cerebral blood volume (CBV), CMROz, and oxygen ex traction in 32 ... and superior temporal regions, whereas only left-sided. CBF values were lower for ...
Tournai o/Cerebral Blood Flow and Metabolism 7:64-67 © 1987 Raven Press, New York
Regional Asymmetries of Cerebral Blood Flow, Blood Volume, and Oxygen Utilization and Extraction in Normal Subjects
Joel S. Perlmutter, William J. Powers, Peter Herscovitch, Peter T. Fox, and Marcus E. Raichle Department of Neurology and Neurological Surgery, Edward Mallinckrodt Institute of Radiology, McDonnell Center for Studies of Higher Brain Function, Washington University School of Medicine, St. Louis, Missouri, U.S.A.
Summary: Positron emission tomography (PET) and \50_ labeled radiotracers were used to measure regional CBF, cerebral blood volume (CBV), CMROz, and oxygen ex traction in 32 right-handed subjects at rest. Mean left hemispheric CBF (46,2 ± 6.8 mlllOO g/min) and CMROz (2,60 ± 0.59 ml/IOO g/min) were significantly lower than right hemispheric values (47.4 ± 7.2 and 2.66 ± 0.61 ml/IOO g/min, respectively; p < 0.0001 for both), whereas left and right hemispheric CBV and oxygen extraction were not significantly different. We further investigated these asymmetries by comparing left- and right-sided values for specific cortical and subcortical regions. We found that left-sided CBF and CMR02 were significantly lower than right-sided values for sensorimotor, occipital,
and superior temporal regions, whereas only left-sided CBF values were lower for anterior cingulum. CBV was asymmetric for the anterior cingulate and midfrontal re gions, and oxygen extraction was asymmetric for the sensorimotor area. No asymmetries were observed in in ferior parietal cortex, thalamus, putamen, or pallidum. Knowledge of these normal physiological asymmetries is essential for proper interpretation of PET studies of phys iology and pathology. Furthermore, the ability to detect asymmetries with PET may lead to a better under standing of the lateralization of specific functions in the human brain. Key Words: Brain lateralization-Positron emission tomography-Regional asymmetries.
Positron emission tomography (PET) permits in vivo measurements of CBF, cerebral blood volume (CBV ) , and metabolism ( P helps et aI. , 1982; Raichle, 1983). These PET measurements have been used to investigate normal physiology (Lam mertsma, 1984; Phelps and Mazziotta, 1985) as well as to demonstrate abnormalities in neurological (Engel et al. , 1982; Rhodes et al. , 1983; Theodore et al., 1983; CaIne et al. , 1985; Powers and Raichle, 1985; Wolfson et al., 1985) and psychiatric (Reiman et al. , 1984) diseases. Past studies have utilized two different methods to define a regional abnormality in a group of subjects: one that uses comparisons with a control population (Perlmutter and Raichle,
1985) and one that assumes that the normal brain is symmetrical (Engel et al. , 1982; Rhodes et al. , 1983; Theodore et al. , 1983). Validity of the latter depends on the accuracy of the assumption of sym metry. The purpose of this study was to investigate this assumption. We recently have demonstrated that left hemi spheric blood flow and oxygen metabolism were significantly less than right hemispheric values in 24 right-handed normal subjects (Perlmutter et al. , 1985). In that study hemispheric values were calcu lated using a mean value from 21 homologous re gions in each hemisphere. The regions were se lected to provide representative samples of gray and white matter while avoiding nonbrain struc tures such as ventricles and major venous sinuses. The regions did not represent specific anatomical structures. We undertook the present study to de termine if there are specific regional asymmetries in normal resting subjects.
Received April 23, 1986; accepted August 5, 1986. Address correspondence and reprint requests to Dr. J. S. Perlmutter at Edward Mallinckrodt Institute of Radiology, 510 South Kingshighway, St. Louis, MO 63110, U.S.A. Ab b reviations used: CBV, cerebral blood volume; PET, posi tron emission tomography.
64
NORMAL ASYMMETRIES OF FLOW AND METABOLISM
METHODS We studied 32 right-handed subjects (1 3 men and 1 9 women) between 2 0 and 8 4 years o f age (mean 37 years) who had no evidence of neurological or psychiatric dis eases. Handedness was determined using the Edinburgh Handedness Inventory (Raczkowski et aI., 1 974). All eight subjects older than 50 years had normal x-ray com puted tomography scans of the brain. PET studies were performed with the PETT VI system in the low-resolution mode (Ter-Pogossian et aI., 1 982; Yamamoto et aI., 1 982). Data were recorded simulta neously for seven slices with a center-to-center separa tion of 1 4.4 mm. Subject preparation included the percutaneous inser tion of a radial arterial catheter under local anesthesia to permit frequent sampling of arterial blood and the inser tion of an intravenous catheter for isotope injection in the opposite arm. Twenty subjects had the arterial catheter on the right side. The head was positioned with a special head holder that utilized an individually molded plastic face mask to prevent movement during the study. A laser permanently attached to the wall projected a line onto the mask that corresponded to the position of the lowest PET slice. A lateral skull radiograph with this line marked by a radiopaque wire provided a record of the subject's exact position in relation to the PET slices. The overlapping position of the radiopaque markers placed in the external auditory canals confirmed that the head was not rotated about the anterior-posterior or vertical axes. After the head was in place, a transmission scan used for individual attenuation correction was performed with a ring source of activity containing germanium-68/gallium-68. During each scan the room was darkened and the subject's eyes were closed. The ears were not occluded. Ambient room noise during the scans was almost entirely from cooling fans for the electronic equipment. Measurements of CBF (Herscovitch et aI., 1 9 83; Raichle et aI., 1 983), CBV (Grubb et aI., 1 978; Mintun et aI., 1 9 84), and CMROz and oxygen extraction (Mintun et aI., 1 984) were made using rapid administration of150-labeled radiotracers. Arterial blood gases were measured during each study. These studies were approved by the Human Studies Committee and the Radioactive Drug Research Com mittee (U.S. Food and Drug Administration) of the Wash ington University School of Medicine. Informed consent was obtained prior to the PET studies. Mean regional hemispheric and global values were cal culated using a standard set of 42 regions as previously described (Perlmutter et aI., 1 985). Values for nine specific cortical and subcortical regions in each hemisphere also were analyzed. Eight regions were defined anatomically according to coordinates (Table 1) in a stereotaxic atlas of the brain (Talairach et aI., 1 967). The atlas coordinates for the sensorimotor re gion were defined from an analysis of the regions of max imal blood flow response to repetitive opening and closing of the hand during a scan measuring blood flow in eight subjects, as previously described (Perlmutter and Raichle, 1 985). All nine regions in each hemisphere were located in each subject's PET scan using a stereotaxic localization technique that is independent of the appear ance of the physiological images (Fox et aI., 1 9 85). Each PET region of interest was 13.5 x 1 3.5 mm (5 x 5 pixels).
65
TABLE 1. Stereotaxic atlas coordinates for regions
of interest Axes
Region of interest
Vertical
Sensorimotor Inferior parietal Midfrontal Anterior cingulate Occipital Superior temporal Thalamic
5.0 4.0
Putaminal
0. 3 0.0
Pallidal
3.0 2.0 1.6 1.3 0.8
Rightleft
Anteriorposterior
4.6 3.0 3.6 0.7 3.0
-0.6 -3.8 3.2 4.0 -5.3
5.5 1.0
-1. 2 -0.5
2.3 1.6
1.5 1.0
Coordinates (cm) are from a stereotaxic atlas of the human brain (Talairach et a!. , 1967) and are for regions in the left hemi sphere. Homologous right-sided regions have the same vertical and anterior-posterior coordinates with inverse right-left coor dinates. An anatomical localization technique for use with posi tron emission tomography (PET) images that employs propor tionate measurements for each subject was used to calculate the PET coordinates for these regions (Fox et a!. , 1985). All PET regions of interest were 13.5 x 13.5 mm in size.
Left-sided and corresponding right-sided values were compared using two-tailed, paired t tests with Bonfer roni's correction for multiple comparisons. Statistical sig nificance for comparisons of regional values was defined as p < 0.0014, since there were nine pairs of regions and four measurements (CBF, CBV, CMROz, and oxygen ex traction) for each region [i.e., 0.05/(9 x 4) 0.0014]. Regions found to have statistically significant differences between the two sides were defined as asymmetric for that measurement. Statistical significance for differences of left and right hemispheric values was defined as p < 0.0125 (i.e., 0.05/4). =
RESULTS The mean (± SD) arterial carbon dioxide tension during the PET studies for all of the patients was 37 ( ± 4) mm Hg, probably reflecting a modest degree of acute hyperventilation in these subjects. Global mean left hemispheric and mean right hemispheric values of CBF, CBV, CMR02 and oxygen extrac tion are listed in Table 2. Mean left hemispheric CBF and CMR02 were significantly lower than mean right hemispheric values (p < 0. 0001) , whereas mean left hemispheric and corresponding right-sided values for CBV and oxygen extraction were not significantly different. Mean values of CBF, CBV, CMR02, and oxygen extraction for the specific regions of the 32 normal subjects are listed in Table 3. Left-sided values were significantly lower than corresponding right sided values of CBF and CMR02 for sensorimotor, occipital, and superior temporal regions, whereas only the left-sided CBF values were lower for the anterior cingulum. The asymmetries of the sensori motor region did not change depending on the side
J Cereb Blood Flow
Metab, Vol. 7, No.1, 1987
66
1. S. PERLMUTTER ET AL. tween left and right homologous regions, whereas in other cases the statistical significance reflects a small (e. g. , 5% for the occipital region) but consis tent difference that might not have physiological importance. Furthermore, the number of significant asymmetries found depended on the number of re gions that were analyzed as well as the number of patients studied. For example, statistical signifi cance using Bonferroni' s correction for multiple comparisons would be defined as p < 0. 000625 if the number of regions analyzed in each hemisphere were 20 instead of 9 [i. e. , 0.05/(20 x 4)]. Some of the significant asymmetries defined in this study then would not be statistically significant. These nine regions were chosen arbitrarily, and other re gions not chosen also could have been asymmetric. Thus, we do not claim to have identified all possible regional asymmetries within the brain. This study did not investigate the underlying physiological basis for these regional asymmetries. At the least, that would require a higher-resolution scanner as well as precise knowledge of the ana tomical structure of each of the regions of interest for each subject. Partial volume averaging limits the correlation of a PET measurement to the anatom ical region contributing to that measurement (Maz ziotta et al. , 1981). Regional asymmetries of CBF and CMR02 for the sensorimotor region, for ex ample, could result from asymmetries in a nearby region that influence the PET measurements for the sensorimotor area. Given these caveats, what could be the physio logical importance of these asymmetries? They could reflect gross anatomical, cytoarchitectonic,
TABLE 2. Global and hemispheric values for 32 normal
right-handed subjects at rest CBP
Global
CMRO,
(mlllOO gl
CBV
(mlllOO gl
Oxygen
min)
(m11100 g)
min)
extraction
46.8 (7.0)
4.85 (1.00)
2.63 (0.60)
0.357 (0.089)
46.2 (6.8)
4.84 (0.99)
2.60 (0.59)
0.358 (0.089)
47.4 (7.2) 0.18
4.86 (1.04) 0.05
2.65 (0.61) 0.01
0.356 (0.089) 0.002
Left hemispheric Right hemispheric SOD
Values are means (SO) for all 32 subjects. Mean left hemispheric CBF and CMRO, were significantly different from corresponding right-sided values (p < 0.0001, both). Left- and right-sided hemispheric values of cerebral blood volume (CBV ) and oxygen extraction were not signifi cantly different. SOD represents the standard deviation of the differences between paired left- and right-sided values.
of the arterial catheter. Left-sided CBV was signifi cantly lower for the anterior cingulum, but was sig nificantly higher for the midfrontal region. Left sided values of oxygen extraction for the sensori motor region were significantly lower than right-sided values. No other region was asymmetric for oxygen extraction. These asymmetries are un likely to result from systematic error in the place ment of regions as this would have also resulted in similar regional asymmetries of CBV. DISCUSSION The purpose of this study was to determine if there are regional asymmetries of blood flow, blood volume, and oxygen utilization and extraction in normal resting subjects. The data demonstrate sta tistically significant regional asymmetries. In some cases the significant asymmetry represents at least a 10% difference (e. g. , sensorimotor region) be-
TABLE 3. Regional CRF, cerebral blood volume (CRV), CMR02, and oxygen extraction in 32
right-handed subjects at rest CBP (m1/100 glmin)
CBV (m11100 g)
CMRO,
Oxygen
(mlllOO gimin)
extraction
Region
Left
Right
SOD
Left
Right
SOD
Left
Right
SOD
Left
Right
SOD
Sensorimotor
44.2 (8.9)a
49.5 (9.2)
0.770
5.24 (1.23)
5.33 (1.32)
0.113
2.53 (0.57)a
2.98 (0.69)
0.061
0.364 (0.087)
0.384 (0.092)
0.0052
48.8 (8.6) 49.5 (8.5)
48.7 (8.0) 49.5 (8.0)
0.681 0.874
4.53 (1.22) 4.13 (1.04)b
4.43 (1.18) 3.92 (1.06)
0.222 0.057
2.88 (0.68) 2.79 (0.70)
2.92 (0.74) 2.82 (0.070)
0.042 0.062
0.375 (0.098) 0.356 (0.085)
0.378 (0.102) 0.359 (0.091)
0.0048 0.0051
45.6 (9.0)a 39.0 (5.2)a
50.5 (9.1) 41.2 (6.6)
0.911 0.613
4.38 (1.16)a 4.04 (0.90)
4.89 (1.13) 4.57 (1.72)
0.098 0.182
2.43 (0.69) 2.45 (0.54)a
2.57 (0.74) 2.63 (0.70)
0.045 0.047
0.335 (0.089) 0.398 (0.094)
0.319 (0.088) 0.404 (1.00)
0.0045 0.0062
49.5 53.6 55.4 51.5
53.9 (8.5) 53.8 (7.6) 54.6 (7.6) 51.2 (7.0)
0.700 0.591 0.756 0.668
5.16 5.43 5.23 5.66
4.% 5.48 5.17 5.80
0.120 0.115 0.173 0.155
2.78 2.70 2.99 2.62
3.00 2.71 2.97 2.52
0.039 0.045 0.055 0.066
0.354 0.322 0.344 0.323
0.348 0.321 0.345 0.310
0.0060 0.0050 0.0045 0.0061
Inferior parietal Midfrontal Anterior cingulate Occipital Superior temporal Thalamic Putaminal Pallidal
(9.4)a (7.4) (8.8) (7.7)
(1.11) (1.31) (1.48) (1.83)
(1.08) (1.22) (1.60) (2.00)
(0.69)a (0.79) (0.78) (0.85)
(0.77) (0.80) (0.77) (0.90)
(0.090) (0.099) (0.096) (0.108)
(0.093) (0.099) (0.093) (0.109)
Values are means (SO) for all 32 subjects with one exception: only 27 subjects had measurements for the sensorimotor regions, since 5 subjects were placed too low in the scanner to include these regions. SOD represents the standard deviation of the differences between paired left- and right-sided values. Comparisons of left- and right-sided values were made using two-tailed, paired t tests with Bonferroni's correction for multiple comparisons. Statistical significance for comparisons of left and right regional values was defined as p < 0.0014 since there were nine pairs of regions and four measurements (CBP, CBV. CMRO, and oxygen extraction) for each pair of regions [i.e., 0.051(9 x 4) = 0.0014]. a Left-sided values are significantly lower than right-sided values (p < 0.0014).
b Right-sided values are significantly lower than left-sided values (p < 0.00\4).
J
Cereb Blood Flow Metab, Vol. 7, No.1, 1987
NORMAL ASYMMETRIES OF FLOW AND METABOLISM
biochemical, or functional asymmetries. The exis tence of asymmetries in the human brain is well documented (Geschwind and Galaburda, 1985). For example, decreased blood flow and metabolism in the left superior temporal region (Table 1) could re sult from the known gross anatomical asymmetry observed in the region of the planum temporale, thought to reflect the left hemispheric dominance for language in most right-handed individuals (Geschwind and Levitsky, 1968; Geschwind and Galaburda, 1985). One might expect that left-sided blood flow and metabolism would be higher than corresponding right-sided values because the left planum is larger than the right. In fact, Mazziotta et al. (1982) found higher fluorodeoxyglucose uptake in the left temporal regions in 22 right-handed normal volunteers with eyes closed and/or ears par tially occluded. This contrasts with the decreased flow and oxygen utilization that we found on the left. We had the advantage of locating the anatom ical regions of interest on the PET scans with a ste reotaxic method that is entirely independent of the appearance of the physiological images. This may in part account for the differences between our findings and those of the previous study. The un derlying cause of our findings remains unclear. However, the ability to detect physiological asym metries in vivo with PET should eventually con tribute to a better understanding of the lateraliza tion of cerebral functions. Most importantly, knowledge of these asymme tries in normal resting subjects is essential for proper interpretation of PET studies of pathology. Demonstration of either a regional abnormality or an abnormal asymmetry requires a measured or calculated value that is significantly different from an appropriate control value and should not assume that the normal brain has right-left symmetry for any given region. Acknowledgment: We thank Martha Storandt, Ph.D., for expert statistical advice. This research was supported by NIH grants HLl3851 AG/NS03991 and NS06833 and
Teacher-Investigator Awards NS00929 (lS.P.), NS00647 (W.lP.), and NS00904 (P.T.F.).
REFERENCES Caine DB, Langston JW, Martin WRW, Stoessl AJ, Ruth T J, Adam MJ, Pate BD, Schulzer M (1985) Positron emission tomography after MPT P; observations relating to the cause of Parkinson's disease. Nature317:246-248 Engel J, Kuhl DE, Phelps ME, Crandall PH (1982) Comparative
67
ological mechanisms, associations and pathology: 1. A hy pothesis and a program for research. Arch Neurol 42:428459 Geschwind N, Levitsky W (1968) Human brain: left-right asym metries in temporal speech region. Science161:186-187 Grubb RJ Jr, Raichle ME, Higgins CS, Eichling JO (1978) Mea surement of regional cerebral blood volume by emission to mography. AnnNeurol4:322-328 Herscovitch P, Markham J, Raichle ME (1983) Brain blood flow measured with intravenous H2150. 1. Theory and error anal ysis. J Nuci Med 24:782-789 Lammertsma AA (1984) Positron emission tomography of the brain: measurement of regional cerebral function in man. ClinNeurol Neurosurg 86:1-11 Mazziolta JC, Phelps ME, Plummer D, Kuhl DE (1981) Quanti tation in positron emission computed tomography: V. Phys ical-anatomical effects. J ComputAssist Tomogr5:734-743 Mazziotta JC, Phelps ME, Carson RE, Kuhl DE (1982) Tomo graphic mapping of cerebral metabolism: sensory depriva tion. AnnNeurol 12:435-444 Mintun M, Raichle ME, Martin WRW, Herscovitch P (1984) Brain oxygen utilization measured with 0-15 radiotracers and positron emission tomography. J Nuci Med25:177-187 Perlmutter JS, Raichle ME (1985) Regional blood flow in hemi parkinsonism. Neurology 35:1127-1134 Perlmutter JS, Herscovitch P, Powers WJ, Fox PT, Raichle ME (1985) Standardized mean regional method for calculating global positron emission tomographic measurements. J Cereb BloodFlow Metab 5:476-480 Phelps ME, Mazziotta JC (1985) Positron emission tomography : human brain function and biochemistry. Science 228:799809 Phelps ME, Mazziolta JC, Huang SC (1982) Study of cerebral function with positron computed tomography. J Cereb BloodFlow Metab 2:113-162 Powers W J, Raichle ME (1985) Positron emission tomography and its application to the study of cerebrovascular disease in man. Stroke 16:361-376 Raczkowski D, Kalat JW, Nebes R (1974) Reliability and validity of some handedness questionaire items. Neuropsychologia 12:43-47 Raichle ME (1983) Positron emission tomography. Annu Rev Neurosci 6:249-267 Raichle ME, Martin WRW, Herscovitch P, Mintun MA, Markham J (1983) Brain blood flow measured with intrave nous H2150: II. Implementation and validation. J Nuci Med 24:790-798 Reiman EM, Raichle ME, Butler FK, Herscovitch P, Robins E (1984) A focal brain abnormality in panic disorder, a severe form of anxiety. Nature 310:683-685 Rhodes CG, W ise RJS, Gibbs JM, Frackowiak RSJ, Hatazawa J, Palmer AJ, Thomas DGT, Jones T (1983) In vivo distur bance of the oxidative metabolism of glucose in human cere bral gliomas. AnnNeurol 14:614-626 Talairach J, Szikla G, Tournoux P, Prossalentis A, Bordas-Ferrer M, Covella L, lacob M, Mempel E (1967) Atlas dA ' natomie Stereotaxique du Telencephale. Paris, Masson Ter-Pogossian MM, Ficke DC, Hood JT, Yamamoto M, Mullani NA (1982) PET T VI: a positron emission tomograph uti lizing cesium fluoride scintillation detectors. J ComputAs sist Tomogr 6:125-133 T heodore W H, Newmark ME, Sato S, Brooks R, Patronas N, De La Paz R, DiChiro G, Kessler RM, Margolin R, Manning RG, Channing M, Porter RJ (1983) [18F1Fluorodeoxy glucose positron emission tomography in refractory complex partial seizures. Ann Neurol 14:429-437 W olfson LI, Leenders KL, Brown LL, Jones T (1985) Alter
localization of epileptic foci in partial epilepsy by PCT and EEG. AnnNeuroI12:529-537
ations of regional cerebral blood flow and oxygen metabo
Fox PT, Perlmutter JS, Raichle ME (1985) A stereotactic method of anatomical localization for positron emission tomog
Yamamoto M, Ficke DC, Ter-Pogossian MM (1982) Performance study of PETT VI, a positron emission tomograph with 288 cesium fluoride detectors. IEEE Trans Nuci Sci NS29:529533
raphy. J ComputAssist Tomogr4:727-736 Geschwind N, Galaburda AM (1985) Cerebral lateralization. Bi-
lism in Parkinson's disease. Neurology 35:1399-1405
J Cereb
Blood Flow Metab. Vol. 7, No. 1, 1987
