| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
PRESIDENTIAL ADDRESS |
From the Departments of Psychiatry and Psychology, Western Psychiatric Institute and Clinic, University of Pittsburgh, Pittsburgh, Pennsylvania
Address reprint requests to: J. Richard Jennings, PhD, E1329 WPIC, 3811 O’Hara St., Pittsburgh, PA 15213. Email: JenningsJR{at}msx.upmc.edu
ABSTRACT
OBJECTIVE: This presentation seeks to demonstrate the use of brain imaging techniques for understanding the interaction between hypertension and psychosocial function.
METHODS: The historical background for the study of brain function among hypertensive patients is reviewed. An initial and a current project examining rCBF with 15O water radiotracer and PET in unmedicated hypertensives and normotensives are described. The rCBF response is assessed during the performance of spatial and verbal working memory tasks of increasing memory load. The assessment also addresses the influence on rCBF and performance of white matter hyperintensities and the presence of carotid artery thickening.
RESULTS: Initial results suggest that hypertensives relative to normotensives show less CBF and less posterior parietal rCBF in response to increases in memory load. Hypertensives, however, increase lateral prefrontal (Broca’s area)/insula and amygdala/hippocampal rCBF more than normotensives.
CONCLUSION: Initial results are sufficient to show that hypertension induces changes in rCBF. A tentative hypothesis is that a relatively general decrease in rCBF responsivity induces specific compensatory cognitive strategies as well as subcortical activation. The rCBF changes appear to have implications for information processing and, as such, hold promise for understanding prior reports relating hypertension to affective regulation and cardiovascular reactivity. Imaging techniques provide a powerful tool for psychosomatic research.
Key Words: hypertension, • information processing, • brain imaging, • positron emission tomography, • magnetic resonance imaging, • white matter hyperintensities, • carotid atherosclerosis.
Abbreviations: PET = positron emission tomography;; MRI = magnetic resonance imaging;; CBF = cerebral blood flow;; rCBF = regional cerebral blood flow;; ROI = region of interest;; SPM = statistical parametric mapping
Founding members of the American Psychosomatic Society, such as Helen Flanders Dunbar and Franz Alexander, identified hypertension as a disease with a significant psychosocial component. Franz Alexander (1) organized a symposium on hypertension that was printed in the first issue of Psychosomatic Medicine. Helen Flanders Dunbar’s (2) influential book, Psychosomatic Diagnosis, devoted a chapter to hypertension (as have later influential reviews (3)). A variety of different approaches have supplanted the psychoanalytic approach used by these authors, but the question remains of how psychological factors influence hypertension and how hypertension influences psychological processes. The primary aim of this presentation is to show how a relatively new tool of the neurosciences, brain imaging, can be brought to bear on the question of how hypertension may influence psychological processes. I will review our work in progress (see Acknowledgments for collaborators), presenting only preliminary findings that are illustrative of the promise of this technique. First, however, I will trace the origin of our work and the hypothesis that is guiding the research.
The origins of our current research lie in the psychosomatic research on hypertension performed at the University of Pittsburgh by Drs. Shapiro (4) and Sapira and collaborators (5). Our research grew directly from research by Shapiro (4), where Miller and King examining the hypothesis that hypertension would influence cognitive function. Initial results were positive and suggested that attentional, perceptual-motor, and short-term memory processes might be influenced by the disease (6). These findings were extended and corroborated by others (7–10) and led to a number of studies that refined sample selection and neuropsychological measurement. My contribution to this work was in the refinement of cognitive tasks; a skill that fit with my ongoing work on influence of information processing on autonomic nervous system responses (11). In 1991, Waldstein et al. (12) reviewed progress in this area. This review identified and evaluated the various confounding variables, reviewed the consistency of findings, and concluded that the most consistent and defensible finding was a mild impairment in working or short-term memory among hypertensives. The magnitude of the memory effect was small, however, and a number of the studies found impairments (albeit inconsistently across studies) in areas other than memory.
Figure 1 is illustrative of the magnitude of the neuropsychological differences between hypertensives and controls. The figure is based on results from a study on 123 unmedicated, relatively young hypertensives (age range 27–56 years, see 13). Patients and age-matched controls (N = 50) performed a test in which three letters had to be memorized and then transformed to letters three later in the alphabet, eg, "a r g" transforms to "d u j." Note that both mental transformation and the holding of the basic information while doing the transformation (working memory) are tested. The figure shows the number of sets of three completed in a 2-minute period. In the two bars on the left, hypertensives are shown to perform slightly but significantly fewer transformations than normotensives. The two bars on the right provide an intuitive benchmark for the size of this effect. The sample was split at its median age (ignoring hypertensive status) and the same analysis performed. A similar result showed that older individuals performed slightly, but significantly, fewer transforms than young individuals. The age differential is slight in this sample; the figure compares people in their mid-30s to people in their mid-40s (the sample was split at age 40 years). Few of us routinely notice any major cognitive changes across this age span although they are present and of approximately similar magnitude to the difference found between hypertensives and controls. We and others (14) had also tried a different tack—examining the influence of antihypertensive medications on these mild impairments. Our results (14, 15) were typical: although some cognitive functions improved with the reduction of blood pressure, others showed slight relative impairments associated with treatment. Furthermore, different medications yielded differing cognitive patterns.
|
DEVELOPMENT OF VASCULAR HYPOTHESIS
The establishment of a neuroimaging capability at the University of Pittsburgh School of Medicine provided us with the ability to proceed using a tool that had the potential of rather directly investigating the physiological basis of the neuropsychological deficit in hypertensives.
We first, however, developed a hypothesis based on prior work on cerebral blood flow and hypertension. Most brain imaging studies examine blood flow-related changes in the brain. The earliest studies related to blood pressure and brain blood flow were done in the late 1800s. At this time, a common opinion on the possibility of vascular adjustments within the brain was based on gross anatomy. The brain is the only organ totally encased in a nonexpandable lining—the skull. Given this, physiologists of the time thought that blood pressure changes within the brain would not be possible. Working with smoked drum kymographs and sensitive springs during this pre-electrical amplifier age, Roy and Sherrington (17) injected a vasoactive brain extract into the brain and measured its expansion as well as arterial and venous pressure. The brain expanded proving that its volume was not fixed by the skull. Bayliss and Hill (cited in 17) did a similar experiment, but measured a cerebral vein as well as systemic and cardiac pressure. They observed parallel change in brain and systemic pressures. These pioneering physiologists disproved the theory that the skull prevented vascular change in the brain, but they missed another interesting point. By concluding that the brain and systemic circulation were similar, they threw the "baby out with the bath water."
More recent investigations have established the concept of brain autoregulation of blood flow. Sensitive measures of cerebral artery diameter permitted Fog (18) to observe decreases in arterial diameter when pressure in the artery was increased by stimulation of the splanchnic nerve, carotid sinus stimulation, or injection of epinephrine. Note the paradoxical result—an increase in pressure in a potentially expandable tube is associated with a compression of the tube. By 1966 investigators such as Harper (19) and Reinmuth, Beteta, and Scheinberg (20) had established that cerebral blood flow (and presumably vessel diameter) increased rather linearly with blood pressure as mean arterial pressure increased from 0 to about 80 mm Hg. As mean arterial pressure increased beyond 80 mm Hg, however, flow was maintained at a steady level despite increasing pressure. These observations formed the basis of the current concept of pressure-flow regulation in the brain. In the range of normal human blood pressure, the brain appears to autoregulate, that is maintain a relatively constant blood flow despite variation in peripheral blood pressure. Chillon and Baumbach (21) diagrammed the concept elegantly and Figure 2 shows a redrawing of their diagram (see also 22). Note the role of vessel diameter. The autoregulatory plateau is maintained in the lower ranges of arterial pressure by a relatively increased vessel diameter. In the higher ranges of pressure, the plateau is maintained by relatively decreased vessel diameter, ie, constricted blood vessels. Hypertensive patients operate in the higher range of blood pressure. Thus, they must maintain constricted blood vessels to regulate flow. Indeed, hypertensives typically have relatively normal cerebral blood flow, albeit their blood flow is slightly less than that of comparable normotensives (a small, but typically statistically significant, difference) (23–30). Autoregulation is also altered in hypertensives—the plateau starts and ends at greater values along the blood pressure axis. Among hypertensives relative to normotensives, brain hypoperfusion is more likely to result from equivalent transient systemic hypotension.
|
An additional factor is a reduction of responsivity to dilative effects, as shown by damped responses to nitric oxide, acetazolamide, and carbon dioxide in hypertensives (25, 35). Patients with hypertension relative to age-matched controls respond, however, with greater vasoconstriction to hypocapnia (36). Dilative changes, not constrictive changes, are impaired in hypertensives (37–39). In short, a number of factors suggest that the cerebral vessels should be less capable of enhancing regional cerebral blood flow, rCBF, among hypertensives relative to normotensive age-matched controls (28, 39).
The observations just reviewed form the basis for our vascular hypothesis of cognitive changes with hypertension. They must first be combined with the observation that heightened neural activity in the brain creates a metabolic need that in turn engenders enhanced blood flow to the active tissue—the basis for most current neuroimaging (40). In our hypothesis, hypertension is first assumed to initiate a shift to autoregulation of cerebral blood flood at higher pressures. This shift maintains cerebral blood flow and is accomplished through vessel remodeling, but cerebral blood flow is presumed to be at a slightly lower level than normal. The decreased flow and reduced capability to vasodilate (due to the remodeling) lead to less rCBF increase to active neural areas. This reduction in activated flow in turn induces compensatory increases in flow in homologous regions, eg, expansion of areas of activation, bilateral activation of structures rather than unilateral activation. This compensation is presumed to be largely effective except when sustained neural activation is required, such as for tasks requiring the maintenance and transformation of items in working memory. Note that such compensatory changes are not presumed to be specific to hypertension because the literature suggests similar changes in both pathological conditions, eg, Alzheimer’s disease, as well as normal aging (41–43). Reduced rCBF and compensatory changes among hypertensives should be evident in a substantial amount of brain tissue (which should be evident in imaging studies). Cognitive deficits may be harder to detect. The success of compensation is consistent with the mild and variable pattern of deficits observed in hypertensive individuals. We might also expect that vulnerable physiological areas might demonstrate more striking relative reductions in flow during activation. The watershed areas are such vulnerable areas. These are areas that are supplied by the most distal portions of the cerebral arteries—distal portions of the distributions of the anterior and middle cerebral arteries and the middle and posterior cerebral arteries (29, 44) . These areas might be expected to be the first to show reductions in activated flow due to hypertension.
INITIAL STUDY OF rCBF IN HYPERTENSIVES AND CONTROLS
Armed with this vascular hypothesis, we completed an initial study (45) looking at nine hypertensive patients and five controls of comparable age (59–68 years old, right-handed except for one patient). Hypertensive patients were unmedicated for a minimum of 8 weeks and had less than a 2-year lifetime history of taking antihypertensive drugs. We decided to focus on the memory performance implicated in the review by Waldstein et al. (12) using PET with a radiotracer with a short half-life. The short half-life permits multiple injections/brain scans in a session allowing us to examine different levels of cognitive challenge. The radiotracer [15O] water has a 2-minute half-life and permitted scans separated by 8 minutes. Rest was compared with activation at each of two levels of difficulty of two tasks. A continuous performance task required subjects to monitor the screen for either a single target letter (level 1) or for the repetition of a letter one back, ie, "b" is a target following "b a" but not following "a b" (level 2). A free-recall task required subjects to recall a word list presented at a rate of 1 per second followed by a 15-second recall period. Level 1 of this task was trivial—a list length one—and level 2 rather challenging—a list length of 12.
The results of this study both supported our vascular hypothesis and indicated other possible influences of hypertension on rCBF. Hypertensives and controls showed rCBF changes to the memory tasks that were very consistent with changes observed in similar tasks in younger, college-aged volunteers (46). Robust changes were observed bilaterally in prefrontal cortex (Brodmann’s areas 9,44,46), including the dorsolateral prefrontal cortex, which is implicated particularly in working memory, as well as posterior parietal cortex (Brodmann’s areas 39,40). Activations were somewhat more significant in the continuous performance task relative to the verbal recall task. Our vascular theory was supported in that comparisons directly between hypertensives and controls showed that prefrontal and parietal increases in rCBF with increasing memory load were large for controls but minimal for hypertensives. There were two surprises. First, significant differences were confined to right hemisphere rCBF changes—controls performed the memory task with enhanced right hemisphere changes whereas hypertensives showed rCBF changes that appeared more bilateral (left hemisphere activation was marginally greater in hypertensives relative to controls). Second, hypertensives showed greater rCBF changes with increasing memory load than controls in an area that included a portion of the hippocampus.
These results encouraged us to continue the study and we were fortunate to obtain National Institutes of Health funding to do so. The initial study was clearly limited. First, hypertension is a complex disease and the sample size did not address the question adequately. Second, we had only obtained relative measures of rCBF. Our hypothesis was one of compensation for a vascular deficit; thus, quantitative measures of blood flow would be desirable. Third, we had not assessed plausible co-occurring conditions that might modify our interpretations, eg, the presence of atherosclerosis or minor lesions indicated on MRI. The initial study also suggested the value of the continuous performance task as a probe and provided us with regions of interest to test prospectively in our planned study. We had not found any clear deficits related to watershed areas in the preliminary work, but we also wanted to designate these areas as regions of interest in the planned work.
HYPERPET
The project, which we developed, was given the title of Hyper for hypertension and PET for positron emission tomography. Our goal was to examine a substantial sample of hypertensives and age-matched controls to verify our preliminary work and test our vascular hypothesis. Based on the preliminary work, we used a continuous performance task but varied whether the task focused on the spatial or verbal content of the displays. We also collected the data using both an arterial and venous catheter so that arterial blood could be sampled to estimate the tracer amount presented to the brain. Application of blood flow models (47, 48) for PET studies derived from the basic Fick principle could then be applied to estimate quantitative blood flow (49). We also added an extensive neuropsychological battery, measures of an index of peripheral atherosclerosis (Doppler ultrasound assessment of the carotid artery wall), and MRI scans optimized to detect white matter hyperintensities [assumed to likely arise from minor white matter lesions (50)]. Basic medical history and psychometric data were also collected.
Structural MR imaging was performed as well as ultrasound examination of the carotid artery before the PET examination. The MR images were used in the process of transforming all images of PET activation to a common neuroanatomical set of coordinates, to define regions of interest for extraction of quantitative blood flow measures, and to assess white matter hyperintensities. Trained radiologists examined the MR images and performed overall ratings of degree of white matter hyperintensity as well as specific ratings by brain area. The size of sulci and ventricles was also rated as global index of structural brain integrity. Dr. Sutton-Tyrell of our graduate school of public health supervised the ultrasound assessments of the carotid artery used as indices of atherosclerosis. Each subject had a sequence of segments of the external carotid artery, the carotid bulb, and the initial internal carotid artery assessed. The thickness of the intima and media of the artery was assessed at each segment. These data were summarized into mean intima-medial thickness and plaque occurrence and severity were also derived from the data.
A subset of the obtained sample characteristics is presented in Tables 1 to 3. The hypertensives were again unmedicated with less than a 6-month lifetime history of any antihypertensive medication. As shown in Table 1 hypertensive and control groups were quite similar in their basic characteristics. Despite the relatively small size of the differences, hypertensives were significantly older and drank more per week than controls. Analyses had to control for these factors statistically. Table 2 shows that overall, the volunteers were not depressed and were in reasonably good moods during our testing. Hypertensives actually reported being in a more positive mood than normotensives during the testing. Table 3 shows the measures from the carotid artery wall and the MR examination of white matter lesions and brain structures. Overall, the indices suggest reasonably healthy arteries and brain structures, but the expected relative increase in carotid wall thickness and incidence of white matter hypertensities is present among the hypertensives (50).
|
|
|
The tasks lasted 5 minutes, commencing slightly before tracer injection so that assessment of radiation occurred during the middle 2 minutes of task performance. Eight minutes separated task presentations. The radiotracer, 11 mCi of 15O water, was injected through a venous catheter for each of 12 scans. Arterial blood was sampled through a radial artery catheter on 8 of the 12 scans—the verbal and spatial static memory conditions were excluded to avoid excessive blood withdrawal from the subject.
INITIAL NEUROPSYCHOLOGICAL AND QUANTITATIVE PET RESULTS
Initial analyses have been completed on our HyperPET project, but it is a complex data set. We are not yet satisfied that we have caught all the errors of data entry or curve fitting, and we certainly have not completed all the analyses that we have planned. I doubt that the primary features of our results will change as we finish analyses, but details may shift and our interpretation may be altered somewhat. I hope that the primary changes will be in providing more convincing empirical evidence for our current interpretation. In any event, the best course is to consider our current results as primarily illustrative of the type of information relevant to psychosomatic medicine available with brain imaging techniques.
Our neuropsychological results supported our choice to focus on working memory. Subjects responded to a 4-hour battery of neuropsychological tests that emphasized mnemonic skills, executive function, and attention, but included basic tests of perceptual-motor function as well. Significant differences between hypertensive and control subjects were only found, however, for tests of short-term/working memory. Figure 3 shows the short-term memory results from the Wechsler Memory Scale (52). Hypertensives remembered slightly fewer items than controls at all of the retention intervals. Only the tests involving working memory showed any discrimination between hypertensives and controls. This result is heartening as it suggests the appropriate choice of tasks for the neuroimaging examination and it also supports the conclusion from the Waldstein et al. review (12).
|
The quantitative results were supportive of our vascular hypothesis. First, these patients, like prior ones in the literature (23), did show significantly reduced overall CBF relative to controls. During our checkerboard task, which was designed to produce a consistent activation in occipital and motor cortex, blood flow in the normotensives averaged 0.44 ml/min per ml of tissue whereas hypertensives averaged 0.41 ml/min per ml. This small difference was consistent across our analysis of all the tasks. Figure 4 shows more specific support for our vascular hypothesis. For the posterior parietal ROI, controls showed a brisk activation during the spatial 2-back working memory (WM) task relative to our no-load control task, but hypertensives showed little change. The verbal 2-back task yielded virtually the same result albeit with slightly less statistical significance. Comparable trends were shown for the prefrontal ROI for both spatial and verbal tasks, but these results were statistically marginal.
|
|
RELATIVE rCBF (SPM) RESULTS
The quantitative results were quite supportive of our vascular hypothesis. They both showed decreased overall CBF, less activated rCBF, and a possibly compensatory rCBF increase in the amygdala/hippocampus among hypertensives performing well. In this context, our results were surprising from the whole-brain analyses of relative rCBF using SPM analysis programs. These analyses standardize total CBF flow for each scan to 50 and then assess relative rCBF for each 2 x 2 x 2 mm area (voxel) within the brain (containing about 145,000 voxels). The data are smoothed, however, to facilitate fitting all subjects/scans to a common template and to reduce inhomogeneities in the data. This reduces the actual resolution of an area to closer to 7 x 7 x 7 mm (although the nominal voxel size remains 2 x 2 x 2). The analyses were done by first assessing the rCBF for each individual that approximated the increasing vector of working memory load defined as -3 for the control/no load condition, 1 for the static 3-item memory load, and 2 for the dynamic 3-item 2-back task. This was done separately for spatial and verbal working memory. The results of these individual analyses were then combined in a random-effects analysis that compared the rCBF areas across subjects first within their groups and then between the normotensive and hypertensive groups.
Hypertensive and normotensives most clearly differed in their response to increasing spatial working memory in these SPM analyses. Both groups showed clear posterior parietal and prefrontal rCBF increases, replicating our initial work and again confirming this pattern of activation seen previously in younger samples (45, 46). A visual comparison between the two groups of the rCBF changes during both spatial and verbal working memory suggested substantial similarity in the areas showing increase, but a greater extent of significantly increased rCBF among hypertensives. Figure 6 illustrates this for spatial working memory. Direct statistical comparison of the groups failed to show any significant differences for verbal working memory, but did show differences for spatial working memory. Hypertensives relative to normotensives showed greater rCBF increases with increasing spatial memory load in the left prefrontal cortex. The cluster of significant voxels encompassed portions of the insula and Broca’s area—Brodmann’s areas 1,13,44, and 46 (Talaraich coordinates of the most significant voxel were –42 12 18, t value = 5.0). Figure 7 shows the average rCBF change within a 5-mm sphere surrounding this significant voxel. The greater rCBF change among the hypertensives is clear.
|
|
In the first approach, we tried to map our quantitative ROIs into an SPM analysis. Specifically, we sampled similar brain areas between the quantitative and SPM analyses by placing a 1-cm sphere around the approximate center of the regions of interest drawn for the quantitative analysis. Regions of interest (prefrontal, posterior parietal, and amygdala/hippocampal) were then analyzed within SPM. We failed to find interpretable differences for the prefrontal and posterior parietal regions, but did find an effect within the amygdala/hippocampal region. For both verbal and spatial working memory, rCBF increases in this region were significantly greater for hypertensive subjects. The result is consistent with our initial study as well as with our quantitative results. We anticipate that precise specification of areas between quantitative and SPM analyses may yet yield more consistency between analyses.
The second approach seems more promising and is consistent with our vascular hypothesis. The SPM analysis adjusts each individual’s rCBF voxel-by-voxel data by the estimate of global flow obtained from the total radiation count. Each condition/scan in the experiment is adjusted for the relationship across all the scans between global flow and rCBF at that voxel. To be consistent with this SPM procedure, we returned to our quantitative data for the parietal ROI and covaried the quantitative estimates of global flow obtained during the control and the dynamic memory task. In the presence of the global flow covariate, the F values for the group by task activation factor were reduced in half. More importantly, the covariate was a significant contributor to the variance; and, plotting the covariate, global flow, showed that control individuals increased their global flow with memory load, but that hypertensives decreased their global flow. A direct test of the difference between groups in response to the memory demand was suggestive, but not significant (p = .12). The implications for the SPM analysis are important, however. The voxel rCBF values for memory activation among the controls are likely to be scored as a difference from an increase in global CBF; whereas the voxel rCBF values for memory activation for the hypertensives are likely to be scored as a difference from a decrease in global CBF. In other words, the difference in the response of global flow between groups will be removed in SPM and this will result in a greater opportunity to find relative increases above global flow changes among the hypertensives. If we find direct support for this assessment, it will mean that the quantitative analyses are showing a relatively general decrease in CBF with cognitive performance among hypertensives while the SPM analyses are showing the specific brain areas in which increases relative to their (decreasing) CBF occur.
TENTATIVE INTERPRETATION
A tentative interpretation of the current results guides us in our additional analyses of these results. First, we maintain our vascular hypothesis. The quantitative results suggest that the brains of hypertensives have adapted to maintain a normal, but slightly reduced, CBF. This in turn seems to induce less rCBF increase in response to large regions of parietal and prefrontal cortex during memory performance. The quantitative results support this conjecture for parietal cortex in particular. The spread of rCBF activation in parietal and prefrontal cortex observed in the SPM results may also be consistent with a spread in voxels with relatively heightened rCBF countering the CBF decrease inferred from the quantitative results. That is, the hypothetical deficit in activated blood flow is compensated for by activation of regions that are not used as extensively by normotensives to process spatial and (less well supported empirically) verbal working memory tasks. Most specifically, from both SPM and quantitative analyses, hypertensives appear to show relatively enhanced blood flow in the amygdala/hippocampal area relative to comparably performing normotensives. From the SPM analysis, hypertensives show enhanced rCBF in a left prefrontal area encompassing Broca’s area and underlying insula. The involvement of Broca’s area in the spatial task is particularly interesting. Broca’s area is defined by control of articulatory speech. We might hypothesize that hypertensives adopt a verbal strategy to keep track of the spatial stimuli in compensation for reduced activation in areas used by normotensives to perform this task. Thus, we are able to continue to maintain a vascular hypothesis that incorporates compensatory adjustments. At this point, our statistical support for this working hypothesis is present, but spotty. Furthermore, our vascular hypothesis is not specific about brain regions influenced—the observed effects do not seem as general as might be expected from the hypothesis; and problems in the watershed regions, which were not found, would be expected from the hypothesis.
Our observations and maintenance of the vascular hypothesis is supported somewhat by recent brain imaging findings focusing on aging and the effects of disease other than hypertension. Psychology and Aging recently published a special section on aging, cognition, and neuroimaging. Three of these articles emphasized compensatory rCBF changes with aging that typically involved an increase in the extent of areas activated (41, 42, 46) . Becker et al. (43) made similar arguments in their interpretation of activation changes in early stages of Alzheimer’s disease. The ubiquity of such compensatory adjustments to challenges to brain function is interesting, but clearly requires substantially greater specification. Our colleagues assessing recovery from brain injury have long been concerned with compensation concepts (53, 54). Their reviews separate types of compensation, those a) based on task strategy—such as the verbalization strategy in hypertensives just described; b) based on the use of nonspecialized or idle brain tissue—perhaps, related to the growth in area of activation and subcortical activation in hypertensives; c) based on structural growth; and d) based on functional changes in neurotransmitter or other mechanism. Although we are not likely to investigate functional compensations as in d) soon, we will have to examine more closely anatomical changes such as those that may either contribute to hypertensive deficits or play a role in compensation (55).
We also need to be aware of individual differences among hypertensives (and normotensives) that may modulate our findings. To date, our search for important individual difference dimensions has been unsuccessful, but we have only scratched the surface of possible biological and psychological modulators. For example, a tantalizing possibility is suggested by our preliminary results. We observed rCBF increases in an ROI that included the amygdala, an area that has been widely related to affective processing (56). Furthermore, interactions between amygdala and prefrontal cortex, which was also implicated in our results, have similarly been widely discussed in the context of the regulation of affect (57, 58). Indeed, our interpretation of compensatory activity in the hippocampus/amygdala might alternatively be considered the result of changes in inhibitory/excitatory effects on this area from the prefrontal cortex. These possibilities are further engaged by our finding that hypertensives reported greater positive affect during performance than normotensives on a self-report scale. This combination of findings suggests that individual differences in affect regulation might relate to the pattern of results observed and relate to the well-known finding that hypertensives report positive affect in situations in which normotensives do the opposite (5). Neither the resolution of our data nor our design are optimal for exploring these possibilities, but they do suggest the value of examining affect regulation and brain function among hypertensives.
CONCLUSION
I have presented a work in progress, therefore, firm conclusions on our research project cannot be given. Stepping back from the project, however, permits some general conclusions. First, after more than 60 years, hypertension remains a psychosomatic disease and deserves our continued attention. Second, our research has shown that hypertension influences brain blood flow early in the natural history of the disease. Clinically, initial treatment of hypertension should consider the relative efficacy of treatments in normalizing brain blood flow as well as systemic pressure. In HyperPET, we are trying to define the influence of hypertension on cognitive information processing. The brain is not confined to cognitive information processing, however. Earlier psychosomatic work focused on the affective regulation of hypertensives (1); recent psychosomatic work has focused on prospective and concurrent differences in cardiovascular reactivity between hypertensives and normotensives (59–62). Given that hypertension alters brain blood flow, it seems reasonable to expect that functional consequences may extend to affective and, perhaps, autonomic nervous system regulation. Indeed, our vascular hypothesis of the cognitive deficit in hypertensives may relate interestingly to other vascular hypotheses, such as that suggesting a vascular basis for late-onset depressive disorder (63). Imaging technology provides a window on the organ system primarily responsible for thoughts, actions, and affect. Psychosomatic medicine relates these thoughts, actions, and affects to the development and maintenance of health and disease. Psychosomatic examination of the brain need no longer be limited to interviews and questionnaires.
ACKNOWLEDGMENTS
I gratefully acknowledge the collaborators whose work I describe: Drs. Stephen Manuck, Carolyn Meltzer, Matthew Muldoon, Julie Price, Christopher Ryan, Kimberly Sutton-Tyrrell, and Frederik van der Veen. I also thank for editorial and technical assistance Mary Assenat, Peter Gianaros, Michael Eddy, Judy Thompson, and David Townsend. This work was supported by National Heart, Lung, and Blood Institute Grants HL57529 and HL40962.
NOTES
This article is based on the Presidential Address given at the Annual Meeting of the American Psychosomatic Society, Barcelona, Spain in March 2002.
Received for publication July 8, 2002.
REFERENCES
This article has been cited by other articles:
 |
|
 |
|
  R. D. Lane, S. R. Waldstein, H. D. Critchley, S. W. G. Derbyshire, D. A. Drossman, T. D. Wager, N. Schneiderman, M. A. Chesney, J. R. Jennings, W. R. Lovallo, et al. The Rebirth of Neuroscience in Psychosomatic Medicine, Part II: Clinical Applications and Implications for Research Psychosom Med, February 1, 2009; 71(2): 135 - 151. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. R. Jennings, M. F. Muldoon, C. Ryan, J. C. Price, P. Greer, K. Sutton-Tyrrell, F. M. van der Veen, and C. C. Meltzer Reduced cerebral blood flow response and compensation among patients with untreated hypertension Neurology, April 26, 2005; 64(8): 1358 - 1365. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. A. Suhr, J. C. Stewart, and C. R. France The Relationship Between Blood Pressure and Cognitive Performance in the Third National Health and Nutrition Examination Survey (NHANES III) Psychosom Med, May 1, 2004; 66(3): 291 - 297. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
--) and control (--
--) groups.
) and controls (
).
