Pain processing during three levels of noxious stimulation... : PAIN (original) (raw)
1. Introduction
Changes in regional cerebral blood flow (rCBF) have been demonstrated in a number of cortical and subcortical regions in response to an experimental noxious stimulus, or during clinical pain. The central region most commonly activated is the anterior cingulate cortex (ACC), present in 12 of 14 recently published experiments using positron emission tomography (PET) and rCBF to measure responses to noxious stimulation or clinical pain (Jones et al., 1991; Talbot et al., 1991; Casey et al., 1994; Coghill et al., 1994; Derbyshire et al., 1994; Rosen et al., 1994; Hsieh et al., 1995a; Hsieh et al., 1995b; Weiller et al., 1995; Craig et al., 1996; Hsieh et al., 1996; Vogt et al., 1996); Rosen et al. (1994) reported decreased rCBF in area 24 of the ACC and increased rCBF in area 25 of the ACC. The two studies not reporting ACC activity (Di Piero et al., 1991; Iadorola et al., 1995) both involved region of interest (ROI) analysis not including the ACC. Thus the ACC has been shown to be active in all studies which have included the region for analysis.
Increased insula and thalamic activity are the second most commonly reported responses but are noticeably less consistent. Eight of the papers reviewed reported activation of the insula cortex (Casey et al., 1994; Coghill et al., 1994; Derbyshire et al., 1994; Hsieh et al., 1995a; Hsieh et al., 1995b; Craig et al., 1996; Hsieh et al., 1996; Vogt et al., 1996); Iadorola et al. (1995)and Di Piero et al. (1991) did not examine insula responses. Increased thalamic activity has been reported in seven of the 14 recent papers (Jones et al., 1991; Casey et al., 1994; Coghill et al., 1994; Rosen et al., 1994; Craig et al., 1996; Vogt et al., 1996); Talbot et al. (1991) did not report thalamic activation in their original paper but later reported it in a letter (Duncan et al., 1992). Hsieh et al. (1995b), Iadorola et al. (1995)and Di Piero et al. (1991) reported decreased thalamic activation during chronic pain. No obvious methodological variability between the studies accounts for the lack of thalamic response in the remaining four reports (Derbyshire et al., 1994; Hsieh et al., 1995a; Weiller et al., 1995; Hsieh et al., 1996). For example, Derbyshire et al. (1994) used largely the same methodology as Jones et al. (1991); the former reported no thalamic activation, while the latter demonstrated a contralateral response.
Structures other than the ACC, thalamus and insula are far less consistently activated. Six of the 14 papers reviewed reported activity of the prefrontal cortex (Derbyshire et al., 1994; Rosen et al., 1994; Hsieh et al., 1995a; Hsieh et al., 1995b; Hsieh et al., 1996; Vogt et al., 1996) (Iadorola et al., 1995 did not examine prefrontal responses); six reported activity in the lentiform nucleus (Jones et al., 1991; Coghill et al., 1994; Derbyshire et al., 1994; Hsieh et al., 1995a; Craig et al., 1996; Hsieh et al., 1996) (Iadorola et al., 1995and Di Piero et al., 1991 did not examine lentiform responses); five reported activity in the somatosensory cortex (Talbot et al., 1991; Casey et al., 1994; Coghill et al., 1994; Hsieh et al., 1995a; Craig et al., 1996) (Iadorola et al., 1995 did not examine somatosensory responses); five reported changes in the inferior parietal cortex (Derbyshire et al., 1994; Hsieh et al., 1995a; Hsieh et al., 1995b; Hsieh et al., 1996; Vogt et al., 1996) (Hsieh et al., 1996and Vogt et al., 1996 reported decreases in rCBF, while Iadorola et al., 1995and Di Piero et al., 1991 did not examine inferior parietal responses); and four reported changes in the secondary somatosensory cortex (S2) (Talbot et al., 1991; Casey et al., 1994; Coghill et al., 1994; Craig et al., 1996).
What emerges is that there is a matrix of structures including anterior cingulate, prefrontal, insula, inferior parietal and somatosensory cortices, thalamus and lentiform nucleus which are variably activated during pain experience. This variability may be due to technical or biological factors. The type, duration and location of stimulus varies considerably between studies. Patients with migraine, neuropathic pain, idiopathic pain, cancer pain and post stroke pain have all been studied. Experimental stimuli include ethanol injection, hot probes and cold pressor. Each pain stimulus varies with respect to intensity and quality. The analysis techniques employed also vary between laboratories, as do the types of PET scanners and the number of subjects comprising each study. These problems are confounded when the different thresholds for reporting ‘statistically significant’ results are taken into account.
The basis for biological variability can be found in the ‘multidimensional’ nature of pain experience resulting in a ‘biopsychosocial’ model of pain (Waddell, 1987). Within this model, pain is not regarded as merely a physical sensation related to a noxious stimulus or disease, but is seen as a conscious experience modulated by mental, emotional and sensory mechanisms. Pain has been described as a multidimensional phenomena for some time (Melzack and Casey, 1968), and this continues to be reflected by the current International Association for the Study of Pain definition of pain as “‘n unpleasant sensory and emotional experience associated with actual or potential tissue damage, or described in terms of such damage’ (Merskey, 1991).
The understanding of pain as a multidimensional phenomenon suggests that the concept of a specific ‘pain centre’ is unlikely to adequately account for central pain processing (Jones et al., 1991). The inherent plasticity of pain is likely to be reflected in a matrix of neuronal structures rather than in a fixed ‘pain centre’. A ‘neuromatrix’ (Melzack, 1989) of regions, incorporating for instance the anterior cingulate, prefrontal and insula cortices may be involved in the processing of pain without any single region in itself being necessary and sufficient for the pain experience. Rather, each region would contribute a limited and discrete aspect of the final experience (Vogt et al., 1996).
This study is designed to focus on one of the principle sources of methodological and experiential ambiguity, specifically stimulus intensity. Typically, imaging studies have used two levels of stimulation: one non-painful the other painful. These stimuli are either fixed in advance (Talbot et al., 1991; Casey et al., 1994; Coghill et al., 1994; Craig et al., 1996) or are chosen by the subject (Jones et al., 1991; Derbyshire et al., 1994; Vogt et al., 1996) and often there is considerable variability in the final experience. Table 1 illustrates the range of areas reported with differing stimulation techniques and associated intensity ratings.
Areas active in previous studies related to the stimulus intensity used
Clearly there is little discernible relationship between the central regions activated and the intensity of the stimulus used. While some of the variability is likely to be contingent on factors associated with the biopsychosocial model of pain, some of it is also bound to be associated with technical factors such as the increasing sensitivity of PET imaging over the last 6 years. There is no means by which these factors can be disentangled post hoc. The relationship between stimulus intensity, pain rating and rCBF response can only be assessed using the same techniques in the same subject group.
The study reported here uses a laser stimulator which is capable of delivering multiple levels of energy with a high degree of accuracy, thereby producing varied experiences from warmth to moderate pain. Twelve subjects were studied; each subject received six scans corresponding to a warm stimulus, six scans of just painful or pain threshold, six mild pain and six moderate pain. In this way, changes in the neuronal activity as the experience increased from warm to moderate pain were identified.
2. Methods
2.1. Subjects
Data were analysed from 12 normal volunteers (six male, six female) with a mean age of 28.8 (range 21–41). Permission to carry out these studies was obtained from the internal review board and the Radioactive Drug Review Committee. All subjects gave written, informed consent. Subjects had no history of psychiatric illness or substance abuse. All female volunteers had negative serum _β_-human chronic gonadotrophin (pregnancy) test results.
2.2. Design
The responses of all subjects to an intermittent laser pulse which was either experienced as warm, just painful, mildly painful or moderately painful were compared.
The dependent variable investigated was the rCBF response to each 1-min period of phasic laser stimulation measured by PET. This change in rCBF was used as a measure of change in synaptic activity (Fox and Mintun, 1989; Friston and Frackowiak, 1991). rCBF responses to each level of stimulation of the right hand were measured six times in each volunteer.
2.3. Apparatus
The stimulus was produced using a 50-W continuous wave CO2 laser (Synrad) modulated by a computer controlled voltage to give maximum flexibility of pulse parameters. The laser was mounted on a stable base with beam steering optics, a visible pointing and alignment laser, and an articulated arm to deliver the stimulus to the subject.
Analogue power modulation of the laser output was achieved through a control box (Synrad UC-1000 universal controller) which received a 0–10-V signal with 0 V being laser off and 10 V being laser fully on. The 0–10-V control signal was derived from a digital to analogue converter card (DAC) mounted on a personal computer.
Six subjects scans were scanned using a PET scanner; Siemans model 951R/31 (Siemans Medical Systems, Hoffman Estates, IL; 3D axial resolution 5.0 mm) and for the other six subjects scans were obtained with an Automated Rotating PET scanner (ART; 3D axial resolution 5.0 mm).
2.4. Procedure
2.4.1. Stimulus parameters
Prior to the PET studies, laser energies which corresponded to the sensations warm, just pain, mild pain and moderate pain were determined for each subject. The laser was positioned such that the spot size would be 1 cm in diameter with the subject positioned in the scanner as if to be scanned. The subject then experienced a series of 100-ms laser pulses to the back of the right hand every 2 s which started at a very low level and gradually rose in energy. The subject was asked to rate each stimulus on an arbitrary scale of 1–10 (with 0 corresponding to no sensation; 2 to warm; 4 to just painful; 6 to mild pain; 7 to moderate pain; and 10 to unbearable pain) designed to reflect the different sensory experiences (Chen and Bromm, 1995). This was repeated six times. The energies corresponding to warm, just pain, mild pain and moderate pain were then averaged for the last three trials. The averaged energies were recorded and used for the experimental procedure. These energies were adjusted when necessary during the experiment to maintain a consistent pain intensity rating. Changing the energy was necessary to overcome sensitisation and habituation effects. This solution was adopted because the laser was not able to mechanically change position and volitional movement was considered to introduce an uncertain motor component.
2.4.2. PET procedure
Each subject was positioned in the scanner so that the axis of the scanner was approximately parallel to the glabellar-inion line. A transmission scan was performed using an external ring source of positrons to provide an image of regional tissue density for the correction of emission scans for tissue attenuation.
rCBF in each subject was measured 24 times by recording the distribution of cerebral radioactivity following intravenous bolus infusion of the freely diffusible positron emitting 15O-labelled tracer, H215O. For each measurement, individuals received a 7-mCi bolus of H215O through an automated injector. A 60-s scan commenced 30 s after the start of the bolus. Eight minutes between each injection allowed for the decay of background radiation to less than 10% of the recorded peak. The transmission scan and 24 rCBF measurements were completed in a single session lasting 3.5 h.
Each laser stimulus commenced at the start of the bolus injection, i.e., 30 s prior the start of the scan. This was to account for early sensitisation effects. An intermittent and precisely reproducible laser pulse was applied to the back of the right hand every 2 s for a total of 90 s (45 laser pulses).To avoid any possible order effects the series commenced with each stimulation in a rotated fashion. During the time of stimulation the lights were dimmed and silence maintained. The subjects were instructed to stay motionless and fix their gaze on cross-hairs projected on a computer screen. After each measurement verbal confirmation was obtained that subjects had appropriately experienced the laser pulses as either warm, just painful, mildly painful or moderately painful.
2.4.3. PET data analysis: subtraction
As each scan for each subject results in approximately 1 million radio detections, analysis was completed off line using previously published parametric techniques (Friston et al., 1990; Friston et al., 1991). The object of the analysis of these studies was to compare changes in rCBF between the different stimulation conditions so that the effect of heat intensity without pain could be controlled for in relation to increasingly painful thermal stimulation. To make these comparisons the following procedures were carried out. Correction for head movement between scans was carried out by aligning them all with the first one, using Automated Image Registration (AIR) software (Woods et al., 1992). Each re-aligned set of scans from every subject was reoriented into the standardised anatomical space of the stereotactic atlas of Talairach and Tournoux (1988). To increase the signal to noise ratio and accommodate variability in functional anatomy, each image was smoothed in x, y and z dimensions with a Gaussian filter of 10 mm (FWHM). A correction was made for global changes in blood flow between scans (Friston et al., 1990). These procedures allow flow values for each stimulus condition to be pooled across subjects. Finally a statistical comparison of blood flow distributions between conditions was performed to identify sites of significantly changed regional flow (Friston et al., 1991). Images were coded and ‘binned’ according to the subjective pain report.
2.4.4. PET data analysis: correlation
The verbal rating of stimulus intensity was correlated with adjusted rCBF at each pixel and the profile of correlation coefficients presented as a statistical parametric map (SPM). Brain regions with a high coefficient correspond to structures whose rCBF shares a substantial amount of variance with stimulus intensity. In principle this analysis is more powerful than subtraction as it engages all the variance associated with reported stimulus intensity. The methods of SPM data analysis are described in detail elsewhere (Friston et al., 1995; Friston et al., 1996).
3. Results
3.1. Pain ratings
Table 2 shows the averaged energies used for the experiment and the associated intensity ratings for each stimulus category: warm, just painful, mild pain and moderate pain. The differences between the energies used for each condition and the intensity ratings were statistically significant (_F_=3.18, P<0.05 and _F_=312, P<0.0001, respectively). There were also significant effects of subject (_F_=4.45, P<0.001) and the repeated measure (_F_=7.51, P<0.0001) for the energies used. There was no subject effect or effect of the repeated measure for the intensity ratings (F<1). In other words, there was a significant variation in the energy required to produce the same rating in each subject, which is indicated by the large standard deviations for energy in Table 2, and a significant variation in the energy required during each individual experiment in order to maintain consistent subjective report. This ‘upward drift’ is illustrated graphically in Fig. 1.
Average energies used to obtain the indicated sensory experiences and the average intensity ratings for each of the sensory experiences (_n_=12)
The change in energy required to maintain the same sensory experience during the time course of the experiment (i.e., over the 24 measurements of rCBF), averaged for the 12 subjects. Error bars are omitted for clarity.
3.2. rCBF changes in response to stimulation
The rCBF from the 12 normal controls were compared during warm and three levels of painful experimental heat stimulation of the right hand. Table 3 and 4 show the areas of significant rCBF increase and decrease obtained with the following comparisons: ‘just pain’ versus warm; ‘mild pain’ versus warm and; ‘moderate pain’ versus warm. These changes in rCBF are displayed as SPM(t) in Fig. 2, Fig. 3 and Fig. 4 and as SPM(r) in Fig. 5. The voxels in the standard orthogonal displays are thresholded at _Z_>2.33 (P<0.01). As this threshold does not protect against false positives (Bailey et al., 1991) only previously reported areas, or areas with a _Z_-value in excess of 3.09 (_P_<0.001) are listed in Tables 3 and 4. Selected axial slices derived from each SPM comparison present the increased or decreased rCBF in areas identified according to a priori hypothesis or with _Z_-value>3.09.
rCBF as normalised flow during warm experience, just pain, mild pain and moderate pain in areas with significantly increased blood flow during pain as compared with warm
rCBF as normalised flow during warm experience, just pain, mild pain and moderate pain in areas with significantly decreased blood flow during pain as compared with warm
(A) The comparison pain threshold versus warm. The display format is standard with three orthogonal views of the brain from the top, from the right and from the back (the largest voxel along any line of view is shown). The stereotactic space is that defined by the atlas of Talairach and Tournoux (1988). Voxel are thresholded at P<0.01 (_Z_>2.33). The gray scale is arbitrary with white=max. The right panel shows the design matrix of the general linear model used to partition the data (12 blocks corresponding to the 12 subjects, 24 scan effects and global activity as a covariate of no interest). (B) Selected axial slices derived from the analysis shown in (A). The activations are shown as statistical parametric maps that show the areas of rCBF increase with a _Z_-value coded according to the colour bar shown in the right hand corner. The SPM is superimposed on a standard anatomical reference image derived from a T1-weighted MR image. The numbers above each axial slice refer to the relative distance to the ac-pc line (joining that anterior and posterior commisures), which is situated at 0 mm. The anterior part of the brain corresponds to the top of the image, the posterior parts to the bottom. The left side of each image is the left side of the brain. The red numbers 1–8 indicate regions that are active in this or later comparisons. These regions are numbered in the same way for each of the axial figures. 1 indicates the hippocampus, not active in this comparison; 2 indicates the anterior insula. There is an active region in this area which lies between anterior insula and Brodmann area (BA) 44; 3 indicates the thalami, not active; 4 indicates the prefrontal cortex (BA 10/46) which shows bilateral activation at 8 mm; 5 indicates the perigenual portion of the anterior cingulate cortex, not active; 6 indicates the inferior parietal cortex (BA 39/40) which shows bilateral activation at 32 mm; 7 indicates the midcingulate region, not active; and 8 indicates ipsilateral activation of the premotor region (BA 6) and the prefrontal region (BA 9). (C) The comparison warm versus pain threshold, i.e., the reverse of (A). The display format is as for (A). (D) Selected axial slices derived from the analysis shown in (C). The activations are shown as before for (B). The number 1 indicates the amygdala and hippocampal region not showing any rCBF decrease for this comparison. Temporal decrease in the region of the PHG (BA 37) can be seen as well as bilateral occipital cortex and cerebellum.
(A) The comparison mild pain versus warm. Detail is as for Fig. 2A. (B) Selected axial slices derived from the analysis shown in (A). The activations are shown as before. All indicated regions are now active except for the cingulate regions (BA 24). Additional to the indicated regions there is significant rCBF increase in the contralateral lentiform nucleus and caudate at 4–8 mm and in the medial frontal region (BA 32) anterior to the midcingulate region at 28–40 mm. (C) The comparison warm versus mild pain. Detail is as before. (D) Selected axial slices derived from the analysis shown in (C). The activations are shown as before. The amygdala and hippocampal region also show no rCBF decrease for this comparison. Temporal decrease in the region of the PHG (BA 37) and bilateral occipital cortex and cerebellum are observed as before. In addition, decreased rCBF can be observed in the medial frontal cortex (BA 32).
(A) The comparison moderate pain versus warm. Detail as before. (B) Selected axial slices derived from the analysis shown in (A). The activations are shown as before. All indicated regions are now active except for the hippocampal and midcingulate region (BA 24). Additional to the indicated regions there is significant bilateral rCBF increase in the lentiform nucleus (4–8 mm) and medial frontal region (BA 32) anterior to the midcingulate region. The contralateral somatosensory region is indicated with the inferior parietal cortex (region 6). (C) The comparison warm versus moderate pain. Detail is as before. (D) Selected axial slices derived from the analysis shown in (C). The activations are shown as before. The amygdala region has decreased rCBF for this comparison. Decreases in occipital cortex and cerebellum are observed as less bilateral than before. The decreased rCBF in the temporal region of the PHG (BA 37) and in the medial frontal cortex (BA 32) is still visible but is less significant than previously. In addition to previous comparison, there is a bilateral decrease in the SMA (BA 4) and S1 region.
(A) The pixels showing a positive correlation with subjective pain rating. (B) Selected axial slices derived from the analysis shown in (A). The positive correlations are shown as statistical parametric maps that show the areas of rCBF increase with a _Z_-value coded according to the colour bar shown in the top right hand corner. Detail is as before. All indicated regions are active almost exactly as for Fig. 4A except for the regions being generally larger with slightly improved significance. (C) The pixels showing a negative correlation with subjective pain rating. (D) The selected axial slices derived from the analysis shown in (C). The correlations are shown as before. As for the subtraction displayed in Fig. 4D, the amygdala indicates decreased rCBF correlation (deactivation) with pain rating. Decreases in occipital cortex are observed as for Fig. 4D. Deactivation in the temporal region of the PHG (BA 37), however, is not detectable. The medial frontal cortex (BA 32) indicates a large deactivation. At 12 mm there is a deactivation at the juncture of orbitofrontal (BA 47) and temporal cortex (BA 38) in the left hemisphere which was not detected using the subtraction displayed in Fig. 4D. Deactivation in the bilateral SMA and S1 region is less pronounced on the contralateral side than that seen in Fig. 4D. At 48 mm there is a deactivation in the frontal eye fields (BA 8) not seen previously.
3.3. Cognitive subtractions: rCBF increases
Fig. 2A,B show rCBF increases during the noxious stimulation described as ‘just painful’ compared with the warm stimulation. This comparison revealed significant rCBF increases in the contralateral prefrontal cortex (BA 10/46/44, bordering with anterior insula), bilateral inferior parietal cortex (BA 39/40) and ipsilateral premotor cortex (BA 6/9).
Fig. 3A,B show the comparison of ‘mild pain’ with warm. In addition to the activations revealed in the previous comparison (which are increased in extent and significance) there is additional rCBF increase in the contralateral hippocampus, lentiform nucleus, insula cortex and premotor cortex (BA 6), bilateral superior medial frontal cortex (BA 32) and ipsilateral thalamus and prefrontal cortex (BA 10/46).
Fig. 4A,B show ‘moderate pain’ compared with warm. All previously observed regions are active with increased extent and significance except for the hippocampal region which is no longer active. In addition there is increased rCBF in the contralateral thalamus and primary somatosensory cortex (S1) and in the ipsilateral prefrontal cortex (BA 44/45) and perigenual cingulate cortex (BA 24).
3.4. Cognitive subtractions: rCBF decreases
Fig. 2C,D show rCBF decreases during the noxious stimulation described as ‘just painful’ compared with the warm stimulation (i.e., the opposite of Fig. 2A,B). This comparison revealed significant rCBF decreases in the region of the contralateral para-hippocampal-gyrus (PHG) (BA 37) and bilateral occipital cortex (BA 18).
Fig. 3C,D show the comparison of warm with ‘mild pain’. This comparison also revealed significant rCBF decreases in the region of the contralateral PHG (BA 37) and bilateral occipital cortex (BA 18). An additional decrease was observed in the contralateral inferior medial frontal cortex (BA 32/10).
Fig. 4C,D show the comparison of warm with ‘moderate pain’. This comparison revealed a contralateral rCBF decrease in the region of the amygdala, a bilateral decrease in the region of the SMA (BA 4) and S1 cortex and an ipsilateral decrease in the occipital cortex (BA 18). A decrease in the PHG region and in the medial frontal region can be observed, but these are now below the level of significance (Z<3.09).
3.5. Positive correlation
Fig. 5A,B indicate the positive rCBF correlation with increasing stimulus intensity (activation). There is activation in all regions described for Fig. 4A,B, although there is some attenuation of the ipsilateral lentiform response.
3.6. Negative correlation
Fig. 5C,D indicate the negative rCBF correlation with increasing stimulus intensity (deactivation). There are two regions of deactivation not observed in Fig. 4C,D, one region at the juncture of orbitofrontal (BA 47) and temporal cortex (BA 38) and a second region in the frontal cortex (BA 8).
4. Discussion
The hypothesis under investigation was that different regions of brain may process various aspects of pain and that these different regions may activate and deactivate according to stimulus intensity. Our data indicate that as the stimulus rating increases in intensity an increasing number of brain regions become active. Except for the hippocampus and the ipsilateral lentiform nucleus, central responses to noxious stimulation appear linearly related to subjective intensity. In contrast, decreases in rCBF were more variable and less predictable based on intensity. This result raises the possibility that the inconsistent results observed across imaging studies may be in part a consequence of using variable stimulus intensities. It should be noted, however, that there exists considerable variability in PET results both between groups (as illustrated in Table 1) and subjects (Vogt et al., 1996). As indicated in Section 1, some of this variability will be due to differences in the scanning technology and analysis software rather than to physiological factors. In addition, the techniques used for subjective pain measurement vary across studies and are liable to reflect different aspects of the multi-dimensional pain experience. The differences observed between stimulation and studies will, in part, be a product of these technical factors and should be viewed with due caution.
In this study, a CO2 laser delivered thermal sensation to the back of the right hand. The laser energy was fixed according to each subject's report of warm, pain threshold, mild pain and moderate pain. The energy required to produce these sensations varied considerably both across individuals and scans. There was pronounced habituation to the stimulus indicated by the upward slope in energy needed for each successive scan (Fig. 1). However, intensity rating of the pain remained fairly consistent across scans and individuals, indicating that the energy adjustments made during the procedure were successful in maintaining the desired sensations. Each intensity rating produced rCBF changes in a number of cortical and subcortical areas which will be considered for discussion.
5. Hippocampus and amygdala
This study is the first to identify rCBF changes in the hippocampus and amygdala during the experience of pain, although Hsieh et al. (1995a) previously reported subsignificant increases in the amygdala. Based on the function and connections of the amygdala, its involvement in pain is not surprising. The amygdala has an intimate involvement in the formation of ‘fear avoidance’. For example, rats with amygdala lesions do not respond to contextual or explicit cues indicating a danger such as the imminent arrival of a painful shock (Selden et al., 1991; Kim and Fanselow, 1992). Removal of the amygdala along with the temporal lobes in primates results in their exhibiting behaviour that includes the loss of fear inhibition (Kluver and Bucy, 1937). Lesioned monkeys will approach humans, animals and inanimate objects without hesitation. Bernard et al. (1990) have described a complicated nociceptive specific response in the central amygdala of anaesthetised rats, involving inhibitory and excitatory responses, which does not show somatotopic organisation. They interpret this response as indicating possible emotional, autonomic or behavioural reactions to noxious events. LeDoux has published extensively on the involvement of amygdala in emotional responses using mostly animal models (see LeDoux (1995)and LeDoux (1996) for reviews). Recent support for amygdala involvement in processes associated with human fear responses comes from PET (Morris et al., 1996). Amygdala rCBF correlates positively with facial images depicting an increasing expression of fear.
In contrast, the hippocampus does not appear to have a direct role in the formation of pain or fear avoidance but does have an important role in the consolidation of long-term memories (Zola-Morgan et al., 1989). In particular, hippocampal damage inhibits the assimilation of contextual information preventing animals solving tasks that require the use of mapping strategies or from responding to contextual cues indicating danger (Morris et al., 1982).
Gabriel et al. (1980)and Gabriel (1993) have suggested that the memory function of the temporal hippocampal system emerges from interactions between the medial temporal lobe, cingulate cortical and diencephalic limbic circuits. Several neuroanatomical studies have shown that ACC receives substantial input from the amygdala (Amaral and Price, 1984; Vogt et al., 1987). In addition, widespread hippocampocortical connections have been described including connections with the orbitofrontal cortex, ACC and the PHG (Van Hoesen et al., 1993). The connections between ACC, insula and amygdala-hippocampal-prefrontal circuits constitute a network within which fear and contextual information relevant to pain can be integrated.
It is unclear why the hippocampal response we observed increases only during mild pain and not during moderate pain and further why this increase occurs with PHG and amygdala decreases. Currently there is too little information to warrant speculation. Both the hippocampus and amygdala are small structures with many complicated clusters of nuclei that may produce variable patterns of rCBF increase and decrease, the basis of which is yet to be understood.
6. Anterior cingulate cortex (ACC)
Possibly the most surprising result of this study is the lack of a midcingulate response to pain. Vogt (1993) defines the midcingulate region as the ventral division of the ACC, denoted area 24′. The division of the ACC into midcingulate and rostral cingulate (the more anterior portion of ACC; also called perigenual cingulate) is motivated by the different cytoarchitecture and connections of these regions. These differences are described in detail elsewhere (Vogt et al., 1992; Vogt, 1993; Devinsky et al., 1995). In summary, there is a transition region at the border between the agranular rostral cingulate and the granular midcingulate. From a connection point of view, rostral area 24′ in the rat receives most mediodorsal thalamic (Krettek and Price, 1977) and amygdala (Sripanidkulchai et al., 1984) afferents. Area 24′ has major and reciprocal connections with cingulate premotor areas, the pontine nuclei and area 46 of the prefrontal cortex (Vogt, 1993). Thus the midcingulate region is liable to be involved in functions relating to motor response and has been implicated in response selection during divided attention tasks (Pardo et al., 1990; Corbetta et al., 1991; Bench et al., 1992). Processes associated with affect, by contrast, are liable to involve the perigenual cingulate region. Electrical stimulation in the perigenual region can evoke fear and other emotional responses (Bancaud and Talairach, 1992) and perigenual cingulate activation has been reported with PET when subjects recalled sad events during the scan (George et al., 1995). Area 32 of the frontal cortex lies immediately adjacent to the anterior portion of the ACC (area 24) and its function is often elided with the functions of perigenual ACC (e.g., Vogt et al., 1996). The suggestion that medial frontal area 32 is involved in emotional functions receives empirical support (Frystak and Neafsy, 1991) however it is also involved in spatial regulation (Poucet and Herrmann, 1990).
The lack of midcingulate response is discordant with our previous reports (Jones et al., 1991; Derbyshire et al., 1994; Vogt et al., 1996) and with most other PET studies with pain where the position of the cingulate response can be assessed (Casey et al., 1994; Coghill et al., 1994; Craig et al., 1996). Two studies have reported activity in perigenual cingulate without midcingulate increases (Rosen et al., 1994; Weiller et al., 1995). Both involved clinical pain that could not be avoided by a motor response. Rosen et al. (1994) reported decreased rCBF in the midcingulate during their study of angina. This was accounted for by patients learning to actively suppress movement during an angina attack to prevent increased pain. The inescapable, frightening or worrying aspect of angina pain was considered to explain the activation of perigenual cingulate.
An important behavioural aspect of the present PET paradigm is the need to restrain movement in spite of a noxious but bearable somatic stimulus. This suppression of movement is liable to result in activation and deactivation of motor regions including midcingulate, lentiform nucleus, premotor areas and somatomotor (SMA) regions. An important difference between laser stimulation and stimulation with mechanical heat probe is the lack of physical contact with the subject. During laser stimulation there is no physical prevention of movement meaning the subjects have to actively suppress movement as for the angina patients. It is possible that the deliberate self infliction of pain, with a surrender of control, is a more emotive experience than when pain is perceived to be inflicted by others and may account for the differential activation of rostral and midcingulate cortex. In addition, the laser stimulation resulted in greater subject concern regarding skin damage than previously encountered with a heat probe (e.g., Derbyshire et al., 1994; Vogt et al., 1996). Certainly the present paradigm produced extensive rostral cingulate activation that was evident during mild and moderate pain experience. The activation was also restricted to the contralateral (right) side which Hsieh et al. (1995b) suggested to be exclusively associated with affect.
This interpretation, however, based on speculation and anecdote, should be viewed cautiously. The ACC response shown anterior to the midcingulate area is contiguous with area 24′ and may involve this region, especially during the higher level of pain (Fig. 4B). These relatively close-packed regions may not be resolvable using PET, although the difference in comparison to our previous work and other studies is striking. Further studies, using more specific activation paradigms, are necessary to address the precise roles of particular regions of the ACC in the central processing of pain.
7. Thalamus
In this study thalamic activation was extensive, highly significant and bilateral. Although the difference is small, the ipsilateral thalamus indicated a more significant response and incorporated a greater number of thalamic regions particularly during mild pain. Although the spatial resolution of PET is insufficient to discriminate the separate neuronal regions within the thalamus, it is clear that the ventral posterior (VPL) and medial (dm) thalamic regions showed bilateral positive correlation with pain intensity (Fig. 5B). The subtraction of warm from moderate pain, however, suggested a more specific response to pain in the contralateral dm complex and anterior thalamic region. Ipsilateral thalamic responses were certainly more posterior suggesting involvement of the pulvinar on the right side.
Casey et al. (1994) also reported ipsilateral thalamic responses to pain and suggested the responses may be related to nociceptive inputs from ipsilateral projections of the spinothalamic tract. Their study emphasised the role of the VPL complex and cited evidence of differential or exclusive VPL responses to noxious stimuli recorded contralaterally in anaesthetised cats and awake monkeys and humans. Evidence that microstimulation within the VPL can produce painful sensations and that lesions can produce a marked contralateral loss of somatic sensations, including pain, was cited to suggest the excitation of neurones within the contralateral VPL thalamus as ‘necessary for the sensation of pain’. Although the VPL region and its subsequent projection to S1 cortex have an undoubted involvement in processes associated with pain, such as sensory-discrimination, they have no role in affect and learning processes that mediate long-term avoidance of noxious stimuli. It is possible that VPL activity is necessary for pain, but it seems unlikely that VPL activity is sufficient for pain, at least under normal circumstances (Jones and Derbyshire, 1996).
The activation of structures such as the ACC, insula and prefrontal cortex cannot be explained from VPL activation (Vogt et al., 1987). Projections to these regions are from the medial thalamic nuclei and form the basis of the ‘medial pain system’ as a theoretical construct (Vogt and Pandya, 1987). The medial pain system incorporates the limbic thalamic and telencephalic structures involved in processing the motivational-affective features of noxious stimuli as well as the motor system interactions needed for generating relevant behaviours.
The involvement of bilateral thalamus with differential anterior/medial activation on the left and posterior (pulvinar)/VPL activation on the right suggests an intriguing lateralised parallel processing associated with sensory integration via the right posterior thalamus and emotional integration via the left anterior thalamus. This speculation would be more compelling, however, if the VPL activation was contralateral. As for divisions within the ACC, further studies are imperative to address the precise roles of different thalamic structures with the additional requirement of more sensitive imaging techniques.
8. Insula
The current paradigm indicated contralateral mid-insula activity in response to mild pain that spread in an anterior direction during moderate pain. rCBF in a large portion of the anterior and mid-insula correlated positively with stimulus intensity. The insula connects reciprocally with S2, receives input from spinothalamically activated posterior thalamic nuclei and projects to the amygdala and perirhinal cortex (Burton and Jones, 1976; Friedman and Murray, 1986). Thus the insula is well situated for the integration of information relating to the affective and reactive components of pain (Casey et al., 1994; Coghill et al., 1994; Hsieh et al., 1995a) and is included by Gabriel (1993) as part of the circuitry related to fear avoidance.
As for the ACC, however, it is wrong to consider the insula as a homogenous area. Perigenual cingulate has stronger connections with the more posterior portion of insula than the anterior region (Vogt et al., 1996) and there is some tentative evidence that the anterior insula is involved in functions relating to sadness (Drevets et al., 1995b) while the mid-insula is involved in autonomic regulation (Lane et al., 1995).
9. Lentiform nucleus
The contribution of lentiform nucleus to pain processes remains uncertain. Outside of our own group, activation of the lentiform region is rare (Coghill et al. (1994), Hsieh et al. (1995a)and Hsieh et al. (1996) are exceptions) and even more rarely commented on. It is possible that lentiform activation is being overlooked. The study of Craig et al. (1996), for example, has an activation of lentiform nucleus that is clearly visible in their Fig. 1 but is not listed in their results nor commented upon.
The lentiform nucleus is most often associated with planned action (Brooks et al., 1993) and movement (Golebatch et al., 1991). Thus the increased lentiform rCBF observed here may relate to an alerting or priming mechanism as has been observed in primates (Alexander, 1987). Such a mechanism may be expected to operate bilaterally as was the case in response to the more intense, moderate pain. It is possible that the contralateral response to the mild pain indicated a preparation to move the affected hand, whereas with increasing intensity preparation to use the other arm or to flee resulted in bilateral activation. This scenario suggests lentiform activation as part of an ancillary pain mechanism related to escape behaviour.
The major output pathways of the lentiform nucleus project to the ventral lateral, ventral anterior, centromedian and mediodorsal nuclei of the thalamus which in turn project to the prefrontal cortex, the premotor cortex, the supplementary motor area (SMA), the motor cortex and the ACC (Cote and Crutcher, 1991; Neafsey et al., 1993). This provides a potential basis for a circuit associated with motor preparation or response selection.
10. Prefrontal cortex and inferior parietal cortex
Two foci of activation in the dorsolateral prefrontal (DLPF) region have been identified bilaterally. The first region broadly corresponds to area 10 and area 46. This region has previously been reported by this group on the contralateral (Vogt et al., 1996) and ipsilateral side (Derbyshire et al., 1994). Two reports from the Karolinska group have shown bilateral responses (Hsieh et al., 1995a; Hsieh et al., 1995b). The second region of prefrontal activation broadly corresponds to area 44 and area 45 which has not been previously reported by our group but has been reported elsewhere (Rosen et al., 1994; Hsieh et al., 1995a; Hsieh et al., 1996). The association functions of the various regions of the frontal cortex are too diverse to be easily summarised. In general the prefrontal responses are interpreted as a consequence of ‘planning’, involving behavioural and attentional organisation during pain (Derbyshire et al., 1994; Hsieh et al., 1995a; Hsieh et al., 1995b; Vogt et al., 1996). This will include plans to move and to speak which are probably represented in different regions of the DLPF cortex.
On the contralateral side prefrontal activation in all active regions (10/46/44/45) was evident in the presence of the lowest pain stimulation (‘just pain’). The response became bilateral in area 10/46 in response to mild pain and was bilateral in all areas in response to moderate pain. Broadly speaking, this response supports the role of the prefrontal cortex as being associated with ‘vigilance’. Anatomical studies suggest that the DLPF cortex works closely with the posterior parietal cortex; the two regions are the most densely interconnected areas of association cortex, they both project to numerous common cortical and subcortical regions, and prefrontal and parietal responses to target stimuli often operate in concert (Kupfermann, 1991). Posner and Rothbart (1991) have divided the brain into two broad attentional networks, the posterior attentional system responsible for preconscious orientation and the anterior attentional system responsible for focal attention. The prefrontal supervisory system is seen as regulating information between these two attentional networks. The activation of inferior parietal cortex (area 40) associated with the posterior attention system and the early prefrontal response suggests orientation of attention towards incoming sensory stimulation from the hand, with inhibition of motor and verbal responses.
11. Lateral premotor cortex
Previous studies have reported primary motor (MI), SMA or premotor activation (Coghill et al., 1994; Derbyshire et al., 1994; Hsieh et al., 1995a). Lesions of the premotor cortex and SMA cause more complex movement disorders than do lesions of the primary motor cortex. Loss of the premotor cortex impairs the ability to develop an appropriate strategy for movement. The premotor cortex receives its principal input from the posterior parietal cortex and from thalamic neurones relaying information from the globus pallidus. Weinrich and Wise (1982) demonstrated neuronal firing in the premotor cortex when the animal receives an instruction telling it to move. Thus the inferior parietal, lateral premotor, prefrontal and lentiform regions constitute a network within which orientation and preparedness for action can take place without any actual movement occurring. It is possible that projections from DLPF (areas 10/46) and from premotor cortex to ACC (area 24′) are inhibiting the motor responses in area 24 resulting in a lack of activity from the midcingulate region (Van Hoesen et al., 1993).
12. Somatosensory cortex
The changes in rCBF in the S1 region are confusing. Contralateral increased rCBF occurs in response to the moderate pain at 28–32 mm from the ac-pc line. Decreased rCBF occurs on the ipsilateral side at 48–52 mm and on the contralateral side at 56 mm (though these changes border with SMA cortex). Examination of the sensory homunculus (Martin and Jessell, 1991) suggests the increase is possibly a little below the hand area and towards the face region of S1 cortex whereas the decreases are just above the hand region in the leg and trunk area of S1. The decreases can be interpreted with reference to Apkarian et al. (1992)and Drevets et al. (1995a) as indicating the redundancy of the leg and trunk region, especially on the ipsilateral side. Changes in blood flow in those regions reflects the organisation of the CNS towards the stimulated region.
The Coghill study, the Casey study and the Talbot study all suggest an S1 hand/forearm region increase at 50–60 mm above the ac-pc line. These results, therefore, place our S1 increase in the face area and our S1 decrease in the hand area or, alternatively, place their S1 increase in the leg and trunk region. Hsieh et al. (1995a) reported bilateral activation of the S1 face region (at 20–30 mm above the ac-pc line) during an acute injection of ethanol to the upper arm. It is possible that the main determinants of direction of rCBF responses in S1 are not predominantly determined by the presence or absence of pain but rather the sensory-discriminant components of the stimulus, such as edge detection and pressure detection, that are difficult to control for in studies using a physical probe. However, the cortical forearm representation may be rather high up on the convexity, close to the trunk representation, and PET may not be able to resolve these relatively closely aligned somatotopic representations.
13. Other comments and conclusions
This study has identified a large number of areas involved in the central processing of acute pain resulting from stimulation with a CO2 laser. Two of the major problems facing PET researchers are the analysis and display of the data. Early studies used very strict criteria for the thresholding of PET data (P<0.05 with a correction for multiple comparisons) meaning that positive results could be interpreted with confidence while negative results had to be interpreted cautiously (e.g., Jones et al., 1991; Talbot et al., 1991). More recently these thresholds have tended to drop; it is now usual to report PET data without correcting for multiple comparisons at P<0.001 (e.g., Casey et al., 1994; Coghill et al., 1994; Derbyshire et al., 1994) and becoming increasingly usual at P<0.01 (e.g., Hsieh et al., 1995a; Hsieh et al., 1995b; Hsieh et al., 1996); the default P<0.01 in the latest SPM packages encourages this trend). Once this effect is combined with the increased sensitivity of a new generation of PET cameras it becomes clear that attempting to compare new studies with old is like trying to compare apples with oranges.
In addition to this is the inevitable increase in the number of regions activated and the possibility of reporting single subject analyses (e.g., Vogt et al., 1996). Reporting all available data in 12 subjects, recorded from a modern camera, at a threshold of P<0.01 would likely fill a journal of this size. Clearly, the report of PET data requires some form of rationalisation. The results from this study suggest that the correlation analysis may be a reasonable guide to the overall results. It is striking that rCBF correlates significantly with perceived stimulus intensity in all the regions previously demonstrated using subtraction analysis: prefrontal, insula, inferior parietal, perigenual cingulate and lateral premotor cortices and thalamus and lentiform nucleus. Inclusion of only the correlation analysis, which is a more powerful technique than subtraction, would have shown the major active regions. However, this would have been at the loss of the hippocampal and amygdala result and much of the detail regarding changing activity in the cortical and subcortical regions identified.
Another alternative is to focus studies from the beginning. It is now clear that pain activation does much more than activate a ‘pain centre’; activation during pain processing incorporates motor function, attentional allocation, emotional processes and sensory localisation. It should be of little surprise that this diversity, which is the basis of the multidimensional understanding of pain, has been demonstrated to produce activation in a large number of neuronal areas which vary with stimulus intensity. To further understand the detail of this ‘neuromatrix’ for pain it is imperative to carry out studies which examine the precise roles of each particular region during the central processing of pain.
Acknowledgements
The authors are grateful to Tom Nichols, Carnegie Mellon University, Pittsburgh, for his guidance with the statistics; Professors Terry King and Mark Dickenson, Laser Photonics and Physics Department, Manchester University, for their expert advice and guidance in the construction of the laser system; and Professor Ken Casey, Chief Neurology Service, V.A. Medical Center, Michigan, for his useful comments and constructive criticisms of this manuscript.
References
Alexander GE. Selective neuronal discharge in monkey putamen reflects intended direction of planned movements. Exp. Brain Res. 1987;67:623-634.
Amaral DG, Price JL. Amygdalo-cortical projections in the monkey (Macaca fascicularis). J. Comp. Neurol. 1984;230:465-496.
Apkarian VA, Stea RA, Manglos SH, Szevernyi NM, King RB, Thomas FD. Persistent pain inhibits contralateral somatosensory cortical activity in humans. Neurosci. Lett. 1992;140:141-147.
Bailey DL, Jones T, Spinks TJ, Gilardi M-C, Townsend DW. Noise equivalent count measurements in a neuro-PET scanner with retractable septa. IEEE Trans. Med. Imaging. 1991;10:256-260.
Bancaud, J. and Talairach, J., Clinical semiology of frontal lobe seizures. In: P. Chauvel, A.V. Delgado-Escueta, E. Halgren and J. Bancaud (Eds.), Frontal Lobe Seizures and Epilepsies, Raven Press, New York, 1992, pp. 3–58.
Bench CJ, Frith CD, Grasby PM, Friston KJ, Paulesu E, Frackowiak RSJ, Dolan RJ. Patterns of cerebral activation during the Stroop colour word interference task: a positron emission tomography study. Neuropsychologica. 1992;31:907-922.
Bernard JF, Huang GF, Besson JM. Effect of noxious somesthetic stimulation on the activity of neurons of the nucleus centralis of the amygdala. Brain Res. 1990;523:347-350.
Brooks, D.J., Jenkins, I.H., Playford, E.D. and Passingham, R.E., Positron emission tomography studies on regional cerebral control of voluntary movement. In: N. Mano, I. Hamanda and M.R. DeLong (Eds.), Role of the Cerebellum in Voluntary Movement, Elsevier, Amsterdam, 1993, pp. 267–274.
Burton, H. and Jones, E.G., The posterior thalamic region and its cortical projection in New World and Old World monkeys, J. Comp. Neurol., 168 (1976) 301–349.
Casey KL, Minoshima S, Berger KL, Koeppe RA, Morrow TJ, Frey KA. Positron emission tomographic analysis of cerebral structures activated specifically by repetitive noxious heat stimuli. J. Neurophysiol. 1994;71:802-807.
Chen, A.C.N. and Bromm, B., Pain related generators of laser-evoked potentials: brain mapping and dipole modelling. In: B. Bromm and J.E. Desmedt (Eds.), Pain and the Brain: From Nociception to Cognition, Advances in Pain Research and Therapy, Vol. 22, Raven Press, New York, 1995, pp. 245–266.
Coghill RC, Talbot JD, Evans AC, Meyer E, Gjedde A, Bushnell MC, Duncan GH. Distributed processing of pain and vibration by the human brain. J. Neurosci. 1994;14:4095-4108.
Corbetta M, Meizin FM, Dobmeyer S, Shulman GL, Petersen SE. Selective and divided attention during visual discriminations of shape, color and speed: functional anatomy by positron emission tomography. J. Neurosci. 1991;11:2383-2402.
Cote, L. and Crutcher, M.D., The basal ganglia. In: E.R. Kandel, J.H. Schwartz and T.M. Jessell (Eds.), Principles of Neural Science, 3rd edn., Elsevier, Amsterdam, 1991, pp. 647–659.
Craig AD, Reiman EM, Evans A, Bushnell MC. Functional imaging of an illusion of pain. Nature. 1996;384:258-260.
Derbyshire SWG, Jones AKP, Devani P, Friston KJ, Feinmann C, Harris M, Pearce S, Watson JDG, Frackowiak RSJ. Cerebral responses to pain in patients with atypical facial pain measured by positron emission tomography. J. Neurol. Neurosurg. Psychiatry. 1994;57:1166-1173.
Devinsky O, Morrel MJ, Vogt BA. Contributions of anterior cingulate cortex to behaviour. Brain. 1995;118:279-306.
Drevets, W.C., Burton, H., Videen, T.O., Snyder, A.Z., Simpson, J.R. and Raichle, M.E., Blood flow changes in human somatosensory cortex during anticipated stimulation, Nature, 373 (1995a) 249–252.
Drevets, W.C., Simpson, J.R. and Raichle, M.E., Regional blood flow changes in response to phobic anxiety and habituation, J. Cerebral Blood Flow Metab., 15 (S1) (1995b) S856.
Duncan GH, Bushnell MC, Talbot JD, Evans AC, Meyer E, Marrett S. Pain and activation in the thalamus (letter; comment). Trends Neurosci. 1992;15:252-253.
Di Piero V, Jones AKP, Iannotti F, Powell M, Perani D, Lenzi GL, Frackowiak RSJ. Chronic pain: a PET study of the central effects of percutaneous high cervical cordotomy. Pain. 1991;46:9-12.
Fox PT, Mintun MA. Non-invasive functional brain mapping by change-distribution analysis of averaged PET images of H215O tissue activity. J. Nucl. Med. 1989;30:141-149.
Friedman, D.P. and Murray, E.A., Thalamic connectivity of the second somatosensory area and neighboring somatosensory fields of the lateral sulcus of the macaque, J. Comp. Neurol., 252 (1986) 348–373.
Friston, K.J. and Frackowiak, R.S.J., Imaging functional anatomy, Alfred Benzon Symposium 31, Munksgaard, Copenhagen, 1991.
Friston KJ, Frith CD, Liddle PF, Dolan RJ, Lammertsma AA, Frackowiak RSJ. The relationship between local and global changes in PET scans. J. Cerebral Blood Flow Metab. 1990;10:458-466.
Friston KJ, Frith CD, Liddle PF, Frackowiak RSJ. Comparing functional (PET) images: the assessment of significant change. J. Cerebral Blood Flow Metab. 1991;11:690-699.
Friston KJ, Holmes AP, Worsley KJ, Poline JP, Frith CD, Frackowiak RSJ. Statistical parametric maps in functional imaging: a general approach. Hum. Brain Mapp. 1995;2:189-210.
Friston KJ, Price CJ, Fletcher P, Moore C, Frackowiak RSJ, Dolan RJ. The trouble with cognitive subtraction. Neuroimage. 1996;4:97-104.
Frystak RJ, Neafsy EJ. The effect of medial frontal cortex lesions on respiration, ‘freezing’, and ultrasonic vocalizations during conditioned emotional responses in rats. Cerebral Cortex. 1991;1:418-425.
Gabriel, M., Discriminative avoidance learning: a model system. In: B.A. Vogt and M. Gabriel (Eds.), Neurobiology of Cingulate Cortex and Limbic Thalamus: A Comprehensive Treatise, Birkhauser, Boston, MA, 1993. pp. 478–523.
Gabriel M, Foster K, Orona E, Saltwick SE, Stanton M. Neuronal activity of cingulate cortex, anteroventral thalamus, and hippopcampal formation in discriminative conditioning. Progr. Psychobiol. Physiol. Psychol. 1980;9:125-231.
George MS, Ketter TA, Parekh PI, Horowitz B, Herscovitch P, Post RM. Brain activity during transient sadness and happiness in healthy women. Am. J. Psychiatr. 1995;152:341-351.
Golebatch JG, Deiber MP, Passingham RE, Friston KJ, Frackowiak RSJ. Regional cerebral blood flow during voluntary arm and hand movements in human subjects. J. Neurophysiol. 1991;65:1392-1401.
Hsieh, J.C., Stahle-Backdahl, M., Hagermark, O., Stone-Elander, S., Rosenquist, G. and Ingvar, M., Traumatic nociceptive pain activates the hypothalamus and the periaqueductal gray: a positron emission tomography study, Pain, 64 (1995a) 303–314.
Hsieh, J.C., Belfrage, M., Stone-Elander, S., Hansson, P. and Ingvar, M., Central representation of chronic ongoing neuropathic pain studied by positron emission tomography, Pain, 63 (1995b) 225–236.
Hsieh JC, Hannerz J, Ingvar M. Right-lateralised central processing for pain of nitro-glycerine-induced cluster headache. Pain. 1996;67:59-68.
Iadorola MJ, Max MB, Berman KF, Byas-Smith MG, Coghill RC, Gracely RH, Bennett GJ. Unilateral decreases in thalamic activity observed with positron emission tomography in patients with chronic neuropathic pain. Pain. 1995;63:55-64.
Jones AKP, Derbyshire SWG. Cerebral mechanisms operating in the presence and absence of inflammatory pain. Ann. Rheum. Dis. 1996;55:411-420.
Jones APK, Brown WD, Friston KJ, Qi LY, Frackowiak RSJ. Cortical and subcortical localization of response to pain in man using positron emission tomography. Proc. R. Soc. Lond. 1991;244:39-44.
Kim JJ, Fanselow MS. Modality-specific retrograde amnesia of fear. Science. 1992;256:676-677.
Kluver H, Bucy PC. ‘Psychic blindness’ and other symptoms following bilateral temporal lobectomy in rhesus monkeys. Am. J. Physiol. 1937;119:352-353.
Krettek JE, Price JL. The cortical projections of the mediodorsal nucleus and adjacent thalamic nuclei in the rat. J. Comp. Neurol. 1977;84:157-192.
Kupfermann, I., Localization of higher cognitive and affective functions: the association cortices. In: E.R. Kandel, J.H. Schwartz and T.M. Jessell (Eds.), Principles of Neural Science, 3rd edn., Elsevier, Amsterdam, 1991, pp. 823–838.
Lane RD, Reiman EM, Ahern GL, Schwartz GE, Davidson RJ, Axelrod B, Yun LS. Neuroanatomical correlates of happiness, sadness and disgust. Hum. Brain Mapp. 1995;S1:212.
LeDoux JE. Emotion: clues from the brain. Ann. Rev. Psychol. 1995;46:209-235.
LeDoux, J.E., The Emotional Brain: The Mysterious Underpinnings of Emotional Life, Simon and Schuster, London, 1996.
Martin, J.H. and Jessell, T.M., Anatomy of the somatic sensory system. In: E.R. Kandel, J.H. Schwartz and T.M. Jessell (Eds.), Principles of Neural Science, 3rd edn., Elsevier, Amsterdam, 1991, pp. 353–366.
Melzack R. Phantom limbs, the self and the brain: the D.O. Hebb memorial lecture. Can. Psychol. 1989;30:1-16.
Melzack, R. and Casey, K.L., Sensory, motivational and central control determinants of pain. In: D. Kenshalo (Ed.), The Skin Senses, Thomas, Springfield, IL, 1968.
Merskey H. The definition of pain. Eur. J. Psychiatry. 1991;6:153-159.
Morris RGM, Garrud P, Rawlins JNP, O'Keefe J. Place navigation impaired in rats with hippocampal lesions. Nature. 1982;297:681-683.
Morris JS, Frith CD, Perrett DI, Rowland D, Young AW, Calder AJ, Dolan RJ. A differential neural response in the human amygdala to fearful and happy facial expressions. Nature. 1996;383:812-815.
Neafsey, E.J., Terreberry, R.R., Hurley, K.M., Ruit, K.G. and Frysztak, R.J., Anterior cingulate cortex in rodents: connections, visceral control functions, and implications for emotion. In: B.A. Vogt and M. Gabriel (Eds.), Neurobiology of Cingulate Cortex and Limbic Thalamus: A Comprehensive Treatise, Birkhauser, Boston, MA, 1993, pp. 206–223.
Pardo JV, Pardo PJ, Janer KW, Raichle ME. The anterior cingulate cortex mediates processing selection in the Stroop attentional conflict paradigm. Proc. Natl. Acad. Sci. USA. 1990;87:256-259.
Posner, M.I. and Rothbart, M.K., Attentional mechanisms and conscious experience. In: A.D. Milner and M.D. Rugg (Eds.), The Neuropsychology of Consciousness, Academic Press, London, 1991, pp. 91–111.
Poucet B, Herrmann T. Septum and medial frontal cortex contribution to spatial problem solving. Behav. Brain Res. 1990;37:269-280.
Rosen SD, Paulesu E, Frith CD, Frackowiak RSJ, Davies GJ, Jones T, Camici PG. Central nervous pathways mediating angina pectoris. Lancet. 1994;344:147-150.
Selden NRW, Everitt BJ, Jarrard LE, Robbins TW. Complementary roles for the amygdala and hippocampus in aversive conditioning to explicit and contextual cues. Neuroscience. 1991;42:335-350.
Sripanidkulchai K, Sripanidkulchai B, Wyass JM. The cortical projections of the basolateral amygdaloid nucleus in the rat: a retrograde fluorescent dye study. J. Comp. Neurol. 1984;229:419-431.
Talairach, J. and Tournoux, P., Co-Planar Stereotaxic Atlas of the Human Brain, Georg Thieme, Stuttgart, New York, 1988.
Talbot JD, Marret S, Evans AC, Meyer E, Bushnell MC, Duncan GH. Multiple representations of pain in human cerebral cortex. Science. 1991;251:1355-1358.
Van Hoesen, G.W., Morecraft, R.J. and Vogt, B.A., Connections of the monkey cingulate cortex. In: B.A. Vogt and M. Gabriel (Eds.), Neurobiology of Cingulate Cortex and Limbic Thalamus: A Comprehensive Treatise, Birkhauser, Boston, MA, 1993, pp. 249–284.
Vogt, B.A., Structural organization of cingulate cortex: areas, neurons, and somatodendritic transmitter receptors. In: B.A. Vogt and M. Gabriel (Eds.), Neurobiology of Cingulate Cortex and Limbic Thalamus: A Comprehensive Treatise, Birkhauser, Boston, MA, 1993, pp. 19–70.
Vogt BA, Pandya DN. Cingulate cortex of the rhesus monkey: II. Cortical afferents. J. Comp. Neurol. 1987;262:271-289.
Vogt BA, Derbyshire SWG, Jones AKP. Pain processing in four regions of human cingulate cortex localized with coregistered PET and MR imaging. Eur. J. Neurosci. 1996;8:1461-1473.
Vogt BA, Pandya DN, Rosene DL. Cingulate cortex of the rhesus monkey: I. Cytoarchitecture and thalamic afferents. J. Comp. Neurol. 1987;262:256-270.
Vogt BA, Finch DM, Olson CR. Functional heterogeneity in cingulate cortex: the anterior executive and posterior evaluative regions. Cerebral Cortex. 1992;2:435-443.
Waddell G. A new clinical model for the treatment of low-back pain. Spine. 1987;12:632-644.
Weiller C, May A, Limmroth V, Juptner M, Kaube H, Schayck RV, Coenen HH, Diener HC. Brain stem activation in spontaneous human migraine attacks. Nat. Med. 1995;7:658-660.
Weinrich M, Wise SP. The premotor cortex of the monkey. J. Neurosci. 1982;2:1329-1345.
Woods RP, Cherry SR, Mazziotta JC. A rapid automated algorithm for accurately aligning and reslicing positron emission tomography images. J. Comput. Assist. Tomogr. 1992;16:620-633.
Zola-Morgan S, Squire LR, Amaral DG. Human amnesia and the medial temporal region: enduring memory impairment following a bilateral lesion limited to the CA1 field of the hippocampus. J. Neurosci. 1989;6:2950-2967.
Keywords:
Pain processing; Noxious stimulation; Stimulus intensity; Regional cerebral blood flow
© Lippincott-Raven Publishers.