| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the Department of Psychiatry, University of Pittsburgh School of Medicine, Pittsburgh, PA.
Address correspondence and reprint requests to Peter J. Gianaros, Cardiovascular Behavioral Medicine Program, University of Pittsburgh, 3811 O'Hara Street, Pittsburgh, PA 15213. E-mail: gianarospj{at}upmc.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Methods: Sixteen participants (7 men, 9 women aged 47 to 72 years) were classified as high (n = 8) or low (n = 8) blood pressure reactors by the magnitude and temporal consistency of their systolic blood pressure (SBP) reaction to a Stroop color-word interference stressor. Both high and low SBP reactors completed this Stroop stressor while their task-related changes in blood pressure and functional neural activity were assessed in a blocked functional magnetic resonance imaging design.
Results: In both high and low SBP reactors, the Stroop-stressor engaged the anterior cingulate, orbitofrontal, insular, posterior parietal, and the dorsolateral prefrontal regions of the cortex, the thalamus, and the cerebellum. Compared with low reactors, however, high reactors not only showed a larger magnitude increase in SBP to the Stroop stressor, but also an increased activation of the posterior cingulate cortex.
Conclusion: A behavioral stressor that is used widely in cardiovascular reactivity research, the Stroop stressor, engages brain systems that are thought to support both stressor processing and cardiovascular reactivity. Increased activation of the posterior cingulate, a brain region implicated in vigilance to the environment and evaluative emotional processes, may be a functional neural correlate of an individual's tendency to show large-magnitude (exaggerated) blood pressure reactions to behavioral stressors.
Key Words: cardiovascular reactivity • functional magnetic resonance imaging • stress
Abbreviations: BMI = body mass index; BOLD = blood oxygen level dependent; DBP = diastolic blood pressure; fMRI = functional magnetic resonance imaging; MANOVA = multivariate analysis of variance; RF = radiofrequency; ROI = region of interest; SBP = systolic blood pressure; TE = time to echo; TR = time to repetition; VAS = visual analog scale.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Consistent results from functional neuroimaging studies indicate that cardiovascular reactions to behavioral stressors—such as demanding mental arithmetic, working memory, and Stroop color-word interference tasks—are associated with the concurrent activation of three cortical brain systems: the orbitofrontal cortex, the anterior cingulate cortex, and the insular cortex (10–14). As supported by lesion, stimulation, and neuroanatomical tracing studies, subdivisions within these cortical brain systems are thought to initiate and represent cardiovascular reactions to behavioral stressors by their reciprocal circuitry and by their projections to cell groups that regulate autonomic and cardiovascular function, such as the amygdala, hypothalamus, and brainstem nuclei (16–26). In addition to supporting cardiovascular reactivity, the orbitofrontal, anterior cingulate, and insular brain systems are also thought to support functional neural processes that are involved in detecting, appraising, and adaptively responding to behavioral challenges (13,15,16,17,21,27–31). Thus, because the orbitofrontal, anterior cingulate, and insular regions may centrally orchestrate evaluative processes, adaptive behavioral responses, and cardiovascular reactions, it has been hypothesized that individuals who show exaggerated cardiovascular reactivity to a behavioral stressor may also show increased stressor-induced functional neural activation in these brain systems (15,32). To date, however, no study had tested this hypothesis using functional neuroimaging methods.
The present study accordingly tested whether individuals who express a consistent pattern of exaggerated cardiovascular reactivity would also show greater functional neural activation in the orbitofrontal cortex, the cingulate cortex, or in the insular cortex than those individuals who express a contrasting pattern of attenuated cardiovascular reactivity to a behavioral stressor. To test this hypothesis, we examined functional neural activity by functional magnetic resonance imaging (fMRI) and cardiovascular reactivity by blood pressure monitoring in a sample of high and low cardiovascular reactive individuals while they completed a Stroop color-word interference task. High and low reactive individuals were selected from a prior cardiovascular reactivity study (33) in which they completed this Stroop task, and they were categorized by the magnitude and consistency of their systolic blood pressure (SBP) reaction to this task. Because it is a demanding cognitive–behavioral challenge that is known to elicit cardiovascular reactivity, we expected that the Stroop task would engage the orbitofrontal, cingulate, and insular regions. By engaging these brain regions of interest and by eliciting cardiovascular reactivity with the Stroop task, we further reasoned that we would be able to test for differences between high and low cardiovascular, specifically SBP, reactive individuals in their patterns of functional neural activation.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Procedure
Before testing, participants fasted and abstained from caffeine, tobacco products, and exercise for 3 hours, and refrained from drinking alcohol and taking nonessential medication for 12 hours. On arrival at the magnetic resonance (MR) research center, participants were provided with a description of the experiment and were given a tour of the MR testing suite. In a control room adjacent to the MR scanner, participants practiced the Stroop color-word task (described next). Participants were then instrumented for noninvasive blood pressure monitoring and were placed in the bore of the MR scanner. After acquiring structural MR images (for approximately 15 minutes), participants completed the Stroop task (for approximately 45 minutes). After the experiment, participants were removed from the scanner, debriefed, and paid ($40.00 US) for participating.
Stroop Color-Word Interference Task
Participants completed a modified version of the Stroop color-word interference task that was administered in REACT and that has been used in prior studies of cardiovascular reactivity and disease risk (eg, 34–36). In the modified version of the Stroop task implemented here, participants were presented with an array of color words (eg, red, blue, yellow, green) on a visual display. In the array, one target word was presented in the center of the display and four identifier words were presented along the bottom. The task of the participant was to identify as quickly and accurately as possible the color in which the target word was shown by selecting the appropriate identifier word. With their right hand, participants made their identifier selections on a four-button response box located in the MR scanner. The selected identifier word was surrounded with a box on the display after each response. Participants were given performance feedback on each trial such that the correct identifier was highlighted after each response or after a variable time delay (see description of the Stroop task next).
Each array of color words was presented as a single trial and held characteristics that defined each trial into two separate conditions. For all of the trials of one condition (the congruent condition), the color of the target word and the identifier words were congruent with the color in which the target word appeared (eg, the target word "blue" appeared in the color blue and all identifier words were shown in blue). For all of the trials of the other condition (the incongruent condition), the color of the target word and the identifier words were incongruent with the color in which the target word appeared (eg, the target word "blue" appeared in the color red and all identifier words were shown in colors that differed from red). To further increase the demand characteristics of the incongruent condition, the participant's performance (accuracy at target word color identification) was titrated to 
60% using the participant's speed and accuracy: faster and more accurate performance in a given incongruent condition resulted in faster trial presentation times (ie, shorter time delays before the correct identifier was highlighted and the next trial was presented). The total number of trials presented in a given congruent condition was matched (yoked) to the total number of trials that were completed in the preceding incongruent condition; this yoking procedure was used to control for the possible confounding effects of motor response differences between conditions. That is, if a participant completed 22 trials in a given incongruent condition, then 22 trials were presented (distributed evenly with fixed trial presentation times) over the course of the subsequent congruent condition. In order to correctly yoke the number of trials between blocks, an incongruent block was always presented before a congruent block. Thus, congruent and incongruent trial types were presented in a blocked and alternating order, such that a 90-second block of congruent trials always followed a 90-second block of incongruent trials.
This blocked fMRI-design adaptation of the Stroop color-word interference task differs from the version that was implemented in REACT and other cardiovascular reactivity studies. In those studies, only one approximately 9-minute block of performance-titrated incongruent trials was presented. Here, participants completed 16 alternating experimental blocks of 8 incongruent and 8 congruent conditions. This design adaptation facilitated the measurement of the hemodynamic fMRI blood oxygen level dependent (BOLD) response and allowed sufficient time for one oscillometric blood pressure measurement per block. In this way, we obtained eight concurrent estimates of the fMRI BOLD response and blood pressure during a behavioral stressor (the incongruent condition) and eight comparable estimates during a less-demanding comparison condition, which had similar stimulus and response characteristics (the congruent condition).
The experimental blocks were administered over a set of four functional imaging runs, such that four experimental blocks (two incongruent, two congruent) were administered for each 6-minute functional run. Between each functional run, participants were given a short rest lasting approximately 1 minute.
Assessment of Regional Brain Activation
All structural and functional images were obtained using a 3-Tesla Signa scanner (General Electric Medical Systems, Milwaukee, WI) and a standard quadrature radiofrequency (RF) head coil. Over the course of each 90-second block, a reverse-spiral pulse sequence (TR = 1,500 ms; TE = 25 ms; 60° flip angle) was used to obtain 34 axial T2*-weighted contiguous functional images, which were parallel to the plane of the anterior and posterior commissures (3.2-mm slice thickness; 20-cm field of view). With this acquisition sequence, we obtained a whole-brain volume of functional images every 1,500 ms, yielding 60 such whole-brain functional images per block. By averaging these 60 images, we derived a single whole-brain functional image for each congruent and incongruent block. This image depicted the average voxel-by-voxel intensity of the BOLD signal for each 90-second congruent and incongruent block, and it served as the principle fMRI-dependent measure. Before presenting the experimental blocks, a T1-weighted axial structural image was obtained from each participant for later co-registration of the functional images.
All images were reconstructed from k-space with the locally developed NeuroImaging Software suite (NIS; http://kraepelin.wpic.pitt.edu/nis/) and were corrected for motion artifacts with Automated Image Registration (AIR; 37). A between-run baseline correction and within-run linear detrending algorithm and high-pass filter (0.003 Hz) were applied the BOLD data to correct for nuances of the scanning environment, such as slow frequency and linear drifts in BOLD signal intensity that are not attributable to the experimental design. A voxel-by-voxel outlier correction algorithm identified voxel values that exceeded four standard deviations (SD) from its overall mean over time and truncated any identified outliers to the four SD value.
After image reconstruction, detrending, and outlier correction, images were co-registered (with a 60-parameter AIR algorithm) to the standard reference brain of the Montreal Neurological Institute (MNI; ftp://ftp.mrc-cbu.cam.ac.uk/pub/imaging/Colin), which was resampled to match the spatial resolution of the functional images. Minor anatomical differences between subjects that remained after co-registration were corrected by smoothing over all voxels in each functional image with a three-dimensional Gaussian filter (6 mm full-width half-max).
Assessment of Subjective Distress and Blood Pressure Reactivity
Before being placed in the bore of the MR scanner, the participant practiced one-to-three blocks of the congruent and incongruent conditions. Immediately after practice and outside of the scanner, the participant then used a 100-point visual analog scale (VAS) (anchored by 0 = not at all and 100 = extremely) to rate how stressful and demanding she or he found the two conditions.
Once inside of the scanner (and at the end of the 15-minute time period in which structural brain images were obtained), an Omega 1400 MR-compatible blood pressure monitor (InVivo Research, Orlando, FL) was used to obtain three oscillometric measures of the participant's SBP and DBP in 2-minute intervals. These three measures were averaged to derive a baseline blood pressure. Also, one blood pressure measurement was obtained during each of the sixteen 90-second blocks, and these blood pressure measures were averaged separately over the eight incongruent and eight congruent blocks to yield one average blood pressure per condition. The average blood pressure obtained during the prefunctional scanning baseline was used as a comparison for the average congruent and incongruent condition-related blood pressures.
Definition of High and Low Blood Pressure Reactors
High and low cardiovascular reactors were identified initially by the magnitude of their SBP reaction to the Stroop color-word interference task that was administered in REACT. In particular, from the upper tertile of the REACT Stroop-task SBP reaction distribution, we recruited and tested 10 individuals who showed an increase in SBP from a resting baseline to the Stroop task that exceeded 19.62 mm Hg; from the lower tertile, we recruited and tested 10 individuals who showed a change in SBP that fell below 5.56 mm Hg.
To be classified as a temporally consistent high or low SBP reactor, individuals had to meet the additional criterion of showing a comparable-magnitude SBP reaction to the incongruent condition of the Stroop task in the present study. Specifically, participants had to show an average change in SBP from the prescanning baseline period to the incongruent condition of the present study that exceeded (high reactors) or fell below (low reactors) the median SBP reaction of all participants (median 
SBP = 11.06 mm Hg). In addition to the data that were excluded from 2 participants with excessive fMRI artifacts, this classification procedure resulted in the exclusion of data from 2 additional participants who showed mixed SBP reactions: 1 showed a SBP increase in REACT of 17.25 mm Hg and then a decrease of 0.50 mm Hg in the present study; the other showed a SBP decrease of 3.27 mm Hg in REACT and then an increase of 13.81 mm Hg in the present study. It is important to note that this classification criterion was used because the aim of the present study was to examine individuals who show a pattern of cardiovascular reactivity that is consistent over time (reliable).
The remaining 8 high reactors showed an average increase in SBP to the Stroop task in REACT of 26.04 (±SE 1.93) mm Hg; the 8 low reactors showed an average increase of 1.22 (±1.33) mm Hg. These two groups did not differ significantly in their DBP reactions to the REACT Stroop task (high reactors M DBP = 6.48 ± 3.70 mm Hg; low reactors M DBP = 3.08 ± 1.12 mm Hg, t < 1). High and low reactors also did not differ in their age (M 66.37 ± 0.94 vs. 65.3 ± 1.46 years), gender (4M, 4F vs. 3M, 5F), handedness (1 self-reported left-handed individual in each group), or baseline SBP (121.56 ± 3.45 vs. 123.13 ± 4.65 mm Hg) or DBP (75.56 ± 2.70 vs. 77.31 ± 2.62 mm Hg) blood pressure at the time of testing, all p values >0.50.
Data Analyses
Assessment of Task Performance, Subjective Ratings, and Blood Pressure Reactivity
Average incongruent and congruent condition-related accuracy (M % of correct trials) and VAS ratings of stress and demand were analyzed with 2 (Group: high, low reactor) x 2 (Condition: congruent, incongruent) multivariate analyses of variance (MANOVAs). Condition-related changes in SBP and DBP were also analyzed with 2 (Group: high, low reactor) x 3 (Condition: prescanning baseline, congruent, incongruent) MANOVAs. Group by Condition interactions were followed with simple effects tests and pairwise comparisons between conditions for each group.
Assessment of Regional Brain Activation
Condition-related brain activation (indexed by BOLD signal intensity or activation) was assessed using voxel-by-voxel 2 (Group: high, low reactor) x 2 (Condition: congruent, incongruent) random-effects ANOVAs. In these ANOVAs, Group and Condition served as fixed explanatory factors and Participant served as a random explanatory factor. The effects of Group, Condition, and Group by Condition interactions on BOLD activation were investigated using both a priori and exploratory analytic strategies. A priori analyses involved restricting consideration of significant effects to those brain regions of interest (ROIs) that have been implicated in stressor processing and cardiovascular regulation by prior research (10–14). These ROIs included a) the bilateral orbitofrontal cortex, b) the bilateral anterior cingulate, c) the right insula, and d) the left insula. Anatomical boundaries for these ROIs were defined using the Automated Anatomical Labeling masks for the MNI brain (38), which are available from the first author. In addition to these ROI analyses, an exploratory random-effects strategy was also used to assess voxel-by-voxel BOLD activation across the whole-brain volume.
To control the Type I error rate in both a priori and exploratory analyses, activated voxels were required to a) meet a corrected threshold of p < .005 and to b) occur within a three-dimensional cluster of contiguous voxels meeting the same threshold. For a priori and exploratory analyses, the number of contiguous voxels comprising a cluster was defined using small-volume contiguity thresholds (39,40), which maintained the cluster-level Type I error rate at 
= 0.05. This cluster-level error rate was maintained by accounting for the spatial autocorrelation between voxels within each ROI (a priori analyses) and within the entire brain volume (exploratory analyses). Spatial autocorrelations and contiguity thresholds were empirically determined by Monte Carlo simulations (40).
Condition and Group main effects and Group by Condition interactions in BOLD activation within each cluster of activated voxels meeting the above Type I threshold criteria were confirmed with random-effects ANOVAs on the cluster-averaged BOLD activation values. The results from these confirmatory ANOVAs are presented as cluster F values with corresponding cluster size thresholds and p values that are corrected for multiple comparisons (40). Results are illustrated in Figures 2 and 3 as activation maps of those brain regions in which changes in BOLD activation met the above statistical and cluster threshold criteria.
|
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
60%. Specifically, there was no difference in the number of target-word color identification trials that were presented and completed during the incongruent condition of the Stroop task by the high reactors (median number of trials = 23; range = 18–26) and the low reactors (median number of trials = 24; range = 17–28), t < 1. These results show that color-identification accuracy was successfully titrated to 60% during the incongruent condition; they also show that task performance (accuracy) and amount of titration (total trials completed) was comparable between groups. Using a VAS that ranged from 0 (not at all) to 100 (extremely), participants rated the incongruent condition as being more stressful and demanding than the congruent condition (M ± SE incongruent vs. congruent VAS ratings: stress = 57.13 ± 5.22 vs. 21.38 ± 4.10; demand = 57.13 ± 5.22 vs. 22.50 ± 4.44, all main-effects-of-Condition F values (1,14) > 23.68, p values <0.001). No main effects of Group or Group by Condition interactions were found for either stress or demanding ratings, meaning that high and low reactors did not differentially rate the two conditions, all F values <1. These results show that, overall, the Stroop task elicited task demand and subjective distress ratings that were comparable in both reactivity groups.
Blood Pressure Reactivity
Across all participants, SBP increased from the baseline period to the congruent (M ± SE SBP = 4.13 ± 3.78 mm Hg) and incongruent (SBP = 12.31 ± 4.05 mm Hg) conditions of the Stroop task, F(2,13) = 17.78, p < .001. As Figure 1 illustrates, however, high reactors showed a larger increase in SBP from the baseline period to the congruent and incongruent conditions than low reactors, Group x Condition interaction F(2,13) = 9.37, p = .003. This interaction was explained with simple effects tests indicating that high reactors showed a significant change in SBP across the conditions (F(2,6) = 17.45, p = .003), whereas low reactors did not (F(2,6) = 1.67, p = .27). Post-hoc comparisons further confirmed that high reactors showed a significant increase in SBP from the baseline period to both the congruent (t = 2.56, p < .037) and the incongruent (t = 4.98, p < .002) condition; by comparison, low reactors did not show a significant change in SBP from the baseline period to either condition (t values <1.2, p values >0.27).
|
Across all participants, DBP also increased on average from the baseline period to the congruent (M ± SE DBP = 1.77 ± 1.41 mm Hg) and incongruent (DBP = 4.19 ± 1.35 mm Hg) conditions of the Stroop task, F(2,13) = 7.78, p = .006. In contrast to SBP, the DBP change from baseline to the congruent and incongruent conditions were comparable in both high and low reactors, Group x Condition F < 1. The lack of a group difference in DBP change parallels the same finding observed in these individuals when they completed the Stroop task in REACT. Therefore, these findings suggest that the two groups differ primarily in one type of cardiovascular reactivity: SBP reactivity.
The absence of a main effect of Group for both SBP and DBP indicated that high and low reactors did not differ in overall blood pressure across the conditions, similar to the finding that they did not differ at baseline in SBP or DBP when the structural MRIs were obtained (F values <1).
Condition-Related Regional Brain Activation
A Priori Analyses
The Stroop task elicited BOLD activation in the anterior cingulate, bilateral anterior insular, and orbitofrontal cortex (Figure 2). These main effects of Condition on BOLD activation in each of these a priori ROIs specifically reflected an increase in activation from the congruent to the incongruent condition of the Stroop task in a contiguous cluster of voxels in Brodmann area (BA) 32 of the right anterior cingulate cortex (cluster F(1,16) = 33.76, contiguity threshold = 23 voxels), Brodmann area 13 of the left and right insula (cluster F values (1,16) >15.69, contiguity thresholds >10 voxels), and a decreased activation of Brodmann area 47 of the right orbitofrontal cortex (cluster F(1,16) = 25.03, contiguity threshold = 8 voxels).
Exploratory Analyses
In addition to the condition-related BOLD activation that was observed in the above ROIs, exploratory whole-brain analyses showed that compared with the congruent condition, the incongruent condition elicited greater activation in the bilateral dorsolateral prefrontal cortex, the right posterior parietal cortex, the thalamus, and the right cerebellum, all cluster F values >27.90, contiguity thresholds >27 voxels.
Table 1 summarizes the findings for both a priori and exploratory analyses and lists the Talairach-Tournoux (41) coordinates for the voxels of peak activation (ie, those with the highest F value) from each cluster showing a condition-related change in BOLD activity. These Talairach coordinates were transformed from their respective MNI coordinates (http://www.mrc-cbu.cam.ac.uk/Imaging/mnispace.html).
|
Individual Differences in Condition-Related Regional Brain Activation
Similar to expressing larger magnitude SBP reactions to the Stroop task than low reactors, a whole-brain exploratory ANOVA showed that high reactors also expressed a greater activation of the posterior cingulate cortex (BA 31) during the incongruent condition compared with the congruent condition, Group x Condition cluster F(1,16) = 58.62, voxel contiguity threshold = 28 voxels (Figure 3).
Simple effects tests that were performed on the average BOLD signal intensity in this cluster of voxels in the posterior cingulate showed that a) high reactors expressed an increase in posterior cingulate activation from the congruent to the incongruent condition (F(1,7) = 19.20, p = .003) and that b) low reactors showed a decrease in posterior cingulate activation (F(1,7) = 46.11, p < .001). Inspection of the posterior cingulate activation patterns for each high and low reactive individual corroborated these group-level findings. To illustrate this individual-level consistency, Figure 3 (right panel) shows the standardized difference in posterior-cingulate BOLD activation values between the congruent and incongruent conditions for each high and low reactor. These difference scores were obtained for each person by a) subtracting the cluster-averaged posterior-cingulate BOLD-signal value in the congruent condition from the corresponding incongruent BOLD value and b) dividing that difference by the pooled SD of the differences. As shown, all high reactors expressed an increase in posterior cingulate activation from the congruent to the incongruent condition. Conversely, all the low reactors showed a decrease in posterior cingulate activation.
None of the ANOVAs on BOLD activation in the a priori ROIs showed a significant Group x Condition interaction. In addition, neither exploratory whole-brain nor ROI ANOVAs revealed a main effect of Group, meaning that the two groups showed comparable overall and condition-related levels of BOLD signal intensity. In addition, supplementary analyses showed that the two groups did not differ significantly in their BOLD signal intensity that was averaged across all brain voxels (M ± SD whole-brain averaged BOLD signal intensities: high reactors = 544.00 ± 68.62 arbitrary units; low reactors = 523.72 ± 56.43 arbitrary units, t < 1). In addition to the lack of a Group main effect in the ROI and exploratory analyses, this lack of an overall group difference suggests that the magnitude of the BOLD signal was comparable across the entire brain of high and low reactors.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Overall, the modified version of the Stroop color-word interference task elicited patterns of brain activation that parallel those observed in prior functional neuroimaging studies that have used different versions of the Stroop task (eg, 12,45). Specifically, the present Stroop task engaged both cortical and subcortical brain systems (listed in Table 1) that are thought to support attention, working memory, and behavioral response processes (for review see 27,45). In addition, the present Stroop task engaged a set of brain systems that are thought to generate and represent cardiovascular reactions to behavioral stressors: the orbitofrontal cortex, the anterior cingulate cortex, and the insula. Findings from a growing number of functional neuroimaging studies (10–14) show that these brain systems are consistently activated by behavioral stressors that also elicit cardiovascular reactions (eg, increases in blood pressure and heart rate). Thus, the changes in functional neural activity and blood pressure elicited by the Stroop stressor in the present study are consistent with these findings. One limitation of the present study, however, is that it was not designed to dissociate the functional neural activity that primarily supported the cognitive (eg, the attentive, working memory, and behavioral-response) processes that were associated with performing the Stroop task from the neural activity that was primarily associated with cardiovascular reactivity. This limitation could be addressed with different types of experimental design, such as the event-related fMRI design, in conjunction with continuous cardiovascular (eg, electrocardiographic or blood pressure) monitoring. These types of design would allow for a better resolution of the temporal order of functional neural and cardiovascular responses to a behavioral challenge. With this experimental approach, a more definitive assessment can therefore be made regarding the short-term functional neural responses that are primarily correlated with engaging in a behavioral challenge and those that closely precede or follow a peripheral cardiovascular reaction to that challenge.
Acting as a network, the orbitofrontal, anterior cingulate, and insular brain systems are held to process the motivational aspects of environmental stimuli and to support adaptive motor and cardiovascular reactions to behavioral challenges or stressors (for reviews see 16,17,21,24, 27,28,30–32). It is also posited that the orbitofrontal, anterior cingulate, and insular brain systems support these functional processes by integrating and representing multimodal external and internal (visceral afferent) sensory information and by regulating cardiovascular reactivity through their reciprocal projections to and from the amygdala, hypothalamus, and other midbrain and hindbrain cell groups (16,17,21,24,27,28,30–32). However, although we found that a demanding behavioral challenge (the Stroop stressor) elicited both cardiovascular reactivity and engaged the orbitofrontal, anterior cingulate, and insular regions of the cortex, we did not observe a difference in the functional neural activation of these regions between individuals who were classified as high and low cardiovascular reactors. Rather, high and low cardiovascular reactors differed markedly in their patterns of posterior cingulate cortex activation, specifically in Brodmann area 31. These differential patterns of functional neural activation can be tentatively interpreted within the context of recent theoretical views of the functional processes that are supported by the posterior cingulate cortex.
In several functional neuroanatomical models of the cingulate cortex, the posterior cingulate region has been viewed to support evaluative processes that are related to both cognition and emotion. These evaluative processes include a) maintaining a generalized form of attention to and representation of the environment (29,44); b) gauging the emotional salience and motivational significance of environmental events (42,43; cf 47); and c) monitoring the visual environment for threatening stimuli (44). These evaluative processes may be supported by reciprocal projections between the posterior cingulate, anterior cingulate, orbitofrontal, and parahippocampal cortices (42–44). Furthermore, two recent meta-analyses of the functional neuroimaging literature offer additional evidence that the posterior cingulate may be involved in evaluative processing of the environment. In the first meta-analysis, which included the results from 132 normal adults, Gusnard et al. (44) showed that functional activity in the posterior cingulate cortex consistently decreased in response to a broad range of cognitive tasks that required goal-directed behavior. The authors interpreted this consistent pattern of posterior cingulate deactivation to indicate that when an individual engages in a goal-directed behavior, then vigilance or generalized attention to the environment is curtailed in order to focus attention on meeting the specific behavioral demands of a given task. Also, to the extent that the posterior cingulate does not participate in focused attention, but rather in generalized attention to the environment, then its activity should decrease during a goal-directed behavior. In the present study, this pattern of decreased posterior cingulate activation was shown by all of the low cardiovascular reactors to the demanding incongruent condition of the Stroop task; conversely, a pattern of increased posterior cingulate activation was shown by all of the high cardiovascular reactors. As suggested by Gusnard et al., this increase in posterior cingulate activation by the high reactors to the Stroop task may have reflected an increase in a generalized form of attention or heightened vigilance that was elicited by this behavioral stressor. Another possibility is that increased activation of the posterior cingulate in the high reactors may have (also) reflected an increase in the emotional or motivational evaluation of the Stroop stressor—as indicated by another meta-analytic review of the functional neuroimaging literature by Maddock (43).
In a meta-analysis of 51 functional neuroimaging studies, Maddock (43) showed that increased activation of the posterior cingulate was consistently observed in response to tasks that required the processing of unpleasant emotional stimuli. These patterns of increased posterior cingulate activation to emotionally unpleasant stimuli led Maddock to conclude that the posterior cingulate may be involved in the evaluative emotional processing of environmental information (cf 46)—a view that parallels another conceptualization of posterior cingulate function by Luu and Tucker (42). Specifically, Luu and Tucker have proposed that functional activity in the posterior cingulate supports the representation of the perceptual and motivational aspects of the environment; also, projections from the posterior cingulate to limbic and hippocampal memory systems serve to adjust motivated behavior to meet the contextual demands of cognitive or emotional challenges. By these converging views, the pattern of increased activation of the posterior cingulate to the Stroop task shown by high cardiovascular reactive individuals could have thus reflected an increase in their evaluative (possibly unpleasant) processing of the Stroop task.
Inconsistent with this interpretation, however, are the present results regarding task-related ratings of subjective distress and demand. These rating results showed that high and low cardiovascular reactors did not differ in their Stroop task ratings of subjective distress or demand; however, these ratings were obtained during a practice session before the participants were placed in the bore of the MR scanner. Consequently, it is possible that these ratings did not accurately reflect the subjective distress and demand that were elicited by completing the Stroop stressor in the MR scanner. Indeed, the scanning environment is not a neutral context: participants are placed in a confined space for an extended period of time and they are exposed to continuous noise (>90 dB) from imaging. This particular limitation of the present study warrants future neuroimaging studies that better assess task-related ratings of subjective distress and demand in the context of the scanner environment. Such an assessment would facilitate a better understanding of the potential evaluative processes that may be associated with individual differences in cardiovascular reactivity and functional neural activation.
We also considered a more parsimonious explanation of the present results: the difference between high and low cardiovascular-reactive individuals in BOLD activation may reflect a difference in central hemodynamic, but not neural, activity. Indeed, changes in BOLD activation reflect changes in both central hemodynamic and neural activity because the BOLD signal results from regional changes in blood flow, blood volume, and oxygen consumption in areas of neural activity (47). Therefore, it is possible that increased BOLD activation in high cardiovascular reactors may simply reflect their tendency to show both increased peripheral and increased central (eg, cerebrovascular) hemodynamic reactivity. However, two findings of the present study do not lend support to this possibility. First, we did not find a main effect of Group on voxel-by-voxel BOLD activation in a priori brain ROI analyses or in exploratory whole-brain analyses. Second, in supplementary analyses, we did not find a difference between high and low cardiovascular reactive individuals in their overall BOLD signal intensity across the entire brain volume. These two findings thus suggest that differences in BOLD activation between high and low cardiovascular reactors were not likely due to the possible confounding effects of group differences in central hemodynamic factors.
To summarize, we have made an initial characterization of a potential functional neural correlate of individual differences in one type of cardiovascular reactivity to behavioral stressors, namely, SBP reactivity. We found that individuals with a tendency to express exaggerated SBP reactions to a Stroop color-word interference task, a behavioral stressor that is widely used in cardiovascular reactivity research, showed an increase in posterior cingulate activation to this stressor. By contrast, individuals with a tendency to express attenuated SBP reactions showed a decrease in posterior cingulate activation. This differential pattern of activation in the posterior cingulate cortex by high and low cardiovascular-reactive individuals a) was unlikely to be the result of group differences in the nature of the hemodynamic BOLD response, b) was consistent for each individual in each group, and c) was observed using a conservative analytical methodology that corrected for multiple statistical tests in the entire brain volume. We speculate that increased posterior cingulate activation by high cardiovascular-reactive individuals may reflect their heightened attentive or emotional processing of this behavioral stressor. This tentative speculation is consistent with current views of the functional role of the posterior cingulate in cognitive and emotional evaluative processes.
In closing, it should be noted that cardiovascular reactivity is not a one-dimensional construct: individuals differ (at minimum) in the magnitude and pattern of their hemodynamic, vascular, and cardiac–autonomic reactions to behavioral stressors (48–50). Also, these different patterns of cardiovascular reactivity may relate to cardiovascular health outcomes in different ways (1,2). Thus, whether there are different functional neural correlates of individual differences in other dimensions of cardiovascular reactivity and whether these functional neural correlates have implications for the prediction or control of cardiovascular health outcomes are open questions for future research.
We thank V. Andrew Stenger and Michael J. Eddy for their technical assistance.
|
NOTES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
This research was supported by an NHLBI-NRSA (1-F32–71333–01), an NIMH-MRSDA (1-K01-MH-070616–01), and by pilot funds awarded through the Pittsburgh Mind-Body Center at the University of Pittsburgh and Carnegie Mellon University (NIH HL65111 and HL65112) to Peter J. Gianaros.
DOI:10.1097/01.psy.0000151487.05506.dc
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
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]
|
|
|
|
 |
|
 P. J. Gianaros, J. R. Jennings, L. K. Sheu, S. W.G. Derbyshire, and K. A. Matthews Heightened Functional Neural Activation to Psychological Stress Covaries With Exaggerated Blood Pressure Reactivity Hypertension, January 1, 2007; 49(1): 134 - 140. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Flaa, H. H. Mundal, I. Eide, S. Kjeldsen, and M. Rostrup Sympathetic Activity and Cardiovascular Risk Factors in Young Men in the Low, Normal, and High Blood Pressure Ranges Hypertension, March 1, 2006; 47(3): 396 - 402. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
