Articles in PresS. J Appl Physiol (December 23, 2004). doi:10.1152/japplphysiol.00653.2004
Cerebral White Matter Blood Flow Is Constant During Human Non-Rapid Eye Movement Sleep: A Positron Emission Tomographic Study
Masahiko Hiroki,1 Takeshi Uema,2 Naofumi Kajimura,3 Kenichi Ogawa,4 Masami Nishikawa,3 Masaaki Kato,3 Tsuyoshi Watanabe,3 Toru Nakajima,6 Harumasa Takano,3 Etsuko Imabayashi,5 Takashi Ohnishi,5 Yutaka Takayama,3 Hiroshi Matsuda,5 Makoto Uchiyama,7 Masako Okawa,8 Kiyohisa Takahashi,3 Hidenao Fukuyama1
1Human Brain Research Center, Kyoto University Graduate School of Medicine, 54 Shogoin, Sakyo-ku, Kyoto, Kyoto 606-8507, Japan; 2Department of Psychiatry, Osaka Prefectural General Hospital, Osaka, Osaka 558-8558, Japan; 3Department of Psychiatry, 4Department of Anesthesiology, and 5Department of Radiology, National Center Hospital for Mental, Nervous, and Muscular Disorders, National Center of Neurology and Psychiatry (NCNP), Kodaira, Tokyo 187–8551, Japan; 6Department of Psychiatry, Teikyo University School of Medicine, Kawasaki, Kanagawa 213-8507, Japan; 7Department of Psychiatry, National Institute of Mental Health, NCNP, Ichikawa, Chiba 272-0827, Japan; and 8Department of Psychiatry, Shiga University of Medical Science, Otsu, Shiga 520-2192, Japan
Running head: Cerebral White Matter CBF in Human Sleep Contact information: Masahiko Hiroki Dept. of Radiology, Massachusetts General Hospital NMR Center Building 149, 13th Street, Mailcode 149-2301, Charlestown, MA 02129-2060 USA
Phone: +1 617 726-3914 Fax: +1 617 726-7422, Email:
[email protected]
Copyright © 2004 by the American Physiological Society.
1
Abstract
This study aimed to identify brain regions with the least decreased cerebral blood flow (CBF) and their relationship to physiological parameters during human non-rapid eye movement (NREM) sleep. Using [15O]-H2O positron emission tomography, CBF was measured for nine normal young adults during nighttime. As NREM sleep progressed, mean arterial blood pressure (MAP) and whole-brain mean CBF decreased significantly; arterial partial pressure of carbon dioxide (Paco2) and, selectively, relative CBF of the cerebral white matter increased significantly. Absolute CBF remained constant in the cerebral white matter – registering 25.9 ± 3.8 during wakefulness, 25.8 ± 3.3 during light NREM sleep, and 26.9 ± 3.0 (ml/100 g/min) during deep NREM sleep (P = 0.592) – and in the occipital cortex (P = 0.611). The regression slope of the absolute CBF significantly differed with respect to Paco2 between the cerebral white matter (slope 0.054, R = cortex (slope
0.776, R =
0.04) and frontoparietal association
0.31) (P = 0.005) or thalamus (slope
1.933, R =
0.47) (P =
0.004) and between the occipital cortex (slope 0.084, R = 0.06) and frontoparietal association cortex (P = 0.021) or thalamus (P < 0.001), and, with respect to MAP, between the cerebral white matter (slope
0.067, R =
0.10) and thalamus (slope 0.637, R = 0.31) (P = 0.044).
The cerebral white matter CBF keeps constant during NREM sleep as well as the occipital cortical CBF, and may be specifically regulated by both CO2 vasoreactivity and pressure autoregulation.
Key words: Cerebral white matter; Carbon dioxide; Non-rapid eye movement sleep;Cerebral blood flow; PET
2
Introduction
Non-rapid eye movement (NREM) sleep accounts for up to 80% of total human sleep (36) and is physiologically and neurophysiologically distinct from wakefulness and rapid eye movement sleep, which are similar (29, 53). While it is known that various physiological parameters change mainly during NREM sleep, the relationship between brain region and each physiological parameter during sleep has remained elusive. It is essential to gain a thorough understanding of this relationship not only for the consideration of confounding factors in sleep research (26, 59), but also for purposes of investigation into the pathogenesis and pathophysiology of circulatory (9, 10), cerebrovascular (15, 25), and neuropsychiatric diseases (24, 33, 41). During human NREM sleep, arterial blood pressure and heart rate fall due to reduced sympathetic nerve system activity (28, 50); and, CBF decreases both globally (23, 30, 49, 58) and regionally (7, 23, 32). It is known that blood carbon dioxide (CO2) is the strongest vasodilator of the brain (18, 28), and is also one of the end products of brain metabolism. It is also known that regional CBF (rCBF) changes linearly with arterial, arteriolar, and precapillary normal Paco2, and is correlated better with Paco2 than with tissue Pco2 (47). Despite blood CO2 increasing CBF up to 7%/mm Hg (22) and accumulating during NREM sleep (51, 63) due to reduced output from medullary respiratory neurons to respiratory muscles (40), a region with increased or maintained CBF relating to Paco2 during the progression of NREM sleep has never been clearly identified. This seems to be a paradox and deserves investigation in search of a neurobiologic explanation.
3
In this study, we hypothesized that the white matter CBF increases or keeps constant, relating to blood CO2, and that this implies a physiological role during sleep. Almost all previous sleep studies, including our previous study (23), have focused mainly on the activation or deactivation of the whole cerebral hemisphere or the gray matter and have paid little attention to the white matter (7, 23, 30, 32, 49, 58), which differs fundamentally from the gray matter in terms of energy requirements and metabolic rate (47). The aim of this study is to identify areas with increased or constant CBF during human NREM sleep and to clarify the relationship of rCBF to blood CO2 and other physiological parameters.
METHODS
Subjects and experimental procedure. Fifteen healthy, right-handed male university students served as study subjects between September 1999 and October 2000. Written informed consent, approved by the Intramural Research Board of the National Center of Neurology and Psychiatry (Kodaira, Tokyo, Japan), was obtained from each subject before participation in the study. Our study did not include sleep deprivation, therefore each subject was instructed to sleep regularly between the hours of 10:00 p.m. and 7:00 a.m. for at least a week prior to the experiment. Subjects’ compliance with the instructions was monitored by actigraph. Those failing to comply with the instructions were excluded from the study. Finally, nine subjects (age, 21.0 ± 1.0 years; range, 20 23 years) completed the positron emission tomography (PET) study during each period of wakefulness and light and deep NREM sleep; their data were used for the analysis. PET scanning started at approximately 9:30 p.m. to obtain some scans during wakefulness. The lights were turned out at 10:30 p.m. and scans for light (stages 1 and 2 sleep;
4
characterized by low voltage, mixed-frequency EEG activity and appearance of K-complex or spindle) and for deep sleep (stages 3 and 4 sleep; characterized by a slow, high amplitude delta wave) were obtained between 11:00 p.m. and 2:00 a.m. The sleep stage scoring was performed according to the standardized sleep manual of Rechtschaffen and Kales (43). Each subject lay on a PET scanner couch with electrodes attached to the head for polysomnography and an individually-molded thermoplastic facemask was used to stabilize the head. A venous line was inserted into the right median antebrachial vein for tracer injection. An arterial line was inserted into the left radial artery for blood pressure monitoring, arterial blood gas analysis, and radioactivity measurements. A flow-through radioactivity monitor (PICO COUNT; Bioscan, Washington, DC) was used to detect the radioactivity of the arterial blood by automatic sampling throughout the scanning period. The length of the catheter was 30 cm from the radial artery to the flow-through radioactivity monitor. The delay and dispersion were simultaneously corrected using the measured arterial time activity curve and the tissue time activity curve (27). Arrival time of the radioactivity between the radioactivity monitor and the brain was not corrected. Mean arterial blood pressure (MAP), pulse pressure, and heart rate were recorded immediately before tracer injection for each scan. Electroencephalograms at F3, F4, C3, C4, P3, P4, Fz, Cz, and Pz with A1 and A2 reference, monopolar electrooculograms from both canthi, and bipolar electromyograms taken from the chin were monitored for sleep stage scoring. PET procedure. A maximum of 12 intravenous injections of 259 MBq (7 mCi)-[15O]H2Owere administered to each subject during relaxed wakefulness, light sleep, and deep sleep. The whole body exposure was 1 mSv, which is the limit in the recommendation of the International Commission on Radiological Protection (38). The [15O]-H2O bolus was automatically flushed intravenously for 15 seconds. With a PET scanner (Siemens ECAT
5
EXACT HR 961; Siemens Medical Systems, Erlangen, Germany) in 3-dimensional mode, scanning started manually 1 second after the initial rise of head counts and continued for 90 seconds. A camera with an axial field of view of 150 mm acquired data simultaneously from 47 consecutive axial planes. An image resolution of 3.8 × 3.8 × 4.7 mm was obtained after backprojection and filtering (Hanning filter; cutoff frequency 0.5, cycles per pixel) and the reconstructed image was displayed in a matrix of 128 × 128 × 47 voxel format (voxel size, 1.732 × 1.732 × 3.125 mm). A 10-minute transmission scan before acquisition of the emission data corrected for tissue attenuation. Functional images of absolute rCBF were produced by with arterial time activity data using autoradiography (20). SPM analysis. Data were analyzed on a Sun Sparc 20 workstation (Sun Computers Japan, Tokyo, Japan) using Analyze version 7.5.4 image display software (Biodynamic Research Unit, Mayo Foundation, Rochester, MN) and SPM (statistical parametric mapping)-99-software (Wellcome Department of Cognitive Neurology, London, UK [http://www.fil.ion.ucl.ac.uk/spm, (last accessed 06/12/02)]) (13), implemented in MATLAB v. 5.3 (The MathWorks, Inc., Sherborn, MA, U.S.A.) for Windows XP (Microsoft Co., USA) on a PC-compatible computer. Spatial normalization was employed to fit each individual set of brain data to a standard brain template in 3-dimensional space in order to correct for differences in brain size and shape and to facilitate inter-subject averaging. The stereotactically normalized scans contained 68 planes (voxel size, 2 × 2 × 2 mm), and smoothing was done with a Gaussian kernel (10 × 10 × 6 mm). SPM uses a standard brain from the MNI (Montreal Neurological Institute, Montreal, PQ), with precise anatomical localizations of significant changes indicated in accordance with the atlas of Talairach and Tournoux (54) by using a numerical transformation formula supplied by http://www.mrc-cbu.cam.ac.uk/Imaging/minispace.html (MRC Cognition and Brain Science
6
Unit, Cambridge, UK, [last accessed 06/12/02]). Global gray matter and global white matter CBF were obtained by tissue segmentation of the smoothed, spatially normalized PET image using SPM analysis. SPM produces probability maps of CSF and gray or white matter, based both on histogram analysis of pixel intensities and on reference to templates of tissue probability distribution created from MRIs of a population of controls. The segmentation is then performed by cluster analysis with a modified mixture model and the likelihoods of each voxel being one of a number of different tissue types (1). The wholebrain mean CBF was determined using both the CBF value and voxel number of each segment of gray matter and white matter. In the analysis of the relative changes in rCBF, the mean whole-brain CBF in each image was normalized to 50 ml/100 g/min by the proportional scaling method, and the gray matter threshold was set at 0.3 to include the white matter. After the appropriate design matrix was specified, the condition of each voxel in each patient was assessed according to the theory of Gaussian fields. The exact level of significance of difference between conditions was characterized by peak amplitude. A cluster threshold was not set in the SPM analysis because we made an a priori hypothesis that the white matter CBF is related to the physiological parameters, which have never been proven to affect a certain localized region of the brain. The exact level of significance of volumes of difference between conditions was characterized by peak amplitude. Voxels that had peak T values greater than 3.45 (uncorrected P = 0.001) were considered to show a significant difference. Finally, this analysis evaluated areas where the relative rCBF significantly increased or decreased among stages of wakefulness-NREM sleep. ROI analysis. In order to assess the absolute rCBF, region of interest (ROI) analysis was applied, based on the results of SPM analysis, to the frontoparietal association cortex (bilateral
7
medial frontal and inferior parietal cortices), thalamus, perirolandic cortex (the primary motor cortex and unimodal somatosensory regions of the parietal cortex), occipital cortex (including the primary and secondary visual cortices in the striate and lateral occipital gyri), and cerebral white matter. Using MRIcro-software (Rorden C, University of Nottingham, Nottingham, UK [http://www.psychology.nottingham.ac.uk/staff/cr1/mricro.html]), each ROI was manually drawn on a prior probability MNI image segmented for gray or white matter. The ROI boundary of the cerebral white matter was defined – on the white matter segmented image of a upper corona radiata slice – so as to include all of the white matter region, sparing the subcortical white matter adjacent to the cerebral cortex in order to exclude the effect of partial volume averaging. The absolute value of each ROI was then automatically calculated for every image. Additionally, based on the distribution of the significantly increased relative CBF in the cerebral white matter during the progression of NREM sleep, we set a ROI at both the frontal lobe white matter and the posterior cerebral white matter including the occipital and parietal lobe. Statistical analysis. Using repeated-measure analysis of variance with Bonferroni’s post hoc procedure, we compared the following: each physiological parameter, whole-brain, global gray matter and global white matter mean absolute CBF, and the absolute CBF in the frontoparietal association cortex, thalamus, perirolandic cortex, occipital cortex, cerebral white matter, frontal lobe white matter, and posterior cerebral white matter. To determine multi-subject correlation between the absolute rCBF of the brain region and each physiological parameter, multiple linear regression analysis was performed. From each linear regression line, average regression slope was calculated for the brain region. Finally, the difference in the average regression slope between brain regions was evaluated using Student’s t test. The level of significance was set at P < 0.05.
8
RESULTS
During the progression of NREM sleep, there were significant reductions in the physiological parameters of MAP, heart rate, arterial pH, and Pao2, and a significant increase in Paco2 (Table 1). No significant difference was found in pulse pressure (P = 0.262) and Pao2 (P = 0.206). Post hoc analysis showed a significant difference between wakefulness and deep sleep in MAP (P = 0.006), heart rate (P < 0.0001), arterial pH (P < 0.0001), and Paco2 (P < 0.0001) and between wakefulness and light sleep in heart rate (P < 0.0001), arterial pH (P < 0.0001), and Paco2 (P < 0.0001). There was a significant difference in whole-brain mean CBF among stages of wakefulness-NREM sleep: 51.1 ± 6.6 during wakefulness, 47.5 ± 6.9 during light sleep, and 47.3 ± 4.7 ml/100 g/min during deep sleep; and in global gray matter mean CBF: 56.1 ± 7.3 during wakefulness, 52.1 ± 7.7 during light sleep, and 51.5 ± 5.0 ml/100 g/min during deep sleep (Fig. 1). Post hoc analysis did not show any significant difference in either whole-brain or global gray matter mean CBF. Global white matter mean CBF did not differ significantly among the stages: 34.4 ± 5.7 during wakefulness, 33.9 ± 4.8 during light sleep, and 34.1 ± 3.6 ml/100 g/min during deep sleep (Fig. 1). SPM analysis showed significantly decreased relative rCBF (Fig. 2A) during both light and deep sleep compared with wakefulness; in the pons, cerebellum, thalamus, and bilateral neocortical regions including the medial frontal and inferior parietal gyri. Furthermore, significantly decreased relative rCBF during deep sleep compared with wakefulness was found in the midbrain and the basal ganglia. No significantly decreased area was found during deep sleep compared with light sleep. Significantly increased relative rCBF during both light and deep
9
sleep compared with wakefulness was shown in the perirolandic cortex, occipital cortex, and – extensively – in the cerebral white matter. Comparing deep sleep with light sleep, significantly increased relative rCBF was restricted to the cerebral white matter (Fig. 2, B and C). Absolute rCBF was significantly different among the wakefulness-NREM sleep states in the frontoparietal association cortex (Fig. 3A, upper left): 54.2 ± 7.9 during wakefulness, 48.8 ± 7.5 during light sleep, and 48.2 ± 5.3 ml/100 g/min during deep sleep (P = 0.004); and in the thalamus (Fig. 3A, upper middle): 61.5 ± 7.2 during wakefulness, 48.5 ± 6.3 during light sleep, and 46.6 ± 4.7 ml/100 g/min during deep sleep (P < 0.0001) (Fig. 4). In the perirolandic cortex (Fig. 3A, upper right), repeated-measure analysis of variance did not show a significant difference in absolute CBF: 50.8 ± 6.4 during wakefulness, 47.1 ± 6.7 during light sleep, and 47.8 ± 5.9 ml/100 g/min during deep sleep (P = 0.067), while post-hoc analysis showed a significant difference between during wakefulness and light sleep (P = 0.036) (Fig. 4). No significant difference in absolute rCBF was found in the occipital cortex (Fig. 3A, lower left): 55.2 ± 6.7 during wakefulness, 54.3 ± 8.4 during light sleep, and 56.7 ± 7.9 ml/100 g/min during deep sleep (P = 0.611); or in the cerebral white matter (Fig. 3A, lower right): 25.9 ± 3.8 during wakefulness, 25.8 ± 3.3 during light sleep, and 26.9 ± 3.0 ml/100 g/min during deep sleep (P = 0.592) (Fig. 4). There was a significant correlation between the absolute CBF of the frontoparietal association cortex and Paco2 (Fig. 5), between that of the thalamus and both Paco2 and MAP (Figs. 5 and 6), between that of the frontoparietal association cortex and both heart rate (R = 0.42, P < 0.001) and pH (R = 0.49, P < 0.0001), between that of the thalamus and each of heart rate (R = 0.64, P < 0.0001), pH (R = 0.64, P < 0.0001), and Pao2 (R = 0.34, P = 0.006), and between that of the perirolandic cortex and both heart rate (R = 0.31, P = 0.014) and pH (R = 0.33, P = 0.009).
10
No significant correlation was found between either absolute CBF of the cerebral white matter or occipital cortex and any physiological parameter including Paco2 or MAP (Figs. 5 and 6), or in any other comparison. A significant difference in the average regression slope of the absolute rCBF was detected with respect to Paco2 between the occipital cortex and both frontoparietal association cortex and thalamus and between the cerebral white matter and both frontoparietal cortex and thalamus, and, with respect to MAP, between the cerebral white matter and thalamus (Table 2). The absolute CBF of the frontal lobe white matter was significantly higher than that of the posterior cerebral white matter during wakefulness and light sleep (Fig. 3B and 7). No significant difference in the absolute CBF between both white matters was found during deep sleep. There was no significant correlation between either white matter CBF and any physiological parameter through wakefulness-NREM sleep states and no significant difference in any average regression slope between both white matters.
DISCUSSION
Validity of increased relative rCBF. In the SPM analysis, we carefully analyzed increased relative rCBF in order to exclude false-positive areas that can arise from the lesser-reduced rCBF in the normalization of lower whole-brain mean CBF. Therefore, we applied proportional scaling for the whole-brain CBF normalization, since this model is considered appropriate to minimize false-positive areas with less decreased CBF while whole-brain mean CBF decreased. Finally, we confirmed such areas as true-positive by ROI analysis, which showed the absolute CBF of the cerebral white matter where it is considered that the partial volume averaging effect from the gray matter to the subcortical white matter can be minimized. Based on the evidence that CBF
11
increased 30% with a 4.4-mm-Hg increase of Paco2 through the CO2 inhalation method in wakefulness (22), it is reasonable that, even if taking different subjects’ states of consciousness into consideration, areas of the brain with maintained CBF could exist in our study with a 3.8mm-Hg increase of Paco2. Constant absolute CBF of the cerebral white matter during NREM sleep. We showed that absolute CBF of the cerebral white matter was selectively constant during human NREM sleep, despite the blood pressure reduction. From the viewpoint of tissue specificity, what is the physiological role of CBF maintenance of the cerebral white matter during sleep? During the progression of NREM sleep, blood pressure falls, blood CO2 piles up, and the CBF in most gray matters including the cerebral association cortex and thalamus decreases depending on the metabolic rate. Although such deactivation may affect the cerebral white matter where the reciprocal thalamo-cortico-thalamic loop courses, the white matter is basically distinct from gray matter in terms of energy-tolerance and flow vulnerability. It is known that the gray matter, especially the synaptic region on the distal dendritic tree where glutaminergic terminals and astrocytic processes are located, is a highly specialized metabolic unit in which the activation signal is furnished by the neuron or modulated by the astrocyte, and consequently has a very large capacity to change the metabolic rate (31). On the other hand, oligodendroglia/myelin of the white matter probably has a metabolic rate about one-third of that in the gray matter at the Ranvier's nodes where Na+/K+ dependent ATPase operates to restore ionic gradients (55). Both oligodendroglia/myelin and axon have a limited capacity to change the metabolic rate (47). Furthermore, the cerebral white matter is more vulnerable to chronic cerebral hypoperfusion than the gray matter (8, 61), in association with DNA fragmentation in the oligodendroglia (57).
12
From a neurodevelopmental viewpoint, the white matter of young adults seems to be different from that of the elderly adults. MRI studies showed that the volume of the cerebral white matter increased linearly from ages 4 to 20 (3, 14). This increase is probably based on maturational process such as myelination or axonal growth (4, 21). On the other hand, cortical gray matter volume continues to decrease linearly after adolescence (17, 42), probably due to shrinkage of large neurons (56). Therefore, during NREM sleep in young adults, which after all was our study population, the CBF demand on the cerebral white matter may be generally large compared with that on the gray matter. Absolute CBF of the occipital cortex and perirolandic cortex. Absolute CBF of the occipital cortex was also constant during NREM sleep. During this period, it is known that CBF does not change in the unimodal somatosensory regions or sensory regions including the perirolandic cortex and occipital cortex (7, 23). These regions may be spared from the effect of intralaminar thalamus by the brainstem reticular formation during NREM sleep and may be preserved within local circuitry or represent intracortical or transcallosal transfer of information (7). However, in the present study, absolute CBF of the perirolandic cortex showed a significant decrease during light NREM sleep. This is considered to be due to the fact that, contrary to previous studies, subjects in this study did not undergo a sleepdeprivation procedure, which increases the nerve growth factor expression in the pyramidal neurons in the rat somatosensory cortex (6). The SPM result showing an increase in relative rCBF in the perirolandic cortex during light NREM sleep compared with wakefulness is probably due to the normalization artifact caused by decreased whole-brain mean CBF.Mechanism of maintenance of the cerebral white matter and occipital cortex CBF during NREM sleep. Mechanism of maintenance of the cerebral white matter and occipital cortex CBF during
13
NREM sleep. CBF is primarily regulated by chemical stimuli (CO2 vasoreactivity), perfusion pressure (pressure autoregulation), and neural stimuli. During NREM sleep, the cerebral white matter CBF may be regulated by the former two mechanisms and the occipital cortical CBF by CO2 vasoreactivity. In the cerebral white matter, the relevance of CO2 vasoreactivity is based on the fact that, while the absolute CBF of the thalamus and frontoparietal association cortex showed a significant negative correlation to Paco2 through wakefulness-NREM sleep states, the absolute CBF of the cerebral white matter remained constant with change of Paco2 (Fig. 5 and Table 2). Although we could not assess any increased cerebral blood volume or vasodilatation, the constant CBF may be due to the accumulated blood CO2 based on the below reports. Some studies have considered CO2 vasoreactivity to have a global effect during sleep (7, 45), but other studies have reported a difference in CO2 vasoreactivity between the cerebral cortex and cerebral white matter. Using the CO2 inhalation method during wakefulness, the cerebral cortex showed a marked vasoreactivity compared with the cerebral white matter (44, 60). On the other hand, cerebral white matter showed a strong vasoreactivity in response to a 1-mm Hg change from the mean Paco2 (16). During sleep, CO2 vasoreactivity of the cerebral cortex decreased compared with wakefulness (35). These seem consistent with our findings. The occipital cortex resembled the cerebral white matter in the relationship of absolute rCBF to Paco2 (Fig. 5 and Table 2). Regional difference in CO2 vasoreactivity among the cerebral cortices has never been reported under sleep conditions. Since the occipital cortex, especially the primary visual cortex, has the thickest layer 4B (the line of Gennari), which contains the richest myelin concentration (2, 11), it is speculated that there exists an agent interacting between oligodendroglia/myelin and vessels during sleep. Atrial natriuretic peptide, which we could not measure due to methodological limitations, is a possible intermediate agent
14
that acts as a vasodilator with a peak concentration in the late sleep stage (52), is released according to the increase in blood CO2 (12), and has an affinity for the white matter (19). Pressure autoregulation is another possible mechanism for the maintenance of the cerebral white matter CBF. The absolute CBF of the cerebral white matter also remained constant against the change in MAP and showed a specifically lower regression slope with respect to MAP than did absolute CBF of the thalamus (Fig. 6, Table 2). This may reflect the effect of the pressure autoregulation, which works to maintain constant perfusion by contraction or relaxation of the (mainly arteriolar) smooth muscle in the face of blood pressure changes. Since the cerebral white matter area involves a boundary zone with limited collateral flow between the middle cerebral artery perforators and the medullary branches of the pial arteries (37, 46), this area, especially the periventricular region, is vulnerable to reductions in cerebral perfusion pressure (5). In animal experiments, pressure autoregulation of the cerebral white matter is less effective than that of the gray matter (62). However, under the condition that CBF in most gray matters decreases depending on neuronal deactivation, pressure autoregulation may work selectively in the maintenance of the cerebral white matter CBF. Among the cerebral white matter, a regional difference may exist regarding CBF regulation. Based on our results that the frontal lobe white matter increased less in absolute CBF (Fig. 2C and 7), the frontal lobe white matter seemed to change less in relationship with physiological parameters during NREM sleep. This may imply that the frontal lobe activity is physiologically allowed to change the least. Overall, it is speculated that the cerebral white matter is physiologically protected from the gradual hypoperfusion of the brain recurring every night. From a pathophysiological perspective,
15
further investigation is needed to clarify whether the CBF maintenance of the cerebral white matter can be observed in the elderly subjects whose gray and white matter have declined due to degenerative processes, and whether the CBF maintenance of the cerebral white matter has clinical relevance in subjects with the CBF dysregulation that causes independent CBF change of Paco2 or arterial blood pressure (34, 39, 48). Conclusion. We showed a unique state- and region-specific difference in CBF regulation in young adult brains. During NREM sleep, the absolute CBF of the cerebral white matter keeps constant as well as that of the occipital cortex, and may be specifically regulated by both CO2 vasoreactivity and pressure autoregulation, although that remains unproven or some mechanisms should be discovered. Researchers should take this phenomenon into consideration when CO2 or blood pressure is treated as a confounding factor in brain activation/deactivation studies of the sleep state. The cerebral white matter appears to be protected for some biophysiological reasons. To determine whether our results have any clinical relevance, further investigation during sleep is required in the elderly and in subjects with impaired CBF regulation.
16
REFERENCES 1.
Ashburner J and Friston K. Multimodal image coregistration and partitioning–a unified framework. Neuroimage 6: 209–217, 1997.
2.
Barbier EL, Marrett S, Danek A, Vortmeyer A, van Gelderen P, Duyn J, Bandettini P, Grafman J, and Koretsky AP. Imaging cortical anatomy by high-resolution MR at 3.0T: detection of the stripe of Gennari in visual area 17. Magn Reson Med 48: 735 758, 2002
3.
Bartzokis G, Beckson M, Lu PH, Nuechterlein KH, Edwards N, and Mintz J. Agerelated changes in frontal and temporal lobe volumes in men: a magnetic resonance imaging study. Arch Gen Psychiatry 58: 461–465, 2001.
4.
Benes FM, Turtle M, Khan Y, and Farol P. Myelination of a key relay zone in the hippocampal formation occurs in the human brain during childhood, adolescence, and adulthood. Arch Gen Psychiatry 51: 477–484, 1994.
5.
Bladin CF and Chambers BR. Clinical features, pathogenesis, and computed tomographic characteristics of internal watershed infarction. Stroke 24: 1925–1932, 1993.
6.
Brandt JA, Churchill L, Guan Z, Fang J, Chen L, and Krueger JM. Sleep deprivation but not a whisker trim increases nerve growth factor within barrel cortical neurons. Brain Res 13: 105 112, 2001.
7.
Braun AR, Balkin TJ, Wesensten NJ, Carson RE, Varga M, Baldwin P, Selbie S, Belenky G, and Herscovitch P. Regional cerebral blood flow throughout the sleep-wake cycle: an H215O PET study. Brain 120: 1173–1197, 1997.
8.
Cai Z, Pang Y, Xiao F, and Rhodes PG. Chronic ischemia preferentially causes white matter injury in the neonatal rat brain. Brain Res 898: 126 135, 2001.
17
9.
Cuspidi C, Macca G, Sampieri L, Fusi V, Severgnini B, Michev I, Salerno M, Magrini F, and Zanchetti A. Target organ damage and non-dipping pattern defined by two sessions of ambulatory blood pressure monitoring in recently diagnosed essential hypertensive patients. J Hypertens 19: 1539 1545, 2001.
10. Elliott WJ. Circadian variation in blood pressure: implications for the elderly patient. Am J Hypertens 12: 43S 49S, 1999. 11. Fatterpekar GM, Naidich TP, Delman BN, Aguinaldo JG, Gultekin SH, Sherwood CC, Hof PR, Drayer BP, and Fayad ZA. Cytoarchitecture of the human cerebral cortex: MR microscopy of excised specimens at 9.4 Tesla. AJNR Am J Neuroradiol 23: 1313 1321, 2002. 12. Fontana F, Bernardi P, Tartuferi L, Boschi S, De Iasio R, and Merlo Pich E. Mechanisms of hypertension in patients with chronic obstructive pulmonary disease and acute respiratory failure. Am J Med 109: 621–627, 2000. 13. Friston KJ, Holmes AP, Worsley KJ, Poline JB, Frith CD, and Frackowiak RSJ. Statistical parametric maps in functional imaging: a general linear approach. Hum Brain Map 2: 189 210, 1995. 14. Giedd JN, Blumenthal J, Jeffries NO, Castellanos FX, Liu H, Zijdenbos A, Paus T, Evans AC, and Rapoport JL. Brain development during childhood and adolescence: a longitudinal MRI study. Nat Neurosci 2: 861 863, 1999. 15. Goldstein IB, Bartzokis G, Hance DB, and Shapiro D. Relationship between blood pressure and subcortical lesions in healthy elderly people. Stroke 29: 765 772, 1998. 16. Greenberg JH, Alavi A, Reivich M, Kuhl D, and Uzzell B. Local cerebral blood volume response to carbon dioxide in man. Circ Res 43: 324 331, 1978.
18
17. Gur RC, Turetsky BI, Matsui M, Yan M, Bilker W, Hughett P, and Gur RE. Sex differences in brain gray and white matter in healthy young adults: correlations with cognitive performance. J Neurosci 15: 4065 4072, 1999. 18. Guyton AC and Hall JE. Cerebral blood flow, the cerebraospinal fluid, and brain metabolism. In: Textbook of Medical Physiology. Philadelphia, PA: Saunders, 2000, p. 709 715. 19. Herman JP, Dolgas CM, Rucker D, and Langub Jr. MC. Localization of natriuretic peptide activated guanylate cyclase mRNAs in the rat brain. J Comp Neurol 27: 165 187, 1996. 20. Herscovitch P, Markham J, and Raichle ME. Brain blood flow measured with intravenous H215O. I. Theory and error analysis. J Nucl Med 24: 782 789, 1983. 21. Hunter SF, Leavitt JA, and Rodriguez M. Direct observation of myelination in vivo in the mature human central nervous system. A model for the behaviour of oligodendrocyte progenitors and their progeny. Brain 120: 2071 2082, 1997. 22. Ito H, Kanno I, Ibaraki M, and Hatazawa J. Effect of aging on cerebral vascular response to Paco2 changes in humans as measured by positron emission tomography. J Cereb Blood Flow Metab 22: 997 1003, 2002. 23. Kajimura N, Uchiyama M, Takayama Y, Uchida S, Uema T, Kato M, Sekimoto M, Watanabe T, Nakajima T, Horikoshi S, Ogawa K, Nishikawa M, Hiroki M, Kudo Y, Matsuda H, Okawa M, and Takahashi K. Activity of midbrain reticular formation and neocortex during the progression of human non-rapid eye movement sleep. J Neurosci 15: 10065–10073, 1999.
19
24. Kanemaru A, Kanemaru K, and Kuwajima I. The effects of short-term blood pressure variability and nighttime blood pressure levels on cognitive function. Hypertens Res 24: 19 24, 2001. 25. Kario K, Pickering TG, Matsuo T, Hoshide S, Schwartz JE, and Shimada K. Stroke prognosis and abnormal nocturnal blood pressure falls in older hypertensives. Hypertension 38: 852 857, 2001. 26. Krauchi K and Wirz-Justice A. Circadian clues to sleep onset mechanisms. Neuropsychopharmacology 25: 92S 96S, 2001. 27. Lammertsma AA, Cunningham VJ, Deiber MP, Heather JD, Bloomfield PM, Nutt J, Frackowiak RS, and Jones T. Combination of dynamic and integral methods for generating reproducible functional CBF images. J Cereb Blood Flow Metab 10: 675 686, 1990. 28. Levic JR. Control of blood vessels I: Local control. In: An Introduction to Cardiovascular Physiology. London: Arnold, 2000, p. 261 278. 29. Llinas RR and Pare D. Of dreaming and wakefulness. Neuroscience 44: 521 535, 1991. 30. Madsen PL, Schmidt JF, Wildschiodtz G, Friberg L, Hoim S, Vorstrup S, and Lassen NL. Cerebral oxygen metabolism and cerebral blood flow in humans during deep and rapid-eye-movement sleep. J Appl Physiol 70: 2597–2601, 1991. 31. Magistretti PJ. Brain energy metabolism. In: Fundamental Neuroscience, edited by Zignond MJ, Bloom FE, Landis SC, Roberts JL, and Squire LR. San Diego, CA: Academic Press, 1999. 32. Maquet P, Degueldre C, Delfiore G, Aerts J, Peters JM, Luxen A, and Franck G. Functional neuroanatomy of human slow wave sleep. J Neurosci 15: 2807–2812, 1997.
20
33. Matarazzo EB. Psychiatric features and disturbance of circadian rhythm of temperature, pulse, and blood pressure in Wilson's disease. J Neuropsychiatry Clin Neurosci 14: 335 339, 2002. 34. Matsushita K, Kuriyama Y, Nagatsuka K, Nakamura M, Sawada T, and Omae T. Periventricular white matter lucency and cerebral blood flow autoregulation in hypertensive patients. Hypertension 23: 565 568, 1994. 35. Meadows GE, Dunroy HM, Morrell MJ, and Corfield DR. Hypercapnic cerebral vascular reactivity is decreased, in humans, during sleep compared to wakefulness. J Appl Physiol 94: 2197 2202, 2003. 36. Mendelson WB. Human Sleep: Research and Clinical Care. New York: Plenum, 1987. 37. Moody DM, Bell MA, and Challa VR. Features of the cerebral vascular pattern that predict vulnerability to perfusion or oxygenation deficiency: an anatomic study. Am J Neuroradiol 11: 431 439, 1990. 38. Mountford PJ and Temperton DH. Recommendations of the International Commission on Radiological Protection (ICRP) 1990. Eur J Nucl Med 19: 77 79, 1992. 39. Okeda R, Arima K, and Kawai M. Arterial changes in cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL) in relation to pathogenesis of diffuse myelin loss of cerebral white matter: examination of cerebral medullary arteries by reconstruction of serial sections of an autopsy case. Stroke 33: 2565 2569, 2002. 40. Orem J, Osorio I, Brooks E, and Dick T. Activity of respiratory neurons during NREM sleep. J Neurophysiol 54: 1144 1156, 1985.
21
41. Plaschke M, Trenkwalder P, Dahlheim H, Lechner C, and Trenkwalder C. Twentyfour-hour blood pressure profile and blood pressure responses to head-up tilt tests in Parkinson's disease and multiple system atrophy. J Hypertens 16: 1433 1441, 1998. 42. Raz N, Gunning FM, Head D, Dupuis JH, McQuain J, Briggs SD, Loken WJ, Thornton AE, and Acker JD. Selective aging of the human cerebral cortex observed in vivo: differential vulnerability of the prefrontal gray matter. Cereb Cortex 7: 268 282, 1997. 43. Rechtschaffen A and Kales AA. Manual of Standardized Terminology, Techniques and Scoring System for Sleep Stages of Human Subjects. Bethesda, MD: US Department of Health, Education, and Welfare, 1968. 44. Reich T and Rusinek H. Cerebral cortical and white matter reactivity to carbon dioxide. Stroke 20: 453 457, 1989. 45. Reiman EM, Raichle ME, Robins E, Butler FK, Herscovitch P, Fox P, and Perlmutter J. The application of positron emission tomography to the study of panic disorder. Am J Psychiatry 143: 469–477, 1986. 46. De Reuck J. The human periventricular arterial blood supply and the anatomy of cerebral infarctions. Eur Neurol 5: 321 334, 1971. 47. Roland PE. Cellular metabolism and blood flow. In: Brain Activation. New York: WileyLiss, 1993, p. 51 82. 48. Roman GC. Senile dementia of the Binswanger type. A vascular form of dementia in the elderly. JAMA 258: 1782 1788, 1987. 49. Sakai F, Meyer JS, Karacan I, Derman S, and Yamamoto M. Normal human sleep: regional cerebral hemodynamics. Ann Neurol 7: 471–478, 1980.
22
50. Sherwood A, Steffen PR, Blumenthal JA, Kuhn C, and Hinderliter AL. Nighttime blood pressure dipping: the role of the sympathetic nervous system. Am J Hypertens 15: 111 118, 2002. 51. Skatrud JB, Dempsey JA, Badr S, and Begle RL. Effect of airway impedance on CO2 retention and respiratory muscle activity during NREM sleep. J Appl Physiol 65: 1676 1685, 1988. 52. Sothern RB, Vesely DL, Kanabrocki EL, Hermida RC, Bremner FW, Third JL, Boles MA, Nemchausky BM, Olwin JH, and Scheving LE. Temporal (circadian) and functional relationship between atrial natriuretic peptides and blood pressure. Chronobiol Int 12: 106 120, 1995. 53. Steriade M, McCormic DA, and Sejnowski TJ. Thalamocortical oscillations in the 125 sleeping and arousal brain. Science 262: 679 685, 1993. 54. Talairach J and Tournoux P. Three-dimensional Atlas of a Human Brain, Co-planar Stereotaxic Atlas of the Human Brain. Stuttgart: Thieme-Verlag, 1988. 55. Tettamanti M, Paulesu E, Scifo P, Maravita A, Fazio F, Perani D, and Marzi CA. Interhemispheric transmission of visuomotor information in humans: fMRI evidence. J Neurophysiol 88: 1051 1058, 2002. 56. Terry RD, DeTeresa R, and Hansen LA. Neocortical cell counts in normal human adult aging. Ann Neurol 21: 530 539, 1987. 57. Tomimoto H, Ihara M, Wakita H, Ohtani R, Lin JX, Akiguchi I, Kinoshita M, and Shibasaki H. Chronic cerebral hypoperfusion induces white matter lesions and loss of oligodendroglia with DNA fragmentation in the rat. Acta Neuropathol (Berl) 106: 527 534, 2003.
23
58. Townsend RE, Prinz PN, and Obrist WD. Human cerebral blood flow during sleep and waking. J Appl Physiol 35: 620 625, 1973. 59. Trinder J, Kleiman J, Carrington M, Smith S, Breen S, Tan N, and Kim Y. Autonomic activity during human sleep as a function of time and sleep stage. J Sleep Res 10: 253 264, 2001. 60. Vesely A, Sasano H, Volgyesi G, Somogyi R, Tesler J, Fedorko L, Grynspan J, Crawley A, Fisher JA, and Mikulis D. MRI mapping of cerebrovascular reactivity using square wave changes in end-tidal Pco2. Magn Reson Med 45: 1011 1013, 2001. 61. Wakita H, Tomimoto H, Akiguchi I, Matsuo A, Lin JX, Ihara M, and McGeer PL. Axonal damage and demyelination in the white matter after chronic cerebral hypoperfusion in the rat. Brain Res 31: 63 70, 2002. 62. Young RS, Hernandez MJ, and Yagel SK. Selective reduction of blood flow to white matter during hypotension in newborn dogs: a possible mechanism of periventricular leukomalacia. Ann Neurol 12: 445 448, 1982. 63. Zoccoli G, Walker AM, Lenzi P, and Franzini C. The cerebral circulation during sleep: regulation mechanisms and functional implications. Sleep Med Rev 6: 443 455, 2002.
24
Figure legends Fig. 1. Changes in cerebral blood flow (CBF) during human non-rapid eye movement sleep. There was a significant change in whole-brain mean CBF (*, P = 0.041) and global gray matter mean CBF (†, P = 0.030) respectively, but no significant difference in global white matter mean CBF (P = 0.939). Post hoc analysis did not show any significant difference in either whole-brain or global gray matter mean CBF.
Fig. 2. Results of statistical parametric mapping (SPM) analysis. A, B: Orthogonal projections of the brain areas with significantly changed relative rCBF (height P < 0.001). A: Significant decrease was shown in the pons, cerebellum, thalamus, and neocortical regions during light sleep compared with wakefulness. Furthermore, during deep sleep, compared with wakefulness, significant decrease was found in the midbrain and basal ganglia. No such regions were detected during deep sleep compared with light sleep. B: Significant increase was found in the occipital cortex, perirolandic cortex, and extensive cerebral white matter during light and deep sleep compared with wakefulness. Comparing deep sleep with light sleep, significant increase was detected in the cerebral white matter. C: Transverse sections of the standardized brain MRI of the areas showed a significantly increased relative rCBF in deep sleep compared with light sleep. These areas were found to be restricted to the bilateral cerebral white matter.
25
Fig. 3. Region of interest (ROI). Based on the results of the SPM (statistical parametric mapping) analysis showing areas with a significantly decreased or increased relative CBF during the progression of NREM sleep (Fig. 2), ROI was drawn on a prior probability MNI image in each of the following: the bilateral frontoparietal association cortices (z = 30 mm, segmented for gray matter; A, upper left), bilateral thalami (z = 6 mm, segmented for gray matter; A, upper middle), bilateral perirolandic cortices (z = 42 mm, segmented for gray matter; A, upper right), bilateral occipital cortices (z = 12 mm, segmented for gray matter; A, lower left), and bilateral cerebral white matter at the upper corona radiata level (z = 24 mm, segmented for white matter; A, lower right). In the cerebral white matter, we set ROI at the bilateral frontal lobe and posterior cerebral white matter (z = 18 mm, segmented for white matter; B, left and right, respectively).
Fig. 4. Absolute CBF of the frontoparietal association cortex, thalamus, perirolandic cortex, occipital cortex, and cerebral white matter, measured by the ROI analysis. In the absolute CBF of the frontoparietal association cortex, light and deep NREM sleep were significantly lower than wakefulness (*, P = 0.007 and 0.005, respectively). In the absolute CBF of the thalamus, light and deep NREM sleep were significantly lower than wakefulness (†, P < 0.0001, respectively). In the absolute CBF of the perirolandic cortex, light NREM sleep was significantly lower than wakefulness (‡, P = 0.036). No significant difference was found in the absolute CBF of the occipital cortex or cerebral white matter.
26
Fig. 5. Relationship between Paco2 and absolute rCBF during wakefulness-NREM sleep. Significant negative correlation was found between Paco2 and the frontoparietal association cortex and thalamus. There was no significant correlation of Paco2 to the perirolandic cortex, occipital cortex, or cerebral white matter.
Fig. 6. Relationship between mean arterial blood pressure (MAP) and absolute rCBF during wakefulness-NREM sleep. Significant positive correlation was found between MAP and the frontoparietal association cortex and thalamus. There was no significant correlation of MAP to the perirolandic cortex, occipital cortex, or cerebral white matter.
Fig. 7. Regional difference of absolute rCBF of the cerebral white matter. The CBF significantly differed between the frontal lobe and posterior cerebral white matter during wakefulness (*, P = 0.003) and light sleep (†, P = 0.045). No significant difference was found between both white matter CBFs during deep sleep (P = 0.071), or in either frontal lobe or posterior cerebral white matter CBF among stages of wakefulness-NREM sleep (P = 0.765 and 0.731, respectively).
27
Table 1. Physiological parameters Wakefulness
Light sleep
Deep sleep
(42 scans)
(20 scans)
(17 scans)
Mean arterial blood pressure (mm Hg)
84.6 ± 9.1
79.4 ± 9.5
76.8 ± 10.0*
Pulse pressure (mm Hg)
57.8 ± 7.9
58.5 ± 9.7
54.3 ± 7.9
Heart rate (beats per minute)
74.5 ± 10.3
62.7 ± 9.5
60.6 ± 7.6†
pH
7.413 ± 0.020
7.382 ± 0.026
7.369 ± 0.024‡
Paco2 (mm Hg)
41.7 ± 2.0
44.8 ± 2.7
45.5 ± 3.1§
Pao2 (mm Hg)
107.1 ± 17.0
101.5 ± 6.3
101.5 ± 10.5
Arterial blood gas
Values are expressed as mean ± SD. Significant difference was found in mean arterial blood pressure (*, P = 0.011), heart rate (†, P < 0.0001), arterial pH (‡, P < 0.0001), and Paco2 (§, P < 0.0001).
28
Table 2. Difference in average regression slope between brain regions Average slope
1
2
3
4
5
Paco2 1. Frontoparietal association cortex
0.776
—
2. Thalamus
1.933
0.886
3. Perirolandic cortex
0.338
0.890
— 1.062
—
4. Occipital cortex
0.084
2.238*
4.326†
1.022
—
5. Cerebral white matter
0.054
3.011‡
3.120§
1.567
1. Frontoparietal association cortex
0.307
—
2. Thalamus
0.637
0.592
—
3. Perirolandic cortex
0.177
0.821
0.519
—
4. Occipital cortex
0.014
0.579
0.186
0.780
—
0.305
0.088
0.518
0.821
0.115
—
Mean arterial blood pressure
5. Cerebral white matter
0.067
—
Difference of two average regression slopes was expressed as t value. Significant difference in the slopes was found in Paco2 between the occipital cortex and frontoparietal association cortex (*, P = 0.021) and thalamus (†, P < 0.001) and between the cerebral white
29
matter and frontoparietal association cortex (‡, P = 0.005) and thalamus (§, P = 0.004), and in mean arterial blood pressure between the cerebral white matter and thalamus ( , P = 0.044).
30
Fig. 1.
Fig. 2.
31
Fig. 3.
Fig. 4.
32
Fig. 5.
Fig. 6.
33
Fig. 7.
