences in arterial tracer input. The application of a 3-compartment model to regional time-activity data and the arterial input function yielded reasonably reproducible

Reproducibility of [11C]L-Deprenyl-D2 Binding

values for ␭k3 for scan 1 and scan 2, with a difference of ⫺2.84 ⫾ 7.07% averaged over seven brain regions and the global region. This supports the use of the within-subject experimental design and the 3-compartment model using the model term ␭k3 and the LIN method for evaluating the effect of drugs and other interventions on MAO B. This research was carried out at Brookhaven National Laboratory under contract DE-AC02-98CH10886 and with the U.S. Department of Energy and supported by its Office of Biological and Environmental Research and by the National Institutes of Health grant NS 15380. The authors are grateful to Richard Ferrieri, Robert Carciello, Donald Warner, David Alexoff, David Schlyer, Noelwah Netusil, and Payton King for advice and assistance, and to Carol Redvanly and Lois Caligiuri for organization and administration. They are also grateful to the people who volunteered for this study.

49

7.

8.

9.

10.

References 1. Alexoff D. L., Shea C., Fowler J. S., King P., Gatley S. J., Schlyer D. J. and Wolf A. P. (1995) Plasma input function determination for PET using a commercial laboratory robot. Nucl. Med. Biol. 22, 893–904. 2. Blomqvist G. (1984) On the construction of functional maps in positron emission tomography. J. Cereb. Blood Flow Metab. 4, 629 – 632. 3. Carson R. E., Channing M. A., Blasberg R. G., Dunn B. B., Cohen R. M., Rice K. C. and Herscovitch P. (1993) Comparison of bolus and infusion methods for receptor quantitation: Application to [18F]cyclofoxy and positron emission tomography. J. Cereb. Blood Flow Metab. 13, 24 – 42. 4. Feng D., Ho D., Chen K., Wu L.-C., Wang J.-K., Liu R.-S. and Yeh S.-H. (1995) An evaluation of the algorithms for determining local cerebral metabolic rates of glucose using positron emission tomography dynamic data. IEEE Trans. Med. Imaging 14, 697–710. 5. Fowler J. S., Volkow N. D., Wang G.-J., Logan J., Pappas N., Shea C. and MacGregor R. (1997) Age-related increases in brain monoamine oxidase B in living healthy human subjects. Neurobiol. Aging 18, 431– 435. 6. Fowler J. S., Volkow N. D., Wang G.-J., Pappas N., Logan J.,

11.

12.

13.

14. 15. 16.

MacGregor R. R., Alexoff D., Shea C., Wolf A. P., Warner D., Zezulkova I. and Cilento R. (1996) Inhibition of monoamine oxidase in the brains of smokers. Nature 379, 733–736. Fowler J. S. , Wang G.-J., Logan J., Xie S., Volkow N. D., MacGregor R. R., Schlyer D. J., Pappas N., Alexoff D. L., Patlak C. and Wolf, A. P. (1995) Selective reduction of radiotracer trapping by deuterium substitution: Comparison of carbon-11 L-deprenyl and carbon-11 deprenyl-D2 for MAO B mapping. J. Nucl. Med. 36, 1255–1262. Fowler J. S., Volkow N. D., Logan J., Schlyer D. J., MacGregor R. R., Wang G.-J., Wolf A. P., Pappas N., Alexoff D., Shea C., Dorflinger E., Yoo K., Morawsky L. and Fazzini, E. (1993) Monoamine oxidase B (MAO B) inhibitor therapy in Parkinson’s disease: The degree and reversibility of human brain MAO B inhibition by Ro 19 6327. Neurology 43, 1984 –1992. Fowler J. S., Wolf A. P., MacGregor R. R., Dewey S. L., Logan J., Schlyer D. J. and Langstrom B. (1988) Mechanistic positron emission tomography studies: Demonstration of a deuterium isotope effect in the monoamine oxidase catalyzed binding of [11C]L-deprenyl in living baboon brain. J. Neurochem. 51, 1524 –1534. Gur R. C., Gur R. E., Resnick S. M., Skolnick B. E., Alavi A. and Reivich M. (1987) The effect of anxiety on cortical cerebral blood flow and metabolism. J. Cereb. Blood Flow Metab. 7, 173–177. Logan J., Dewey S. L., Wolf A. P., Fowler J. S., Brodie J. D., Angrist B., Volkow N. D. and Gatley S. J. (1991) Effects of endogenous dopamine on measures of [18F]N-methylspiroperidol binding in the basal ganglia: Comparison of simulations and experimental results from PET studies in baboons. Synapse 9, 195–207. Patlak C., Fenstermacher J. D. and Blasberg R. (1983) Graphical evaluation of blood-to-brain transfer constants from multiple timeactivity data. J. Cereb. Blood Flow Metab. 3, 1–7. Press W. H., Flannery B. P., Teukolsky S. A. and Vettering W. T. (1988) Numerical Recipes in C. The Art of Scientific Computing. Cambridge University Press, Cambridge, England. Singer T. P. and Ramsay R. A. (1995) Monoamine oxidases: Old friends hold many surprises. FASEB J. 9, 605– 610. Strolin-Benedetti M. and Dostert P. (1989) Monoamine oxidase, brain aging and degenerative diseases. Biochem. Pharmacol. 38, 555–561. Wang G.-J., Volkow N. D., Overall J., Hitzemann R., Pappas N., Pascani K. and Fowler J. S. (1996) Reproducibility of regional brain metabolic responses to lorazepam. J. Nucl. Med. 37, 1609 –1613.