| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the Departments of Neurology (J.T., N.G., J.L., B.W.), Radiology (J.T., A.F., P.N.), and Radiologic Sciences (R.W.), Thomas Jefferson University/Jefferson Medical College, Philadelphia, PA.
Address correspondence and reprint requests to Joseph I. Tracy, PhD, ABPP (CN), Associate Professor of Neurology and Radiology, Thomas Jefferson University/Jefferson Medical College, 900 Walnut Street, Suite 206, Philadelphia, PA 19107. E-mail: joseph.i.tracy{at}jefferson.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Methods: We measured regional brain activity through functional magnetic resonance imaging and directed subjects (6 men and 11 women) to attend to their own heartbeat, and to a heartbeat played on an external tape.
Results: Statistical parametric brain mapping revealed the importance of right (nondominant) parietal cortex to directing attention internally to one’s visceral state and focusing on a specific body signal.
Conclusions: The parietal activation may be taking advantage of monitoring skills typically utilized for vigilance to the external environment, in addition to working as a higher-level recognition system for signals emerging from the viscera. The finding suggests that the parietal cortex plays a central role in an interoceptive attention system that monitors bodily states.
Key Words: attention • interoceptive processing • visceral processing • somatic processing
Abbreviations: TE = echo time; TR = repetition time; FOV = field of view; hz = hertz; db = decibel; SPM = statistical parametric mapping; ROI = region of interest; BA = Brodmann’s area.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
The purpose of the current study was to map the brain regions implementing a very specific and simple internally directed attentional activity, i.e., attending to one’s own heartbeat. This should provide a clue to the brain network that implements attention to internal visceral state. Posterior parietal cortex (Brodmann’s areas [Bas] 40 and 7) has been implicated in a wide variety of external attention tasks such as spatial location selection, monitoring extrapersonal space, and sustained attention (4). For instance, functional magnetic resonance imaging (fMRI) investigations have revealed that tasks of tactile exploration (5), covert visuospatial attention (6,7), oculomotor search (8), and auditory target detection (9) produce activation in the superior and inferior parietal lobules, the banks of the intraparietal sulcus, and the medial parietal cortex (10,11). Also, patients with lesions in this area are known to manifest deficits such as unilateral neglect and extinction (1). It is clear that right parietal lobe damage produces more dramatic deficits of this kind than damage of the left, and that the right parietal cortex plays a stronger role than the left in attentional vigilance (12) and object-based selection (11). Thus, one can conclude that the right parietal lobe has a significant role in a complex attentional network for external signals.
The somatosensory cortex lies anterior to the posterior parietal cortex. The primary somatosensory region’s role in tactile registration and initial recognition is well known (13). Somatosensory association areas (posterior to BAs 1, 2, and 3) and areas that include anterior sections of the parietal cortex (BAs 5, 7, and 40), however, have a less clear functionality. These regions have been implicated in corporeal awareness, the storage of information about structural descriptions of the body, and the relation of the body and body parts to each other and to extrapersonal space (14–19). Additional studies have shown that the superior parietal cortex is a projection area for visceral information arriving from the periphery (20,21).
Moreover, it is the right hemisphere that is dominant for such corporeal processing as lesions of the right hemisphere appear to be highly disruptive to body image and are associated with body schema or body image disorders (14,15). For instance, allodynia, the perception of pain to stimuli that do not usually produce pain, and asomatognosia, the failure to perceive a body part (personal neglect) or attend to the side of the body contralateral to the lesion, are usually the result of right parietal lobe lesions (19). Anosognosia, a disorder involving denial of deficits such as hemiparesis, has also been associated with deficits in the right parietal region (22–24). There are a variety of other disorders of body schema that may emerge from a right hemisphere deficit in corporeal comprehension and awareness (e.g., conversion disorder, asymbolia for pain, allesthesia, anosodiaphoria, phantom limb phenomena, macrosomatognosia, microsomatognosia) (25). Several studies have shown that conversion symptoms more commonly affect the left side of the body and are more frequent in individuals affected by structural or physiological disorders of the right hemisphere (26–28). A right-hemisphere connection to heart rate and cardiovascular functioning has also been suggested by Hugdahl (29). Heartbeat detection skill seems to be associated with measures that typically implicate a bias toward right hemisphere processing such as emotional responsiveness and conjugate eye movement preference ("left eye movers" (30)). Katkin et al. (31) specifically argue for a right hemisphere bias to heartbeat detection (32). It is also interesting to note that heartbeat evoked potentials have been observed in the right hemisphere during performance of heartbeat tracking tasks (32). Finally, vagus nerve enervation of the heart is preferentially controlled by the right hemisphere (33).
Awareness of our internal visceral state and its relation to other body information and the environment is likely to require integration with current somatosensory information. Therefore, it makes sense that areas proximal to the somatosensory cortex in the parietal cortex (BAs 5, 7, and 40) would be involved in somatic integration processes of this kind. Also, it is clear that the right hemisphere has a commanding role in such processes. An important point to consider is that these activities of corporeal processing have cognitive requirements in terms of monitoring body signals as well as maintaining awareness of the body’s relationship to extrapersonal space and environmental stimuli. This region between the somatosensory and posterior parietal cortex seems aptly situated to take advantage of both the tactile and sensory processing properties of the primary somatosensory cortex, and the attention selection and environmental monitoring skills of the posterior parietal cortex.
As a final point of review, we should mention the important study by Critchley and colleagues (34) who had subjects attend to notes generated either to reflect pulse-derived heartbeat patterns or similarly timed tones with subjects judging whether the stimulus timing was synchronized or desynchronized. Attention to the notes representing the heartbeat activated several regions including the anterior cingulate, lateral sensorimotor, supplementary motor cortex, and bilateral insular cortices. Critchley et al. also found that the right anterior insula cortex best predicted accuracy in making timing discriminations about the notes reflecting the heartbeat, and that the extent of activity in this region correlated with self-rated body awareness. Craig (35) in a recent article interpreted this as evidence for a distinct interoceptive pathway, which may mediate human awareness.
Our goal was to devise a study that would involve attending to an internal visceral signal and use functional neuroimaging to determine the brain regions implementing this form of internally directed attention. We chose as the stimulus of focus the heartbeat, with the rationale that this would likely be the most robust and detectable internal body signal in healthy individuals. A heartbeat tracking paradigm has been used previously (36–39) in studies of heartbeat perception, but none of these studies involved brain imaging.
Based on the potential convergence of an attentional focus and monitoring system with mechanisms that monitor somatic processing, we hypothesized that the region encompassing the anterior parietal cortex (aspects of Brodmann’s area 5/7/40) between the somatosensory and posterior parietal cortex will mediate attention to one’s heartbeat. Given the right hemisphere bias of such corporeal processing effects, we also expected activation during our internal attention task to be biased toward the right hemisphere.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
fMRI Procedure
Whole-brain Blood Oxygen Level Dependent contrast functional images were collected at 1.5 tesla (General Electric LX, General Electric, Milwaukee, WI) with the following parameters: 26 parallel axial-oblique slices, single-shot echoplanar gradient imaging sequence, echotime (TE) = 54 ms, repetition time (TR) = 4.0 seconds (interleaved collection, contiguous slices), field of view (FOV) = 24 cm, 128 x 128 data matrix, flip angle = 90 degrees, bandwidth = 62 kHz. The in-plane resolution was 1.875 mm2 with 4-mm thickness. T1 weighted images (26 slices, FOV = 32 cm, TE = 9 ms, TR = 450 ms, 256 x 256 mm) were collected using standard spin-echo pulse sequence with an axial-oblique angle set to the anterior commissure-posterior commisure line, and this provided an anatomical reference to prescribe the slice locations for the echoplanar images. Two scanning runs were conducted with the initial two volumes of each dropped to allow for T1 equilibration effects.
Experimental Task Procedure
Participants were told that during the course of the study they would be listening to various sounds, including heartbeats and tones, with rest periods. The participants were given an instruction sheet describing the various task conditions and a verbal explanation was also provided. Participants kept their eyes closed during the experiment.
In the key experimental condition ("internal heartbeat"), participants were instructed to place their right, dominant hand over their heart on hearing the verbal cue "chest" (approximately between the second and fifth intercostal spaces, over and to the left of the sternum, with the experimenter demonstrating this). This is the area where physicians place their hand to detect the point of maximum impulse when conducting a cardiovascular examination. It is considered the point at which the ventricular myocardium maximally strikes against the inside of the chest wall during systole. The participants felt the point of maximal impulse through pressure/mechanoreceptors in the hand, but it was also possible that they felt their heartbeat in pressure/mechanoreceptors in the body wall. Participants were instructed to focus their attention inward and attend to their own heart beating, and count the number of beats silently. The participants were given an opportunity to practice this outside the scanner in front of the examiner.
In the key comparison condition ("external heartbeat"), participants were instructed to place their dominant, right hand on their right thigh (ipsilateral quadriceps) when they heard the word "leg" and focus their attention on the sound of a heartbeat from the headphones. All subjects heard the same tape recorded heartbeat taken from the neurology department educational files. This involved an actual heartbeat, beating at an interval of 0.83 seconds, which approximates the heartbeat rate for healthy adults (72 beats per minute). Participants were told to count the number of beats silently until this period ended. The instructions emphasized devoting their attention to the heartbeat sounds arising from the headphones. This procedure provided a means of gauging brain activation associated with external monitoring of a similar stimulus source with the procedural aspects of the tasks matched to the experimental task in terms of tactile stimulation. As a check on compliance with the task, on leaving the scanner, subjects were asked to estimate the total number of heartbeats that they heard "on average" for the individual periods. All subjects provided estimates and these estimates for all subjects were in the range that worked out to a beat every 0.5 to 1.5 seconds for the external condition, and one beat every 0.5 to every 2 seconds for the internal heartbeat condition. These estimates were not considered veridical and were collected as an approximate indication of compliance.
The primary control condition involved silent counting. Here, participants were instructed to count silently to themselves (e.g., "1, 2, 3, 4 . . .") on hearing the verbal cue "number" until the period ended. This condition provided a means to control for the activity of counting that subjects engaged in during the internal and external heartbeat conditions. Finally, the participants were told that during certain periods of the study various sounds would also be played. An auditory control condition composed of distinct periods of auditory white noise [110 hz, 56 dB] was utilized. This noise was constant during the period. A second auditory control was used comprised of a set of alternating, rhythmic neutral tones [randomly chosen: 245 hz, 53dB and 195 hz, 31 dB]) was used as a nonsomatic tone control. The tones occurred at a rate of 0.8 seconds to approximately simulate the rate of the heartbeats. During these periods subjects were instructed to simply listen and focus their attention on the tones played. It was made clear that the verbal cues "leg, chest, number" were prompts to do particular attentional and counting tasks, always silently, and in the other tasks they were to simply listen attentively and lay still. Rest periods (56 and 45 seconds each) occurred at the beginning and end of the experiment. During the entire fMRI session, subjects wore sound reducing/sound canceling headphones (Avotec, Inc., Stuart, FL) that provided the auditory input (instructional cues, auditory control material [white noise or rhythmic neutral tones], and the tape of the heartbeat during the external heartbeat condition).
There were five epochs of internal heartbeat (listening to their own heartbeat), with durations of 12, 16, 18, 20, and 24 seconds, and a total duration of 90 seconds. The same number and length of periods were used for the external heartbeat condition (listening to a tape of a heart beating) with an identical total duration of 90 seconds. There were seven epochs of silent counting lasting 10 seconds, for a total duration of 70 seconds. There were 15 epochs of auditory control where participants attended to either the white noise (8 periods of 10 seconds) or rhythmic neutral tones with epochs varying in length similar to the heartbeat conditions (5 periods of 12, 16, 18, 20, and 24 seconds) for a total duration of 170 seconds. The total run time of each scanning run was 8 minutes and 44 seconds with 131 whole brain volumes collected. Within the scanning runs the key experimental conditions (internal heartbeat, external heartbeat) and the control conditions (silent counting and auditory controls) were presented in a random order.
Image Post-Processing and Statistical Analysis
Volumes within an fMRI run were co-registered or computationally aligned to correct for interscan movement with the third volume of the time series used as the reference (nota bene, the first two volumes were dropped to avoid T1 saturation effects). All volumes were transformed into standard Montreal Neurologic Institute anatomical space using the participant’s T1 image and the statistical parametric mapping normalization procedure (41,42). Volumes then underwent spatial smoothing (Gaussian kernel of 8 x 8 x 8 mm, full width half maximum) to increase signal-to-noise and account for residual intersubject differences in anatomy. A high-pass filter removed low frequency fluctuations. The general linear model procedure of the statistical parametric mapping software (SPM ’99, http//:www.fil.ion.ucl.ac.uk/spm) was used to create models containing sinusoid waveforms representing the key experimental and control conditions (internal heartbeat, external heartbeat, white noise, alternating neutral tone, and rest). One subject’s data were found to be highly unusual and contained artifact inolving motion and out of brain activity. This subject was considered an outlier and the data was excluded from the analyses. As noted, the first two BOLD volumes of each run were dropped to account for T1 equilibration effects.
Voxels associated with activation emerging from the following contrasts were then identified: a) internal heartbeat compared with external heartbeat, b) internal heartbeat compared with rest periods, and c) internal heartbeat compared with silent counting. Analyses focused on the region of interest (ROI) given by our hypothesis, the anterior parietal cortex. Primarily, we used whole brain SPM analysis as one method to test our hypothesis. As a second method, we tried to capture activity in the general region of the anterior parietal cortex using an automatic anatomical labeling system developed by Tzourio-Mazoyer et al. (43). A set of ROIs was imported from a Website (https://sourceforge.net/projects/marsbar) into the Marsbar region of interest toolbox for SPM (44). The ROI that was developed for this experiment was formed separately for the left and right parietal brain regions using the superior parietal, inferior parietal, and supramarginal gyrus volumes (n.b., inferior gray matter structures such as the insula were not included). An ROI for the left hemisphere was included in the analysis to help us judge whether the parietal lobe involvement that we hypothesized about was truly laterlized, and to have a control ROI where no activation was expected. The right hemisphere ROI (Fig. 1) had the following properties: center of mass (Talairach coordinates; x = 41.6, y = –43.8, z = 46.2), anterior boundary (x = 65, y = –10.5, z = 23; BA 40), posterior (x = 12.6, y = –79, z = 48.8; BA 7), superior (x = 22, y = –47.7, z = 73.6; BA 7), and inferior boundary (x = 61.9, y = –18.2, z = 17.5; BA 40). The left hemisphere ROI (Fig. 2) had the following properties: center of mass (x = –38.6, y = –45.63, z = 44.94), anterior boundary (x = –56.6, y = –16, z = 29; BA 40), posterior (x = –12.6, y = –79.1, z = 48.8; BA 7), superior (x = –14.7, y = –42.8, z = 75.8; BA 7), and inferior boundary (x = –57.6, y = –21.3, z = 14; BA 40). In the automated labeling system, the definitions of the inferior parietal region followed the guidelines of Salamon et al. (45). The boundaries of the parietal and occipital lobe followed those given by Dejerine (46). Similar interpolation methods were used for both hemispheres. The authors of the automated labeling system (43) acknowledged that certain regions such as the inferior parietal lobe showed a very complex and highly variable cortical pattern, and that this resulted in imperfect homologues (asymmetries) for these regions across the left and right hemisphere.
|
|
For each ROI the mean parameter estimate (t-statistic) was calculated separately for each of the above contrasts using the appropriate linear compound of parameter estimates with a t-statistic computed at every voxel in the ROI (ROI voxel height threshold of p < .05 corrected). These subject-specific mean parameter data were then entered into a group level random effects analysis that utilized a one-sample t test (two-tailed, test value = 0) to determine statistical significance. This analysis was conducted separately for the left and right ROIs.
A separate whole-brain analysis was conducted with the above three contrasts using the standard SPM approach. The contrasts previously described used the appropriate linear compound of parameter estimates with a t-statistic (two-sided) computed at every voxel utilizing a voxel threshold. The resulting subject-specific contrasts were then entered into a random effects model utilizing a one-sample t test. To assess statistical significance (47), a height threshold of p < .05 (corrected, t = 2.5) was used. Only clusters of activation above this and an extent threshold consistent with image smoothness (i.e., expected voxels per cluster, k = 28) are reported.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
|
The one sample t test for the left hemisphere ROI and the internal heartbeat minus external heartbeat contrast revealed no statistically significant brain activation (Table 1). The internal heartbeat minus rest contrast showed some minimal brain activation that may be considered a trend (p < .06). Lastly, the internal minus silent counting, and internal minus the auditory control conditions (either white noise or the rhythmic tones) produced nonsignificant results.
The SPM whole-brain analysis data with z-maxima and cluster size is presented in Table 2. The contrast comparing the internal and external heartbeat conditions revealed one significant cluster of activation in the right anterior/inferior parietal cortex (BA 40). This cluster is displayed in Figure 3 (in t-units). The anterior extent of this cluster lay in the somatosensory cortex (S1). The posterior, inferior, and superior extents lay within the parietal cortex (BA 40). A similarly located cluster was observed in the internal minus silent counting comparison and the internal minus auditory control (alternating rhythmic tones only) comparisons. This suggests that the activation associated with the internal condition is quite limited and different from the external condition. The contrast comparing the external and internal heartbeat conditions produced no statistically significant clusters. The external condition minus silent counting yielded three significant clusters involving the right middle temporal gyrus/insula (BA 21, k = 574, maxima z = 4.46, p = .006), inferior frontal gyrus (BA 44/6, k = 899, maxima z = 3.04, p < .0001), and superior frontal gyrus (BA 10, k = 490, maxima z = 3.03, p < .05).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Previous studies have identified the parietal operculum (SII) as a brain region responsive to visceral sensations as evoked by rectal or esophageal distension (21,48). There is a debate over whether the areas within SII responsive to visceral versus somatic sensation are spatially distinct (21,49) or overlapping (50), and whether the responses observed in these regions from brain imaging techniques ultimately are driven by the presence of pain or unpleasant affect rather than primary visceral or somatic sensations. Eickoff et al. (51) studied ano-rectal stimulation and argued that the spatially distinct responses found in parietal cortex reflected functional segregation of visceral and somatosensory processing. Interestingly, this functional segregation was not observed in the insula suggesting that this region responds to both and plays a role in integrating these sensations. They did not observe activation associated with pain or affective reactions. Our data are in general agreement with the findings of Anders et al. (52), who utilized an emotional reactivity paradigm (viewing emotional pictures) and found anterior parietal activation that was also correlated with peripheral physiologic responses. The degree to which such activation correlated with the verbal report and arousal varied, most likely with changing levels of attention or cognitive processing during stimulus perception. This suggested the parietal activity did not require either awareness or strong emotional valence. The Anders et al. data (52), as does ours, implicates the anterior parietal region in the processing of internal state and peripheral information. Our study did not test for nor measure awareness or emotional valence, though clearly our findings are based on processing that involved full awareness.
Critchley et al. (34) observed activity in the insula, somatomotor, and cingulate cortices in response to a heartbeat detection task. They also conclude that the right anterior insular cortex is important to the integration of interoceptive (visceral) and external somatic information. Based on the fMRI activation data and additional volumetric data, Critchley and colleagues (34) argue strongly that the right anterior insular/opercular cortex was strongly related to accuracy of heartbeat detection (interoceptive accuracy) and visceral awareness. It is interesting to note that the Critchley study (34) differed from our paradigm in that it used an external, auditory tone as a surrogate for the heart and, thus, subjects were not attending to one’s heart directly as one might do in real life by feeling and "listening" internally to one’s own heartbeat. The right anterior insular cortex activation they observed showed a significant interaction with factors in the experiment reflecting an external attentional focus (a rogue tone among the sequence of notes reflecting the heartbeat, and a delay in the onset of the feedback containing the notes). The authors interpreted this as evidence that the right anterior insular cortex is important to the integration of interoceptive and external information. In contrast, our paradigm and findings allowed us to identify a region more specifically related to the internal monitoring of interoceptive signals without the presence of external information. However, because we did not measure autonomic arousal or affective state, we cannot definitely state that our lack of insular activation reflects a lack of autonomic activation or emotion processing. It is interesting to note that we did observe insular activation during the task condition that did involve an external attention focus (external condition). Consistent with the Critchley et al. (53) interpretation, we may have been observing the role of the insular cortex in the external condition because of the combining or integrating of the external heartbeat signal with the tactile response of touching the thigh. Note, none of the contrasts highlighting activation in the internal condition showed insular activation. Thus, it seems unlikely that such insular activation was common to both the internal and external conditions, and simply cancelled out in the internal minus external statistical contrast described above.
As noted, a different view of the insula suggests that this structure is responsive to autonomic arousal, pain, or affective (unpleasantness) processing (54). For instance, DiGangi and colleagues (20) in an fMRI study of stress urinary incontinence found that after a treatment intervention (pelvic foor muscle training with electromyography biofeedback) a more focused activation in the primary motor and somatosensory cortex emerged but, in addition, there was reduced activation in the insula, right frontal operculum, and the anterior cingulate cortex. The authors interpreted these reductions as a treatment outcome resulting from diminished emotional arousal during micturition. An fMRI study by Lotze et al. (21) of anorectal stimulation in healthy adults produced increased activation in the insula, cingulate, and orbitofrontal lobe. The authors concluded that the insula and cingulate were crucial to attentional regulation and registration of this unpleasant visceral sensation. Note, our study does not distinguish between visceral and somatosensory processing as chest wall tactile sensation and attention devoted to the visceral sensation of the heart could both be at work in our data emerging from the internal condition. In light of the fact that both visceral and somatosensory stimulation were present during the internal condition, the lack of activation in the insula is particularly striking. Given this lack of activation and the presence of insula activation during the external condition, our data can be seen as inconsistent with the notion that the insula integrates visceral and somatic information. In contrast, our data can be seen as consistent with the possibility that the insula responds to affectively unpleasant experiences. The external condition may have evoked unpleasant images of injury or grotesqueness as the circumstances in which the heart is perceived outside the body tend to involve serious injury, etc.
We should note that in our whole-brain analysis (random effects, group model) the only statistically significant region of activation was the parietal cortex. This stands in contrast to the network of activation reported by Critchley et al. (34). Thus, our more modular finding may suggest that internally directed attention is not modulated by other regions to any large degree. This makes sense if one considers that other regions of the brain, such as the frontal lobe (e.g., goal-directed behavior), temporal lobe (e.g., language and memory functions), and posterior visual regions, are generally dedicated to processing external stimuli or externally relevant goal states. Internally directed attention by design may limit communication with these other brain regions in order to function properly and not disrupt any ongoing internal focus. We did observe left parietal activity for the internal condition compared with silent counting and as a trend (versus rest) in the ROI analyses. The nature of this activation is unclear, but seems unlikely to be part of an externally directed attention network. Our task did not obtain measures of heartbeat detection though subjects estimated the average number of heartbeats they counted during the internal and external periods. We do not consider these estimates veridical, but they do provide us with some indication of compliance with the task.
The internal condition did involve more tactile (pressure wave/vibration sense) detection of the heartbeat compared with the auditory input of the external heartbeat condition. This does present a confounding element to our results because the internal heartbeat task is not a pure interoceptive task as it did involve some tactile sensation from the hand. One assurance that the activation we observed was related to the heart and not the tactile sensation of the hand is that the latter would have involved the left somatosensory cortex and no such activation was observed in our statistical images. If sensation from the skin surface was prompting significant activation in the external heartbeat condition, left hemisphere somatosensory activation would have been observed because it too involved tactile perception involving the right hand and right leg, but again no such activation was observed. Nevertheless, we cannot completely rule out the possibility that our finding for the internal heartbeat condition has emerged from a combination of visceral plus chest wall awareness. A future study may need to involve constant hand placement or utilization of some other method for potentially attending to visceral state, and then manipulating attention to be either internally or externally directed.
We acknowledge that factors such as body fat composition (55,56), fitness level (57), and acute (58,59) and baseline (34) anxiety level are known to affect the ability to detect one’s heartbeat. None of these measures are reported here for our subjects and their relation to the activation properties we observed remain unknown. Also, there are other factors of interest such as gender or cerebral lateralization that may alter the findings reported here and are worth investigating.
We also acknowledge that we did not measure mean heart rate during the scanning session. Anxiety, for instance, likely increases heart rate, which in turn might increase BOLD signal in the heartbeat detection areas of the somatorsensory or parietal cortex. Anxiety levels, however, whether caused by scanner noise or other factors related to the artificial environment of the MR scanner would have been common to both the internal and external conditions and therefore cancelled out in our key experimental comparisons. Another confounding factor was that during the internal condition, our subject may have utilized sensory perception of the heart pulse on the fingers and hand, and this was not present during the external condition. The external condition required somatic processing that did not involve this heart pulse aspect. Thus, it is possible that the difference we observed between the internal and external conditions may have arisen from subtle differences in sensory perception not internal attention, per se. Also, it may have been more difficult to find and monitor one’s own heartbeat in terms of limb positioning then finding one’s thigh and listening to a tape of a heartbeat. However, it must be said that subjects knew and practiced exactly where to place their fingers during both the external and internal conditions before entering the scanner, and thus limb positioning and factors related to the processing of extrapersonal space (60) should have been fairly similar.
The generality of our finding is unclear but it raises the possibility that the anterior parietal cortex may play a central role in disorders where thresholds for perceiving bodily signals are aberrant, such as psychogenic pain, psychogenic epilepsy, somatization or conversion disorders, and phantom limb pain. Our results do differ from some studies conducted with these types of patients. For instance, Hakala et al. (61) studied somatization patients using positron emission tomography (PET) and found lower rates of cerebral metabolism than those in healthy volunteers in the bilateral caudate, left putamen, and right precentral gyrus. However, a study using repetitive transcranial magnetic stimulation in phantom limb pain patients demonstrated temporary pain relief with application to the contralateral parietal lobe. This implicates parietal cortex in the re-organization that is the pathophysiologic basis for the perceived pain (62). Also, Appenzeller and Bicknell (63) reported the disappearance of phantom limbs after removal of vascular lesions to the contralateral parietal cortex implicating this structure in the cortical representation of the deafferented limb. Such findings are consistent with our data.
Heart rate perception can be influenced by a wide range of factors such as a person’s ability to discriminate other stimuli well, such as light and sound (64), even beliefs about one’s heartbeat (65), though it does not seem to be influenced by heart stroke volume (66). Healthy individuals who are accurate in detecting heartbeats were found to exhibit significantly lower scores on a scale of somatosensory amplificiation. This suggests that somatosensory amplification processes are not due to heightened sensitivity to body sensations, but a cognitive bias toward misinterpretation of body sensations (67). The later suggests that disorders of somatosensory amplification such as hypochondriasis (68,69) may not be mediated by brain regions that implement physiological sensitivity. Our data would allow for the hypothesis that somatosensory amplification may involve the parietal cortex.
The location of the internal attention activity we observed suggests it may be taking advantage of the monitoring skills typically utilized for vigilance to the external environment that are available through the posterior parietal cortex. In addition, its location in association with areas posterior to the primary somatosensory region, yet still including the somatosensory cortex itself, is consistent with the notion that the activation we observed is working as a higher-level recognition system for signals emerging from the viscera and other aspects of our somatic, corporeal state. Our data provide an initial step in modeling the functional neuroanatomy of an interoceptive attention system that monitors visceral and somatic signals, and internal body states. The next step in this research, in addition to replication, would be to determine the activation properties of this region in individuals who may possess atypical responses to internal body states and signals, such as patients with psychogenic pain, somatization disorder, or psychogenic nonepileptic seizures. These conditions may be driven by many factors (anxiety reduction), but may also relate to perturbed function in brain areas that contribute to registering and interpreting one’s internal state.
|
NOTES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
DOI:10.1097/PSY.0b013e31815b60cf
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
