Abstract

Noxious stimulation of skeletal muscle evokes pain that is often referred into distal areas. Despite referred pain being of significant clinical importance, the brain regions responsible for the perception of referred pain remain unexplored. The aim of this investigation is to define these regions using functional magnetic resonance imaging. We induced muscle pain by hypertonic saline injections (0.5 ml) into the tibialis anterior (TA) or flexor carpi radialis (FCR) muscle. TA injections evoked pain that was referred to the ankle/foot in 10/17 subjects, whereas FCR injections evoked pain that was projected into the wrist/hand in 6/12 subjects. Regional brain responses were statistically tested by convolving the temporal profile of the subjective pain intensity rating with the hemodynamic response function. For all subjects, signal increased in the region of primary somatosensory cortex (SI), which represents the leg or arm, that is, the area corresponding to the injection site. However, for those subjects who reported referred pain, signal intensity increases also occurred in the SI region representing the foot or hand. Interestingly, differential signal changes also occurred in anterior cingulate, cerebellar, and insular cortices. This is the first study to provide evidence of cortical differentiation in the processing of primary and referred pain.

Introduction

Pain that originates in muscle has a dull, aching quality that is diffuse, difficult to localize, and is often referred to sites remote from the original locus of noxious stimulation, that is, referred pain. We have recently shown that the deep dull ache induced by hypertonic saline injection into the tibialis anterior (TA) muscle is often associated with pain that projects into the ankle and dorsum of the foot (Graven-Nielsen et al. 1997b; Henderson, Bandler, et al. 2006). Although the referral of muscle pain was described over half a century ago (Kellgren 1938; Feinstein et al. 1954), the mechanism(s) responsible for this phenomenon are poorly understood.

Because referred pain can be induced during a complete sensory block of the region of the referred pain, it is likely that the pattern of referral results from a centrally mediated mechanism (Feinstein et al. 1954). It has been suggested that pain referred from visceral to cutaneous structures results from the convergence of nociceptive afferents in the dorsal horn, that is, viscero-somatic convergence (Foreman et al. 1984). However, electrophysiological studies reveal only rare convergence at the level of the dorsal horn between nociceptive afferent neurons innervating different muscles or between those that innervate muscles and other deep tissues (Hoheisel and Mense 1990). Given this, it is possible that referral of muscle pain results from the convergence of nociceptive neurons at higher brain centers, such as the thalamus and/or somatosensory cortex. In a previous functional magnetic resonance imaging (fMRI) investigation we found that both muscle and cutaneous pain evoked in the leg resulted in signal intensity increases in the region of the primary somatosensory cortex (SI), which receives sensory input from the region of the injection, that is, the paracentral lobule. In addition, during muscle pain, signal intensity increased in the region of SI that corresponded to the region that receives sensory input from the site of pain referral (Henderson, Bandler, et al. 2006). These data suggest that the SI cortex may encode the referral pattern of muscle pain. However, owing to its location within the sagittal sulcus, the region of SI cortex, which receives sensory input from the leg is difficult to distinguish from the immediately adjacent mid-cingulate cortex, which also displays widespread signal intensity changes during muscle pain. Furthermore, the extent of referral of muscle pain varies between subjects, with some subjects perceiving little or no noticeable referral.

In order to determine whether the referral of muscle pain is reflected in an expansion of the area of activation of SI cortex, we used fMRI to investigate the patterns of signal intensity changes during muscle pain induced by intramuscular injections of hypertonic saline in the leg and arm. We hypothesized that subjects with a large region of perceived muscle pain should have a more extensive activation of SI cortex than subjects with smaller patterns of pain perception.

Methods

Subjects

Twenty-two healthy subjects (11 females, 11 males) aged 22–49 years were recruited for the study. Informed written consent was obtained for all procedures and the study was approved by the Human Research Ethics Committee of the University of New South Wales.

Stimulus and Psychophysics

With each subject lying in a supine position on the scanner bed, a fine needle (23 gauge), connected via a 10 cm cannula to a 1-ml syringe filled with sterile hypertonic (5%) saline, was placed into the right leg: ∼1 cm into the rostral belly of the TA muscle; and/or the right arm: ∼1 cm into the rostral belly of the flexor carpi radialis (FCR) muscle. During scanning, an intramuscular (0.5 ml) injection was administered. Two subjects received both the arm and leg muscle injections during a single scanning session with the injections separated by at least 20 min. In the remaining 20 subjects, injections were made into 1) TA and FCR in separate scanning sessions separated by about one month (n = 9), 2) TA only (n = 9), or 3) FCR only (n = 2). Following a hypertonic saline injection, subjects pressed a buzzer at the onset of pain, just after the peak intensity had been reached and when the pain had ceased. After each pain stimulus, the McGill Pain Questionnaire was read aloud and the subject selected those words, which described the pain. In addition, each subject rated the maximal intensity of the pain on a modified Borg et al. (1981) scale. The Borg scale was as follows: 0 = infinitely small amount of pain, 0.5 = just noticeable pain, 1 = minimal pain, 2 = mild pain, 3 = moderate pain, 4 = considerable pain, 5 = large amount of pain, 7 = very large, 10 = extremely large (maximal) pain. Only subjects that rated the arm or leg muscle pain as 3 (moderate pain) or greater were used for analysis (TA; n = 17; FCR; n = 12). After scanning, each subject plotted the profile of the pain intensity over time and using the time points (onset, maximum, and cessation) indicated by the buzzer, a plot of pain intensity over time was created for each subject. These pain intensity time plots were then averaged across subjects to create a mean pain intensity profile. Each subject also drew the distribution of pain during each stimulus on a standard diagram.

Imaging Procedures and Analysis

Blood oxygen level–dependent (BOLD) contrast gradient echo, echo-planar images were collected using a 3-Tesla whole-body scanner (Philips [Foster City, CA] Intera). One hundred fMRI image volumes were collected continuously over a 5-min period. Each volume covered the entire brain and consisted of 32 axial slices (time repetition = 3 s, time echo = 50 ms, flip angle = 90°, field of view = 220 mm, interslice gap = 0.4 mm, raw voxel size = 1.72 × 1.95 × 4 mm thick). One minute into the scanning period an intramuscular injection was administered. A 3D, T1-weighted anatomical image (voxel size = 0.4 × 0.4 × 0.9 mm) was also collected.

MRI images were analyzed using statistical parametric mapping 2 (SPM2) (Friston et al. 1995). The first 3 image sets were discarded due to nonequilibrium effects. The remaining 97 image volumes were motion corrected, spatially normalized to the Montreal Neurological Institute (MNI) coordinate system, resampled to 1.5 × 1.5 × 4 mm and then segmented into gray matter, white matter, and cerebrospinal fluid, and only the gray matter images were used for further analysis. The gray image sets were then spatially smoothed (5-mm full width at half maximum) and their intensity was normalized. The mean pain intensity profile created from the individual subject's pain intensity profiles was convolved with the canonical hemodynamic response function and then used to search for signal intensity changes. Significant increases and decreases in signal intensity (random effects; uncorrected P < 0.01) were assessed for the arm and leg muscle pain groups. To reduce the risk of false positive activations, a minimum cluster threshold of 10 voxels was used. A subset of the data collected during leg muscle pain has been used in a previous investigation (Henderson, Bandler, et al. 2006).

To assess the effects of pain referral, subjects were divided into those in whom the perceived pain radiated peripherally and those in whom the pain was limited to a diffuse area around the muscle belly. For the leg pain, subjects in whom the perceived pain spread into the ankle and onto the dorsum of the foot were placed into the “TA-referred” group (n = 10) and those in whom the pain was confined to the lower leg were placed into the “TA-nonreferred” group (n = 7). Similarly, for the arm pain, subjects in whom the perceived pain spread into the wrist and into the hand were placed into the “FCR-referred” group (n = 6) and those in whom the pain was confined to the forearm were placed into the “FCR-nonreferred” group (n = 6). Regional signal intensity increases and decreases in each of the TA referred and nonreferred groups and the FCR-referred and -nonreferred groups were determined using a random effects analysis (P < 0.01, uncorrected). To determine regions that responded in a similar fashion, the map of significant signal intensity changes in the TA referred group was used to mask the significant signal intensity change map in the TA-nonreferred group. A similar procedure was performed for the 2 FCR groups. Regions in which signal intensity in the referred groups were significantly greater than the nonreferred groups were also determined (random effects; uncorrected P < 0.01). The percentage changes in signal intensity relative to baseline for each significant voxel in a cluster were calculated over time to form individual voxel time trends. These were then averaged across a cluster to form an average cluster time trend and the cluster time trends from each subject were then averaged to create a mean (±standard error of the mean [SEM]) group cluster time trend. The significance of these trends was verified using a repeated measures analysis of variance analysis (P < 0.05).

The relationship between S1 signal intensity and area of perceived spread of pain was then determined by comparing the maximum signal intensity changes in each significant S1 cluster with the area of perceived pain. Each subject's pain distribution drawing was digitized and the area of perceived pain calculated using ImageJ software (National Institutes of Health, USA). Each subject's total pain area was then expressed as a percentage and plotted against the maximum pain intensity for each of the significant S1 clusters. Significant relationships were then determined using a regression analysis. To illustrate further individual subjects S1 signal intensity changes, a volume of interest (VOI) analysis of the S1 cortex was performed. Using the S1 template provided by SPM2, the S1 cortex was divided into multiple VOIs based on Z-levels. The signal intensity change in each of the S1 VOIs was then calculated for each individual subject.

Results

Pain Perception

Intramuscular injection of hypertonic saline (5%) evoked pain after a few seconds, reaching maximal intensity within 50–60 s and subsiding to preinjection levels after ∼5 min. The most commonly chosen words to describe the muscle pain were “aching” and “cramping.” In all subjects the pain was diffuse and, despite the small volume of the injectate (0.5 ml), encompassed the entire region of the muscle belly. In addition to this “primary pain,” in about half of the subjects the pain was described as spreading distally to encompass the wrist and hand following injection into FCR (arm), or into the ankle and dorsum of the foot following injection into TA (leg). This information was volunteered by the subjects—they were asked to describe where they felt the pain but were not specifically asked whether they felt it projecting distally. The 2 patterns of pain localization for the leg and arm are illustrated in Figure 1. The distally projecting pain will be termed “referred pain.” Following each experiment, subjects were separated into 1 of 2 groups: those who reported referred pain, and those who did not. There were no statistically significant differences in the maximal pain intensity ratings between the “referred” and “nonreferred” groups (P > 0.05; t-test), for pain evoked by intramuscular injection either into the arm (mean ± SEM 6.3 ± 2 for the referred group, 7.5 ± 1 for the nonreferred group) or into the leg (7.0 ± 2 vs. 5.6 ± 2). It should be noted that subjects were asked only to describe the overall pain, not to differentiate between the relative intensities at the primary and referred sites.

Figure 1.

Line drawings of the leg and arm illustrating the extent of perceived pain in each subject following hypertonic saline injection into either the right FCR or TA muscle. Subjects were divided into 2 groups: those in which the pain was referred into the hand or foot (“referred”) and those in which the pain was not (“nonreferred”).

Figure 1.

Line drawings of the leg and arm illustrating the extent of perceived pain in each subject following hypertonic saline injection into either the right FCR or TA muscle. Subjects were divided into 2 groups: those in which the pain was referred into the hand or foot (“referred”) and those in which the pain was not (“nonreferred”).

fMRI Signal Changes

The initial analyses were performed on all subjects, irrespective of whether or not they reported referred pain, to confirm that our earlier observations obtained for muscle pain in the leg (Henderson, Bandler, et al. 2006) held also for the arm. In all subjects significant changes in fMRI signal intensity were observed in multiple, yet discrete areas of the brain following intramuscular injection of hypertonic saline into either the arm or leg. Increases in signal intensity were found in the cerebellum, the posterior and anterior insula bilaterally, the mid-cingulate cortex, the somatotopically appropriate area of the contralateral SI cortex, and the contralateral SII cortex (Fig. 2). Robust decreases in signal intensity occurred in the perigenual anterior cingulate cortex (pACC), confirming our earlier findings (Henderson, Bandler, et al. 2006).

Figure 2.

Significant fMRI signal intensity changes correlated to mean pain intensity profile during forearm and leg muscle pain overlaid onto an individual T1-weighted anatomical image set. The hot and cold color scales (coded for t-value) indicate regional signal intensity increases and decreases, respectively. Slice positions are indicated by MNI coordinates at the bottom right of each image. MCC, mid-cingulate cortex; pACC, perigenual anterior cingulate cortex.

Figure 2.

Significant fMRI signal intensity changes correlated to mean pain intensity profile during forearm and leg muscle pain overlaid onto an individual T1-weighted anatomical image set. The hot and cold color scales (coded for t-value) indicate regional signal intensity increases and decreases, respectively. Slice positions are indicated by MNI coordinates at the bottom right of each image. MCC, mid-cingulate cortex; pACC, perigenual anterior cingulate cortex.

Referred versus Nonreferred

Separation of the subjects into 2 groups, those who reported referred pain in the foot (for injections into the leg) or hand (for injections into the arm) and those who did not, revealed similarities but also widespread differences in BOLD signal changes for the groups. Such differences were masked by pooling the results of the 2 groups in the initial analysis.

Similarities

Irrespective of whether the muscle pain originated in the arm or leg, robust increases in signal intensity occurred in both the referred and nonreferred groups in the mid-cingulate, anterior and posterior insula and secondary somatosensory cortex (SII) bilaterally. This is illustrated by the red shading in Figure 3.

Figure 3.

Significant fMRI signal intensity changes during arm and leg pain in the referred and nonreferred groups. Significant signal intensity increases and decreases in both the referred and nonreferred groups are indicated by the red and blue shadings, respectively. Significant differences are indicated by the yellow shading. The mean percent (±SEM) change in signal intensity over time for a selection of significant clusters is also shown (light green, referred; dark green, nonreferred). The vertical dashed line indicates the onset of the hypertonic saline injection. Slice positions are indicated by MNI coordinates at the bottom left of the lower row of images. MCC, mid-cingulate cortex; pACC, perigenual anterior cingulate cortex.

Figure 3.

Significant fMRI signal intensity changes during arm and leg pain in the referred and nonreferred groups. Significant signal intensity increases and decreases in both the referred and nonreferred groups are indicated by the red and blue shadings, respectively. Significant differences are indicated by the yellow shading. The mean percent (±SEM) change in signal intensity over time for a selection of significant clusters is also shown (light green, referred; dark green, nonreferred). The vertical dashed line indicates the onset of the hypertonic saline injection. Slice positions are indicated by MNI coordinates at the bottom left of the lower row of images. MCC, mid-cingulate cortex; pACC, perigenual anterior cingulate cortex.

Differences

For both arm and leg muscle pain, significant signal intensity differences between the referred and nonreferred pain groups occurred in the SI, pACC, and cerebellar cortices. During arm pain only, a region of significant difference also occurred in the ipsilateral anterior insula.

Within the SI cortex, the signal intensity changes appeared to correspond to the spread of pain described by each subject. That is, during leg muscle pain, in both the referred and nonreferred groups, signal intensity increased in the region of SI representing the leg and increased in the immediately adjacent region (representing the foot) in those subjects who reported a pain referral into the ankle and/or foot. Similarly, during arm muscle pain both the referred and nonreferred groups displayed an increase in signal intensity in the region of SI representing the forearm and, in subjects with pain referred into the wrist and hand, an increase, which spread laterally into the region of cortex representing the hand (Fig. 4). In 2 subjects, the correspondence of the anatomical areas was confirmed by recording significant increases in signal intensity following a 2-min period of light stroking of the hand or dorsum of the foot with a brush, relative to a 2-min baseline period.

Figure 4.

Significant fMRI signal intensity changes during arm and leg pain in the referred and nonreferred groups. Significant signal intensity increases in both the referred and nonreferred groups are indicated by the red shading and significant differences by the yellow shading. The mean percent (±SEM) change in signal intensity over time for significant clusters in the SI is shown to the right (light green, referred; dark green, nonreferred). The vertical dashed line indicates the onset of the hypertonic saline injection.

Figure 4.

Significant fMRI signal intensity changes during arm and leg pain in the referred and nonreferred groups. Significant signal intensity increases in both the referred and nonreferred groups are indicated by the red shading and significant differences by the yellow shading. The mean percent (±SEM) change in signal intensity over time for significant clusters in the SI is shown to the right (light green, referred; dark green, nonreferred). The vertical dashed line indicates the onset of the hypertonic saline injection.

An examination of maximum signal intensity changes compared with the perceived spread of pain revealed a significant positive correlation between signal intensity and pain area in the S1 clusters, which displayed significant differences between the referred and nonreferred groups (Fig. 5). Furthermore, there was no significant correlation between signal intensity and pain area in the S1 clusters, which displayed significant signal intensity increases in both the referred and nonreferred groups. Plots of individual subject's S1 signal intensity changes further emphasize the greater spread of S1 cortex activation in subjects with larger pain referral areas (Fig. 6).

Figure 5.

Maximum fMRI signal intensity changes in the SI plotted against area of perceived pain for each individual subject. The overlay on the left indicates the significant cluster from which the signal intensity changes were calculated. Note that in the region of SI in which both the referred and nonreferred groups displayed similar signal intensity increases (red) there was no significant correlation between signal intensity and area of perceived pain. However, in the SI regions in which the referred and nonreferred groups were significantly different (yellow), there was a significant positive correlation between the maximum signal intensity change and area of perceived pain.

Figure 5.

Maximum fMRI signal intensity changes in the SI plotted against area of perceived pain for each individual subject. The overlay on the left indicates the significant cluster from which the signal intensity changes were calculated. Note that in the region of SI in which both the referred and nonreferred groups displayed similar signal intensity increases (red) there was no significant correlation between signal intensity and area of perceived pain. However, in the SI regions in which the referred and nonreferred groups were significantly different (yellow), there was a significant positive correlation between the maximum signal intensity change and area of perceived pain.

Figure 6.

Individual subject signal intensity changes plotted over time, in 5 VOI's in the SI (A–E) during leg and arm muscle pain. The location of the VOIs is indicated to the top right of each panel. The area of perceived pain is indicated to the left of each set of 5 time trends taken from VOIs (A–E). The vertical dashed line indicates the onset of the hypertonic saline injection. Note that in the subjects in whom the area of perceived pain is greater, the number of VOIs in which signal intensity increases is greater.

Figure 6.

Individual subject signal intensity changes plotted over time, in 5 VOI's in the SI (A–E) during leg and arm muscle pain. The location of the VOIs is indicated to the top right of each panel. The area of perceived pain is indicated to the left of each set of 5 time trends taken from VOIs (A–E). The vertical dashed line indicates the onset of the hypertonic saline injection. Note that in the subjects in whom the area of perceived pain is greater, the number of VOIs in which signal intensity increases is greater.

A similar pattern also occurred in the ipsilateral insula following FCR injections. That is, a region of signal intensity increase in both the referred and nonreferred groups and a discrete region immediately lateral in which signal intensity increased in the referred group only (Fig. 3). In contrast to these signal intensity increases, during both forearm and leg pain, both the referred and nonreferred groups displayed profound signal intensity decreases in the pACC. In addition, in the nonreferred groups only, this region of signal decrease extended into the immediately adjacent medial prefrontal cortex, whereas in the referred group signal intensity increased slightly in this region of prefrontal cortex (Fig. 7). Finally, within the cerebellar cortex, discrete clusters occurred in which signal intensity increased in those subjects reporting referred pain in the hand or foot but decreased in those subjects who did not report referred pain.

Figure 7.

Significant fMRI signal intensity changes in the cingulate cortex during arm and leg pain in the referred and nonreferred groups. Significant signal intensity increases and decreases in both the referred and nonreferred groups are indicated by the red and blue shadings, respectively. Significant differences are indicated by the yellow shading. The mean percent (±SEM) change in signal intensity over time for significant clusters in the perigenual cingulate cortex is shown below (light green, referred; dark green, nonreferred). The vertical dashed line indicates the onset of the hypertonic saline injection. Slice positions are indicated by MNI coordinates at the top left of each image.

Figure 7.

Significant fMRI signal intensity changes in the cingulate cortex during arm and leg pain in the referred and nonreferred groups. Significant signal intensity increases and decreases in both the referred and nonreferred groups are indicated by the red and blue shadings, respectively. Significant differences are indicated by the yellow shading. The mean percent (±SEM) change in signal intensity over time for significant clusters in the perigenual cingulate cortex is shown below (light green, referred; dark green, nonreferred). The vertical dashed line indicates the onset of the hypertonic saline injection. Slice positions are indicated by MNI coordinates at the top left of each image.

Discussion

Using whole-brain fMRI we have confirmed that the deep pain evoked by intramuscular injection of hypertonic saline into the FCR muscle in the forearm is associated with activation of the same pain neuromatrix engaged by intramuscular injection into the TA muscle in the leg, as demonstrated previously (Henderson, Bandler, et al. 2006). In addition, we have demonstrated that somatotopy exists within the SI for deep pain originating in the leg and forearm. More importantly, we have shown for the first time that referred pain is reflected in a widespread increase in neural activity in several discrete cerebral areas, including SI, perigenual anterior cingulate, ipsilateral anterior insular, and cerebellar cortices.

Methodological Considerations

Intramuscular injection of hypertonic saline is a well-established model of deep pain (Graven-Nielsen et al. 1997b) but this is only the second study to have employed fMRI to investigate central processing of pain using this technique. An intramuscular cannula was inserted into the belly of the FCR or TA muscle prior to the scanning session and subjects knew that they would receive an injection that would cause pain. They were also informed that it would subside after a few minutes. And although the expectancy of painful (Ploghaus et al. 1999) or other unpleasant stimuli causes brain activations in its own right, subjects did not anticipate feeling pain that projected into the hand or foot nor did they receive any suggestion prior to the experiment that this could occur. We believe we are recording indicators of cortical neural activity that reflect the perception of referred pain in an area remote from the site of noxious stimulation. It has been shown that intramuscular injection of 0.5 ml of saline into the TA muscle forms a stable pool, diffusing only ∼2 cm from the injection site (Graven-Nielsen, McArdle et al. 1997). Accordingly, we are confident that any referred pain did not result from the spread of hypertonic saline into distal structures. Moreover, the connective tissue sleeve around the muscle would limit spread beyond the muscle into the ankle and foot. Interestingly, it has been shown that intramuscular injection of hypertonic saline into the soleus muscle causes a dull ache that is limited to the small area of the injection site—subjects do not report any projection of the pain into the foot nor do they report an ache that is perceived as extending along the entire muscle belly (Weerakkody et al. 2003). This would suggest that the large area of primary pain emanating from TA (and FCR) following injection of hypertonic saline into this muscle is, of itself, a reflection of the processing of the primary noxious input to the central nervous system—processing occurring within the spinal cord and/or within the brain.

Pain Referred from Deep Somatic Tissues

The pain that projects distally from the primary site of noxious stimulation within muscle has been studied extensively (Arendt-Nielsen et al. 1997; Graven-Nielsen et al. 1997a; Arendt-Nielsen and Svensson 2001). Typically, that evoked by injection of hypertonic saline into the TA muscle refers to the ankle and dorsum of the foot and, like the area of primary pain in the muscle belly, shows both temporal and spatial summation from multiple injections into the same or separate intramuscular sites (Arendt-Nielsen et al. 1997; Graven-Nielsen et al. 1997a). Similar but more expansive sites of projection have been found following injection into the distal tendon or, especially, the proximal bone–tendon junction of TA (Gibson et al. 2006). Referred pain has also been reported following injection of hypertonic saline into the infrapatellar fat pad, though this usually projects proximally, towards the groin (Bennell et al. 2004).

Representation of Referred Pain in the Primary and Secondary Somatosensory Cortices

It has been widely suggested that pain is represented in 2 main systems: a lateral system which includes the lateral thalamus, SI, and SII cortices and, in some cases, the insula, and a medial system which includes the medial thalamus, anterior cingulate, and prefrontal cortices. It is thought that the lateral system mediates sensory-discriminatory aspects and is somatotopically organized, whereas the medial system mediates the affective component of the pain experience and is not somatotopically organized (Ohara et al. 2005). Although there is no doubt SI cortex is crucial for the ability to localize somatosensory stimuli accurately, there is debate as to whether it is also involved in the localization of noxious stimuli. Electrical stimulation of SI rarely evokes painful sensations (Mazzola et al. 2005) or alters pain perception (Head and Holmes 1911). A recent magnetoencephalography study has shown that brief laser-induced heating of the skin evokes a sensation of first pain, mediated by A-delta fibers and engaging SI and SII cortices, and a C-fiber–mediated sensation of second pain that primarily recruits the ACC and SII (Ploner et al. 2002). However, we and others have clearly shown somatotopic activation of SI during noxious stimulation (Bingel et al. 2004; Henderson, Bandler, et al. 2006). Furthermore, our data suggest that the perception of referred pain is represented in the SI cortex. That is, we found that signal intensity increased in the foot area of SI in subjects who reported referred pain in the foot and in the hand area for those subjects who perceived referred pain in the hand. We did not however, find a similar pattern of signal intensity change in the SII cortex. It may be that SII is not critical for the localization of painful stimuli or alternatively, that our data are of too low a resolution to differentiate the fine somatotopic organization of the SII cortex.

Representation of Referred Pain in the Cingulate and Insular Cortices

Similar to SI, during forearm pain we found a pattern of signal intensity change in the ipsilateral anterior insula, which mimicked the pattern of referred pain. We have recently shown that the right anterior insula displays a somatotopically organized pattern of fMRI signal increase during noxious muscle and cutaneous stimulation (Henderson, Gandevia, et al. 2006). It has been suggested that noxious information is transmitted from the dorsal horn of the spinal cord to a pain-specific region of the thalamus—the posterior division of the ventral medial thalamic nucleus (VMpo). Although the existence of the VMpo in humans is still a matter for debate, Craig (2003) has proposed that somatotopically organized noxious information is transmitted from the VMpo to the contralateral posterior insula and finally to the ipsilateral anterior insular cortices. The results of direct electrical stimulation, brain imaging, and single-unit recording investigation have confirmed that the insula is involved in the processing of noxious information (Robinson and Burton 1980; Peyron et al. 2000; Ostrowsky et al. 2002; Mazzola et al. 2005). Magneto- and electroencephalography studies also indicate that SI and the operculo-insula regions are the earliest activated structures following noxious stimulation (Ploner et al. 1999; Frot and Mauguiere 2003). Our data show that, like SI, the right anterior insula displayed a pattern of signal intensity increase, which mirrored the perceived extent of referred muscle pain: subjects who reported significant levels of referred pain displayed an increase in signal intensity in part of the right anterior insula in which no changes occurred in subjects who did not report referred pain. The larger area of insular activation in those subjects reporting referred pain may, like SI, reflect the ability to localize noxious stimuli.

In contrast, it has been reported that lesions encompassing large areas of the insula in humans can result in asymbolia for pain—a condition in which patients can recognize both superficial and deep painful stimuli and distinguish their quality (sharp, dull, etc.) but cannot respond to painful stimuli with either motor withdrawal, grimacing or appropriate emotional responses (Berthier et al. 1988). It is well established that the anterior insula is recruited during emotionally charged challenges. It is activated primarily by negative emotional states such as internally generated sadness, anger, fear, frustration, and disgust, and it has been recently reported to display activity changes that are significantly correlated with the level of unpleasantness associated with visceral and muscle pain (Phillips et al. 1997; Phan, Wager et al. 2004; Abler et al. 2005; Dunckley et al. 2005; Schreckenberger et al. 2005). Furthermore, Phan, Taylor et al. (2004) report that during the appraisal of self-relatedness of emotionally charged stimuli, fMRI signal intensity increases in the same region of the right anterior insula to that reported here during muscle pain. It has been suggested that the insula monitors the body's internal state and acts as an alarm center, with the anterior insula in particular being engaged during recall of negative emotions (Reiman et al. 1997). The “somatic marker” hypothesis suggests that the anterior insula may integrate somatic and external cues of emotional relevance (Damasio 1996). The right anterior insula may integrate somatotopically organized information from muscle nociceptors with the personal relevance of this emotionally charged stimulus, helping to “decide” on an appropriate course of action. In this context, location of the noxious stimulus would be likely to be advantageous in deciding on an appropriate behavioral response.

Surprisingly, the pACC, a region, which has also been implicated in the processing of emotional state changes, also displayed signal intensity change patterns that appeared to reflect the perceived spread of pain. Signal intensity decreases within the pACC during muscle and visceral pain have been reported previously (Dunckley et al. 2005; Henderson, Bandler, et al. 2006) and we have speculated that the decreases seen in this region during muscle pain may reflect the negative affect associated with pain; the pACC is part of the affective division of the cingulate cortex (Devinsky et al. 1995; Bush et al. 2000). Lesions of the pACC result in emotional lability (Hornak et al. 2003), and patients with depression often display decreased pACC metabolism (Drevets et al. 1997; Mayberg et al. 1997). In contrast to the S1 and anterior insular cortices, we found that subjects with a larger referral pattern displayed decreases in single intensity over a smaller area than those subjects who reported little referral. It has been shown previously that the extent of muscle pain referral increases as the perceived intensity of pain increases an that pain intensity is positively correlated with pain unpleasantness (Graven-Nielsen et al. 1997b). the data in the present study do not support the idea that the extent of referral is correlated to perceived intensity—there were no differences in the magnitude of pain between the referred and non-referred groups. however, we did not ask each subject to rate the unpleasantness of the pain. As the extent of pain referral did not correlate with a greater signal decline in the pACC region, we suggest that the right anterior insula, and pACC are involved in different aspects of the emotional response to painful stimuli. Indeed, the sensation of unpleasantness has been ascribed largely to the insula, and the unpleasantness of visceral as well as muscle stimuli involves signal intensity reductions in the pACC and increases in the right insula (Dunckley et al. 2005; Schreckenberger et al. 2005).

Representation of Referred Pain in the Cerebellum

The roles of the cerebellum in the processing of pain have recently been documented (Bingel et al. 2002; Helmchen et al. 2003; Saab and Willis 2003), though any perceptual role in nociception has largely been ignored. In our earlier study, in which we injected hypertonic saline into TA (Henderson, Bandler, et al. 2006), we saw no consistent changes in signal intensity within the cerebellum, but given the divergent responses in the “referrers” and “nonreferrers” we observed in the present study this is perhaps not surprising. A similar lack of cerebellar activation was observed during intramuscular noxious electrical stimulation (Niddam et al. 2002). For pain induced in either the leg or forearm we saw consistent increases in signal intensity within cerebellar cortex in those subjects reporting pain in the ankle/foot or wrist/hand, and decreases in those subjects who did not. As noted above, there were no differences in the perceived intensity of pain in the 2 groups so the disparate responses in cerebellar cortex may reflect subperceptual sensory processing.

This is the first study to have specifically tested whether referred pain reflects an increase in neural activity in the cerebral cortex. We have provided evidence of cortical differentiation in the processing of primary and referred pain following discrete noxious stimuli originating in muscle, with widespread yet discrete differences having been observed in the primary sensory, cingulate, anterior insula, and cerebellar cortices.

We thank Kathy M. Hughes for help in all imaging procedures and Dr Paul Macey, University of California, Los Angeles, for help with image analysis. Scans were performed at the Mayne Clinical Research Imaging Centre. This work was supported by National Health and Medical Research Council grant 350889 to V.M. and L.H. Conflict of Interest: None declared.

References

Abler
B
Walter
H
Erk
S
Neural correlates of frustration
Neuroreport
 , 
2005
, vol. 
16
 (pg. 
669
-
672
)
Arendt-Nielsen
L
Graven-Nielsen
T
Svensson
P
Jensen
TS
Temporal summation in muscles and referred pain areas: an experimental human study
Muscle Nerve
 , 
1997
, vol. 
20
 (pg. 
1311
-
1313
)
Arendt-Nielsen
L
Svensson
P
Referred muscle pain: basic and clinical findings
Clin J Pain
 , 
2001
, vol. 
17
 (pg. 
11
-
19
)
Bennell
K
Hodges
P
Mellor
R
Bexander
C
Souvlis
T
The nature of anterior knee pain following injection of hypertonic saline into the infrapatellar fat pad
J Orthop Res
 , 
2004
, vol. 
22
 (pg. 
116
-
121
)
Berthier
M
Starkstein
S
Leiguarda
R
Asymbolia for pain: a sensory-limbic disconnection syndrome
Ann Neurol
 , 
1988
, vol. 
24
 (pg. 
41
-
49
)
Bingel
U
Lorenz
J
Glauche
V
Knab
R
Glascher
J
Weiller
C
Buchel
C
Somatotopic organization of human somatosensory cortices for pain: a single trial fMRI study
Neuroimage
 , 
2004
, vol. 
23
 (pg. 
224
-
232
)
Bingel
U
Quante
M
Knab
R
Bromm
B
Weiller
C
Buchel
C
Subcortical structures involved in pain processing: evidence from single-trial fMRI
Pain
 , 
2002
, vol. 
99
 (pg. 
313
-
321
)
Borg
G
Holmgren
A
Lindblad
I
Quantitative evaluation of chest pain
Acta Med Scand Suppl
 , 
1981
, vol. 
644
 (pg. 
43
-
45
)
Bush
G
Luu
P
Posner
MI
Cognitive and emotional influences in anterior cingulate cortex
Trends Cogn Sci
 , 
2000
, vol. 
4
 (pg. 
215
-
222
)
Craig
AD
Pain mechanisms: labeled lines versus convergence in central processing
Annu Rev Neurosci
 , 
2003
, vol. 
26
 (pg. 
1
-
30
)
Damasio
AR
The somatic marker hypothesis and the possible functions of the prefrontal cortex
Philos Trans R Soc Lond B Biol Sci
 , 
1996
, vol. 
351
 (pg. 
1413
-
1420
)
Devinsky
O
Morrell
MJ
Vogt
BA
Contributions of anterior cingulate cortex to behaviour
Brain
 , 
1995
, vol. 
118
 (pg. 
279
-
306
)
Drevets
WC
Price
JL
Simpson
JR
Jr
Todd
RD
Reich
T
Vannier
M
Raichle
ME
Subgenual prefrontal cortex abnormalities in mood disorders [see comment]
Nature
 , 
1997
, vol. 
386
 (pg. 
824
-
827
)
Dunckley
P
Wise
RG
Aziz
Q
Painter
D
Brooks
J
Tracey
I
Chang
L
Cortical processing of visceral and somatic stimulation: differentiating pain intensity from unpleasantness
Neuroscience
 , 
2005
, vol. 
13
 pg. 
3
 
Feinstein
B
Langton
JNK
Jameson
RM
Schiller
F
Experiments on pain referred from deep somatic tissues
J Bone Joint Surg
 , 
1954
, vol. 
36
 (pg. 
981
-
997
)
Foreman
RD
Blair
RW
Weber
RN
Viscerosomatic convergence onto T2-T4 spinoreticular, spinoreticular-spinothalamic, and spinothalamic tract neurons in the cat
Exp Neurol
 , 
1984
, vol. 
85
 (pg. 
597
-
619
)
Friston
KJ
Holmes
AP
Worsley
K
Poline
J-B
Frith
CD
Frackowiak
RS
Statistic parametric maps in functional imaging: a general linear approach
Hum Brain Mapp
 , 
1995
, vol. 
2
 (pg. 
189
-
210
)
Frot
M
Mauguiere
F
Dual representation of pain in the operculo-insular cortex in humans
Brain
 , 
2003
, vol. 
126
 (pg. 
438
-
450
)
Gibson
W
Arendt-Nielsen
L
Graven-Nielsen
T
Referred pain and hyperalgesia in human tendon and muscle belly tissue
Pain
 , 
2006
, vol. 
120
 (pg. 
113
-
123
)
Graven-Nielsen
T
Arendt-Nielsen
L
Svensson
P
Jensen
TS
Quantification of local and referred muscle pain in humans after sequential i.m. injections of hypertonic saline
Pain
 , 
1997
, vol. 
69
 (pg. 
111
-
117
)
Graven-Nielsen
T
Arendt-Nielsen
L
Svensson
P
Jensen
TS
Stimulus-response functions in areas with experimentally induced referred muscle pain—a psychophysical study
Brain Res
 , 
1997
, vol. 
744
 (pg. 
121
-
128
)
Graven-Nielsen
T
McArdle
A
Phoenix
J
Arendt-Nielsen
L
Jensen
TS
Jackson
MJ
Edwards
RH
In vivo model of muscle pain: quantification of intramuscular chemical, electrical, and pressure changes associated with saline-induced muscle pain in humans
Pain
 , 
1997
, vol. 
69
 (pg. 
137
-
143
)
Head
H
Holmes
G
Sensory disturbances from cerebral lesions
Brain
 , 
1911
, vol. 
34
 (pg. 
102
-
254
)
Helmchen
C
Mohr
C
Erdmann
C
Petersen
D
Nitschke
MF
Differential cerebellar activation related to perceived pain intensity during noxious thermal stimulation in humans: a functional magnetic resonance imaging study
Neurosci Lett
 , 
2003
, vol. 
335
 (pg. 
202
-
206
)
Henderson
LA
Bandler
R
Gandevia
SC
Macefield
VG
Distinct forebrain activity patterns during deep versus superficial pain
Pain
 , 
2006
, vol. 
120
 (pg. 
286
-
296
)
Henderson
LA
Gandevia
SC
Macefield
VG
Forthcoming
Somatotopic organization of the processing of muscle and cutaneous pain in the left and right insula cortex: a single-trial fMRI study
Pain
 , 
2006
 
doi: 10.1016/j.pain.2006.08.013
Hoheisel
U
Mense
S
Response behaviour of cat dorsal horn neurones receiving input from skeletal muscle and other deep somatic tissues
J Physiol.
 , 
1990
, vol. 
426
 (pg. 
265
-
280
)
Hornak
J
Bramham
J
Rolls
ET
Morris
RG
O'Doherty
J
Bullock
PR
Polkey
CE
Changes in emotion after circumscribed surgical lesions of the orbitofrontal and cingulate cortices
Brain
 , 
2003
, vol. 
126
 (pg. 
1691
-
1712
)
Kellgren
JH
Observation on referred pain arising from muscle
Clin Sci
 , 
1938
, vol. 
3
 (pg. 
175
-
190
)
Mayberg
HS
Brannan
SK
Mahurin
RK
Jerabek
PA
Brickman
JS
Tekell
JL
Silva
JA
McGinnis
S
Glass
TG
Martin
CC
, et al.  . 
Cingulate function in depression: a potential predictor of treatment response [see comment]
Neuroreport
 , 
1997
, vol. 
8
 (pg. 
1057
-
1061
)
Mazzola
L
Isnard
J
Mauguiere
F
Somatosensory and pain responses to stimulation of the second somatosensory area (SII) in humans. A comparison with SI and insular responses
Cereb Cortex
 , 
2005
, vol. 
16
 
7
(pg. 
960
-
968
)
Niddam
DM
Yeh
TC
Wu
YT
Lee
PL
Ho
LT
Arendt-Nielsen
L
Chen
AC
Hsieh
JC
Event-related functional MRI study on central representation of acute muscle pain induced by electrical stimulation
Neuroimage
 , 
2002
, vol. 
17
 (pg. 
1437
-
1450
)
Ohara
PT
Vit
JP
Jasmin
L
Cortical modulation of pain
Cell Mol Life Sci
 , 
2005
, vol. 
62
 (pg. 
44
-
52
)
Ostrowsky
K
Magnin
M
Ryvlin
P
Isnard
J
Guenot
M
Mauguiere
F
Representation of pain and somatic sensation in the human insula: a study of responses to direct electrical cortical stimulation
Cereb Cortex
 , 
2002
, vol. 
12
 (pg. 
376
-
385
)
Peyron
R
Laurent
B
Garcia-Larrea
L
Functional imaging of brain responses to pain. A review and meta-analysis (2000)
Neurophysiol Clin
 , 
2000
, vol. 
30
 (pg. 
263
-
288
)
Phan
KL
Taylor
SF
Welsh
RC
Ho
SH
Britton
JC
Liberzon
I
Neural correlates of individual ratings of emotional salience: a trial-related fMRI study
Neuroimage
 , 
2004
, vol. 
21
 (pg. 
768
-
780
)
Phan
KL
Wager
TD
Taylor
SF
Liberzon
I
Functional neuroimaging studies of human emotions
CNS Spectr
 , 
2004
, vol. 
9
 (pg. 
258
-
266
)
Phillips
ML
Young
AW
Senior
C
Brammer
M
Andrew
C
Calder
AJ
Bullmore
ET
Perrett
DI
Rowland
D
Williams
SC
, et al.  . 
A specific neural substrate for perceiving facial expressions of disgust
Nature
 , 
1997
, vol. 
389
 (pg. 
495
-
498
)
Ploghaus
A
Tracey
I
Gati
JS
Clare
S
Menon
RS
Matthews
PM
Rawlins
JN
Dissociating pain from its anticipation in the human brain
Science
 , 
1999
, vol. 
284
 (pg. 
1979
-
1981
)
Ploner
M
Gross
J
Timmermann
L
Schnitzler
A
Cortical representation of first and second pain sensation in humans
Proc Natl Acad Sci USA
 , 
2002
, vol. 
99
 (pg. 
12444
-
12448
)
Ploner
M
Schmitz
F
Freund
HJ
Schnitzler
A
Parallel activation of primary and secondary somatosensory cortices in human pain processing
J Neurophysiol
 , 
1999
, vol. 
81
 (pg. 
3100
-
3104
)
Reiman
EM
Lane
RD
Ahern
GL
Schwartz
GE
Davidson
RJ
Friston
KJ
Yun
LS
Chen
K
Neuroanatomical correlates of externally and internally generated human emotion
Am J Psychiatry
 , 
1997
, vol. 
154
 (pg. 
918
-
925
)
Robinson
CJ
Burton
H
Somatic submodality distribution within the second somatosensory (SII), 7b, retroinsular, postauditory, and granular insular cortical areas of M. fascicularis
J Comp Neurol
 , 
1980
, vol. 
192
 (pg. 
93
-
108
)
Saab
CY
Willis
WD
The cerebellum: organization, functions and its role in nociception
Brain Res Brain Res Rev
 , 
2003
, vol. 
42
 (pg. 
85
-
95
)
Schreckenberger
M
Siessmeier
T
Viertmann
A
Landvogt
C
Buchholz
HG
Rolke
R
Treede
RD
Bartenstein
P
Birklein
F
The unpleasantness of tonic pain is encoded by the insular cortex
Neurology
 , 
2005
, vol. 
64
 (pg. 
1175
-
1183
)
Weerakkody
NS
Percival
P
Hickey
MW
Morgan
DL
Gregory
JE
Canny
BJ
Proske
U
Effects of local pressure and vibration on muscle pain from eccentric exercise and hypertonic saline
Pain
 , 
2003
, vol. 
105
 (pg. 
425
-
435
)