- Email: [email protected]
Temporo-spatial analysis of cortical activation by phasic innocuous and noxious cold stimuli – a magnetoencephalographic study
Pain 100 (2002) 281–290 www.elsevier.com/locate/pain
Temporo-spatial analysis of cortical activation by phasic innocuous and noxious cold stimuli – a magnetoencephalographic study Christian Maiho¨fner*, Martin Kaltenha¨user, Bernhard Neundo¨rfer, Eberhard Lang Department of Neurology, University of Erlangen–Nuremberg, Schwabachanlage 6, Universitatsstrasse, D-91054 Erlangen, Germany Received 15 February 2002; accepted 1 August 2002
Abstract Clinical findings and recent non-invasive functional imaging studies pinpoint the insular cortex as the crucial brain area involved in cold sensation. By contrast, the role of primary (SI) and secondary (SII) somatosensory cortices in central processing of cold is controversial. So far, temporal activation patterns of cortical areas involved in cold processing have not been examined. Using magnetoencephalography, we studied, in seven healthy subjects, the temporo-spatial dynamics of brain processes evoked by innocuous and noxious cold stimulation as compared to tactile stimuli. For this purpose, a newly designed and magnetically silent cold-stimulator was employed. In separate runs, cold and painful cold stimuli were delivered to the dorsum of the right hand. Tactile afferents were stimulated by pneumatic tactile stimulation. Following innocuous cold stimulation (DT ¼ 5 ^ 0:38C in 50 ^ 2 ms), magnetic source imaging revealed an exclusive activation of the contra- and ipsilateral posterior insular cortex. The mean peak latencies were 194.3 ^ 38.1 and 241.0 ^ 31.7 ms for the response in the ipsiand contralateral insular cortex, respectively. Based on the measurement of onset latencies, the estimated conduction velocity of peripheral nerve fibres mediating cold fell in the range of Ad-fibres (7.4 ^ 0.8 m/s). Noxious cold stimulation (DT ¼ 35 ^ 58C in 70 ^ 12 ms) initially activated the contra- and ipsilateral insular cortices in the same latency ranges as innocuous cold stimuli. Additionally, we found an activation of the contra- and ipsilateral SII areas (peak latencies 304 ^ 22.7 and 310.1 ^ 19.4 ms, respectively) and a variable activation of the cingulate cortex. Notably, neither cold- nor painful cold stimulation produced an activation of SI. By contrast, the evoked cortical responses following tactile stimulation could be located to the contralateral SI cortex and bilateral SII. In conclusion, this study strongly corroborates the posterior insular cortex as the primary somatosensory area for cortical processing of cold sensation. Furthermore, it supports the role of SII and the cingulate cortex in mediating freeze-pain. Therefore, these results suggest different processing of cold, freeze-pain and touch in the human brain. q 2002 International Association for the Study of Pain. Published by Elsevier Science B.V. All rights reserved. Keywords: Magnetoencephalography; Somatosensory evoked magnetic fields; Cold; Freeze-pain; Pain
1. Introduction Cold and pain are phylogenetically old sensations that are necessary to sustain body integrity and survival of living beings. Therefore, pathways of cold and pain are intimately associated in the central nervous system (Craig et al., 2000). Despite a large number of animal studies, the processing of cold and cold-induced pain in humans is not clearly understood. Regarding cold, electrophysiological investigations in non-human primates, cats and rodents have revealed evidence for involvement of the medial and lateral thalamus * Corresponding author. Institute of Physiology and Experimental Pathophysiology, Universita¨tsstrasse 17, D-91054 Erlangen, Germany. Tel.: 149-91-3185-22498; fax: 149-91-3185-22497. E-mail address: [email protected] (C. Maiho¨fner).
(Bushnell et al., 1993), the insula and the primary (SI) somatosensory cortex (Lamour et al., 1983; Norrsell and Craig, 1999). The few reports in the literature on noxious thermal stimulation in animals are limited to electrophysiological measurements of neurons in SI (Lamour et al., 1983) and the cingulate cortex (Sikes and Vogt, 1992). Clinical findings from patients with cerebral lesions have pointed to the parieto-insular cortex as the brain area crucially associated with contralateral thermosensory loss (Boivie et al., 1989; Norrsell, 1989; Schmahmann and Leifer, 1992; Greenspan et al., 1999). In contrast, selective lesions of parietal somatosensory areas have little or no effect on thermal perception (Head and Holmes, 1911; Adams and Burke, 1989). These clinical findings are consistent with lesion studies in monkeys (Porter and Semmes, 1974; Porter, 1987)
0304-3959/02/$20.00 q 2002 International Association for the Study of Pain. Published by Elsevier Science B.V. All rights reserved. PII: S 0304-395 9(02)00276-2
282
C. Maiho¨ fner et al. / Pain 100 (2002) 281–290
Electrophysiological data on the processing of cold in the human brain arise from microelectrode recordings performed during neurosurgical interventions. Electrical stimulation of neurons in the thalamic ventromedial nucleus in awake humans evokes graded sensations of cold. These neurons were also shown to respond to innocuous cold stimulation of the skin (Davis et al., 1999). The progress in non-invasive exploration of human brain functions during the past decade mainly results from studies with positron emission tomography (PET) and functional magnetic resonance imaging (fMRI). However, the role of SI, secondary somatosensory cortex (SII) and the insula in the cortical processing of cold is still not fully understood. In PET studies, cold stimuli consistently activated the insula (Casey et al., 1996; Craig et al., 1996, 2000). However, Davis et al. (1998) found an insular activation for only half of the subjects in an fMRI study. Furthermore, the results regarding activation of SI and the secondary somatosensory cortex (SII) substantially differed between these studies. Magnetoencephalography (MEG) combines both high temporal (millisecond range) and spatial (approximately 3–4 mm) resolution (Kakigi et al., 2000). This technique portrays neuronal electrical activity more directly, rather than using intermediates such as changes in blood flow or metabolism. Recently, we reported on pain-related somatosensory evoked magnetic fields induced by controlled ballistic mechanical impact stimulation (Druschky et al., 2000). We were also able to localise insular activations following electrical stimulation of the oesophagus by MEG (Hecht et al., 1999). Therefore, we assumed that a possible insular activation following cold stimulation should be detected by MEG recordings. To investigate the temporo-spatial dynamics of brain processes evoked by innocuous and noxious cold stimulation, we used MEG recordings and applied a magnetically silent cold-specific stimulator of our own design. Data were compared with the known cortical activation patterns following tactile stimulation of the skin. 2. Methods Seven healthy right-handed male volunteers (mean age 29.4 years, range 24–35 years; mean height 175 cm, range 165–187 cm) participated in the experiment. Informed consent was obtained from all participants prior to the experiment and the study adhered to the tenets of the Declaration of Helsinki. All subjects were medical students or physicians and experienced in psychophysical studies. 2.1. Tactile stimulation An air-puff-derived tactile stimulator (9037-953, Biomagnetic Technologies, Inc., San Diego, CA), which provides a light superficial pressure stimulus to the skin surface, was used for tactile stimulation. The skin contact
area was a circular rubber-bladder, 10 mm in diameter, and the intensity of the mechanical stimulation was 40 g/cm 2. The rise time was 20 ms as measured from 10 to 90% of the intensity increment. No joint movement was observed in this stimulation. The stimulation device was fixed to the dorsum of the hand and the respective position was marked for the following cold stimulation. Two hundred stimuli were applied and the interstimulus interval was randomly jittered, from 980 to 1020 ms. 2.2. Cold stimulation The cold-stimulator was constructed by Christian Maiho¨ fner. A schematic drawing of the device is shown in Fig. 1a. A cylindrical copper thermode (0.5 mm thick, diameter 3.5 cm, baseline temperature 328C) was applied to the dorsum of the right hand in the skin territory of the superficial branch of the radial nerve. The temperature of the thermode was measured by a thermocouple. Coppermade gills rising vertical to the thermode were in contact with a copper chamber (20 cm diameter, 4 cm deep). The chamber was filled with tempered water (328C) to increase the warmth capacity of the system and accelerate re-warming of the thermode after cooling stimuli. A plastic tube connected to an aerosol was centred to the rear of the thermode. The aerosol had an aluminium container and a plastic valve, thus it was totally magnetically silent (kindly gifted by the Deutsche Pra¨ zisionsventil GmbH, Hattersheim, Germany). The aerosol was filled with a freezing agent normally used for electronic requirements (Ka¨ ltespray 75, Kontakt Chemie, Jena, Germany). The aerosol was mounted in a wooden framework and the valve could be triggered by a pneumatically driven piston. The feeding air tube was led through a hole in the shielding of the MEG room and connected outside to an electronic valve. This valve regulated the supply of compressed air (0.7 MPa) to the tubesystem and thereby the activation of the piston. The opening time of the valve could be controlled by the stimulation software of our MEG system (see Section 2.3). Initial psychophysical experiments showed that valve opening for 105 ms produced a reproducible cooling of 5 ^ 0.38C in 50 ^ 2 ms. This was associated with a welldefined local cold sensation (Fig. 1b). Valve opening for 150 ms produced a temperature fall of 35 ^ 58C in 70 ^ 12 ms into the sub-zero temperature range (12 to 288C) (Fig. 1b). In this case, the stimulus evoked a sudden, intense, sharp stinging pain sensation in four of the seven subjects. To minimise habituation, we used long inter-stimulus intervals (randomly jittered, from 25 to 35 s). The stimuli were not applied before thermode temperature had recovered to 328C. One hundred stimuli were applied. Audible noise associated with the stimulation was masked by binaural presentation of white noise applied through plastic tubes and ear moulds. Interference of auditory-evoked brain responses was ruled out by initial control experiments show-
C. Maiho¨fner et al. / Pain 100 (2002) 281–290
283
arrays were placed with the centre above C3 and C4, according to the international 10-20 electroencephalogram (EEG) system. Cerebral evoked magnetic fields were recorded in the time interval between 200 ms before and 3000 ms after the stimulus trigger with a bandwidth of 0–200 Hz. The sampling rate of the analogous signals was set to 1041 Hz. For off-line analysis of the signals, the data were filtered (1– 70 Hz, 50 Hz notch) and visually scanned for artefacts. Only epochs without obvious artefacts were averaged. To detect possible radially oriented current sources, to which MEG is blind, we recorded in three subjects, evoked potentials of the EEG from gold disk electrodes (1 cm in diameter), attached to the scalp with electrode jelly at T3 and T4, according to the international 10-20 EEG system. An electrode placed at Fpz was used as reference. Impedance was maintained less than 5 kV. EEG-data was simultaneously recorded with MEG data. Analysis of EEG data was done with the recording software of the MEG system (see above). A bandpass filter of 0.5–70 Hz was applied offline. Afterwards, data were averaged. 2.4. MEG source localisation
Fig. 1(a). A schematic drawing of the cold stimulator (for details see text).(b) Temperature profiles of the two different stimuli used. The arrow indicates the internal trigger of the MEG system. STIM 1 resembles a temperature fall of 5 ^ 0.38C within 50 ^ 2 ms from a baseline temperature of 328C. STIM 2 is a temperature fall of 35 ^ 58C in 70 ^ 12 ms into the sub-zero temperature range (12 to 288C).
ing absent MEG responses after triggering the cold-stimulator inside the shielded MEG room but without any contact with the subject’s skin.
Magnetic source imaging was performed as described previously (Druschky et al., 2000). Briefly, a sphere locally fitted to the head shape underneath the sensor was used as a volume conductor model. Afterwards, a mathematic process based on the Marquardt algorithm (Marquardt, 1963) was applied if the evoked magnetic fields showed an approximate dipolar distribution. Respective time intervals were analysed using the single equivalent current dipole (ECD) model for each side. Only stable clusters of dipoles of at least 10 ms duration were analysed. For each localisation, a correlation coefficient between the measured and the ideal magnetic dipole field was calculated. The pre-chosen criteria for the final dipole selection were a map correlation and a goodness of fit $0.96. Latencies were calculated relative to the time of stimulation. To visualise results with respect to brain anatomy, the dipole locations were superimposed on magnetic resonance (MR) images. A 1.5 T Magnetom (Siemens, Germany) was used and three skin markers were placed at fiducial points on the subject’s head. The location of the same fiducial points was also recorded relative to the neuromagnetometer position, thus establishing a common spatial reference for the transposition of three dimensional coordinates between MEG and MRI data. 2.5. Conduction velocity of cold excited nerve fibres
2.3. Data acquisition and analysis Cortical responses were recorded with a dual 37-channel neuromagnetometer (Magnes II w; 4 D Neuroimaging, Inc., San Diego, CA) in a magnetically shielded room. The detection coils of the neuromagnetometer were arranged in a uniformly distributed array in concentric circles over a spherically concave surface (144 mm diameter). The sensor
For estimation of the conduction velocity of the cold sensation-mediating peripheral nerve fibres, the distances for central (D1) and peripheral (D2) impulse conduction were measured (Fig. 2). The peripheral conduction distance D2 was measured between the stimulation site and the spinal processes of the sixth cervical vertebra with a tape measure. The central conduction distance D1 was calculated by addi-
284
C. Maiho¨ fner et al. / Pain 100 (2002) 281–290
tion of two distances (D1a and D1b). The spinal and medullar conduction distance (D1a) was approximated from the measured distance between the sixth cervical spinous processes and the projection of the midpoint between the two pre-auricular reference points on the dorsal skull surface by means of a tape measure (usually 1–2 cm below the external occipital protuberance). As confirmed previously in one subject by MRI scans of the head and cervical spine, this distance closely matches the real anatomical distances (11.6 versus 11.1 cm). The conduction distance to the posterior part of the contralateral insula (D1b) was measured from the midpoint of the line between the pre-auricular reference points from respective MRI scans of the subjects. The conduction velocity of the human spinothalamic tract fibres was taken as 9.87 m/s, the mean value reported by Rossi et al. (2000). The central conduction time is calculated as the ratio of D1 and 9.87 m/s. The peripheral conduction
time is calculated as the difference between the onset latency of contralateral insular activation and the central conduction time. The peripheral conduction velocity results from the ratio between D2 and the peripheral conduction time. A calculation for one subject is given in Fig. 2. 2.6. Psychophysical evaluation of the stimulation Subjects were instructed to report their sensations and scale them on a numeric rating scale (NRS). The NRS was graded in decimal steps from 0.0 (no cold sensation) to 0.9 (pronounced cold sensation), and in integer steps from 1 (just barely painful) to 10 (intolerably painful). 2.7. Statistics Comparisons of coordinates, latencies and magnetic field strength of the best fitting ECDs between both hemispheres as well as between innocuous and noxious cold stimulation were performed by Wilcoxon matched pairs sign rank test. All values are given as mean ^ standard deviation (SD). 3. Results 3.1. Psychophysics A rapid temperature drop from the baseline temperature of 328C (DT ¼ 5 ^ 0:38C within 50 ^ 2 ms) elicited a sudden sensation of cold at the stimulation site in all subjects. The intensity of the sensation was estimated as 0.82 ^ 0.06 (mean ^ SD) on the NRS. A temperature fall of 35 ^ 58C in 70 ^ 12 ms into the sub-zero temperature range (12 to 288C) elicited a sudden stinging pain sensation at the stimulation site in four of the seven subjects. Two of these subjects additionally reported a feeling of ‘pinprick’ or electric shock. The mean rating score on the NRS was 4.9 ^ 1.7. The other three subjects reported a sudden intense cold sensation but no feeling of pain. 3.2. Evoked magnetic fields following tactile stimulation
Fig. 2. Estimation of the conduction velocity of nerve fibres mediating cold. The central conduction time (cct) was calculated as the ratio of the central nerve distance D1 (sum of the distance between sixth cervical vertebra and the projection of the midpoint between the two pre-auricular points on the dorsal skull surface [D1a] and the distance from the latter point to the contralateral insula [D1b]) and the presumed conduction velocity of the spinothalamic tract (9.87 m/s, according to Rossi et al. (2000)). The peripheral conduction time (pct) is the difference between the onset latency of the cortical insular activation and cct. The peripheral conduction velocity results then with the ratio between D2 (distance between sixth cervical vertebra and stimulation site) and pct. An example is given for one subject. The table details the values of the seven subjects. In summary, the average conduction velocity of cold-specific nerve fibres is calculated as 7.4 ^ 0.8 m/s.
Tactile stimulation of the skin on the right hand dorsum evoked four clear deflections of the magnetic waveforms from baseline over the contralateral hemisphere (Fig. 3a). The peak latency of the first component (1M) was 35.4 ^ 2.9 ms. The duration and the strength of the magnetic field (root mean square [RMS], specified in femto Tesla [fT]) of 1M were 24.5 ^ 1.8 and 124.5 ^ 1.8 fT, respectively. The corresponding ECDs for the first three responses (1M–3M) were located close together in the hand area of the contralateral SI cortex (Fig. 3b). By contrast, long latency components were detected over both the contralateral (4M) and ipsilateral (4M(I)) hemispheres (Fig. 3a). The peak latency of the contralateral side was 121.4 ^ 2.9 ms (4 M) and of the ipsilateral 123.7 ^ 2,8 ms (4M(I)). In contrast to 1M–3M, the corresponding ECDs could be allocated to the superior bank
C. Maiho¨ fner et al. / Pain 100 (2002) 281–290
285
lapped with the respective MRI scans, they could be located in the left posterolateral insular region and the direction of the dipoles was from posteromedial to antero-latero-caudal (Fig. 4b). In five of the seven subjects, there was a clear signal deflection from baseline with the dewar positioned ipsilateral to the stimulation side, i.e. over C4. This response occurred approximately 45 ms later than the deflection detected over the C3-position (Fig. 4a). The peak latency, duration and strength of the evoked magnetic field (RMS) for the ipsilateral response were 241.0 ^ 31.7 ms, 121.3 ^ 10.1 ms and 97.2 ^ 12.4 fT, respectively. The corresponding field distribution was dipolar (Fig. 4a) and the ECDs could be located in the insular cortex ipsilateral to the stimulation side (Fig. 4b). The reported sensation after cold stimulation did not differ between the subjects with and without detectable ipsilateral insular responses. Average EEGs recorded at T3 (contralateral) and T4 (ipsilateral) showed prominent positive potentials clearly exceeding background noise that corresponded to the cold evoked SEF recorded by MEG. The mean peak latencies, durations and amplitudes for the potentials recorded at T3 were 195.5 ^ 12.7 ms, 137.5 ^ 17.7 ms, 9.9 ^ 2.4 mV and for the potentials at T4, they were 274.5 ^ 12.7 ms, 111.3 ^ 9.1 ms, 5.5 ^ 1.2 mV, respectively. 3.4. Conduction velocity of cold-excited nerve fibres
Fig. 3. (a) Chart showing SEF following tactile stimulation of the dorsum of the right hand. Waveforms recorded at 37 channels of each sensor are superimposed (right and left hemispheres, i.e. C4 and C3 position of the centres of the dewars). Stimulus onset is indicated by an arrow. Four components are identified over the contralateral (1M–4M) and one over the ipsilateral hemisphere (4M(I)).(b) MRIs showing the location and direction of ECDs of 1–4M and 4M(I). Symbols correspond to SEF waveforms in (a) 1M–3M are located close together in the hand area of SI, 4M and 4M(I) in the SII areas, contra- and ipsilaterally, respectively.
The mean onset latency of the contralateral insular activation was 158.4 ^ 13 ms. Using the individual values, we tried to calculate the conduction velocities of peripheral nerve fibres mediating cold (see Section 2.5 and Fig. 2). The average peripheral conduction velocity of the fastest excited cold-specific nerve fibres in our subjects was estimated to be 7.4 ^ 0.8 m/s. In contrast, using the same arithmetic operation, the conduction velocity of tactile excited nerve fibres was calculated to be 74.7 ^ 3 ms.
of the sylvian fissure, an area that is most likely to represent the SII cortex (Fig. 3b).
3.5. Evoked magnetic fields following noxious cold stimulation
3.3. Evoked magnetic fields following innocuous cold stimulation
Three subjects reported no feeling of stinging pain following a temperature fall of 35 ^ 58C. Consistently, the corresponding evoked magnetic fields were comparable to those following a temperature drop of 5 ^ 0.38C. The detected patterns of cerebral evoked responses in the four subjects who reported a sharp stinging feeling of pain were more complex (Fig. 5a). With the sensor position above the contralateral hemisphere, two main deflections were clearly identified in each subject, with peak latencies of 191.2 ^ 35.6 and 300.0 ^ 22.7 ms, respectively. The ECDs of the first response were located in the posterior part of the insula and those of the second in the superior bank of the Sylvian fissure, corresponding to the SII region. Ipsilateral to the stimulation side, in three of four subjects, two deflections were recorded with peak latencies of 235.0 ^ 20.5 and 310.1 ^ 19.4 ms, respectively (Fig. 5a).
Somatosensory evoked magnetic fields (SEF) following non-painful cold stimulation of the dorsum of the right hand with a temperature fall of 5 ^ 0.38C in 50 ^ 2 ms differed markedly from those obtained by tactile stimulation. In all subjects, a clear single deflection of the baseline was detected with the sensor above C3, i.e. contralateral to the stimulation site (Fig. 4a). The mean peak latency of this activation was 194.3 ^ 38.1 ms after onset of the cold stimulation. The duration of the signal was 154.3 ^ 33.1 ms and the strength of the magnetic field (RMS) was 111.1 ^ 12.9 fT. Isomagnetic field maps of this evoked response showed a clear dipolar field distribution (Fig. 4a). When the locations of their ECDs were over-
286
C. Maiho¨ fner et al. / Pain 100 (2002) 281–290
The ECDs of the first response were located in the posterior part of the insula and those of the second response in the SII area as well (Fig. 5b).
The strengths of the magnetic fields representing the contralateral insular activation after noxious cold stimulation (141 ^ 35.5 fT) tended to be greater compared to those obtained following innocuous (111.1 ^ 12.9 fT) cold stimulation. However, the difference was not statistically significant. In two of the subjects, a clear polarity reversal third deflection was observed occurring at 405 and 412 ms, respectively, after onset of the noxious cold stimulation (Fig. 5a). These responses could be allocated into the cingulate gyrus. A small third deflection was also seen in a third subject, but the ECDs did not fulfil the set criteria (goodness of fit , 92, map correlation , 90%). A synopsis of the mean values and SDs of coordinates (cm), peak latencies (ms), and magnetic field strength (RMS, fT) of the ECDs with the best fit to the Marquardt algorithm as well as the durations of the recorded deflections are summarised in Fig. 6a for innocuous cold and in Fig. 6b for noxious cold stimulation.
4. Discussion 4.1. MEG patterns of innocuous cold and touch sensation
Fig. 4. (a) Superimposed averaged waveforms (37 channels of each sensor) over the right and left hemispheres following innocuous cold stimulation of the dorsum of the right hand of a representative subject. Stimulus onset is indicated by an arrow. A clear deflection of the baseline was detected with the sensor above C3, i.e. contralateral to the stimulation site, the peak latency in this subject was 190.2 ms. A second signal deflection was recorded with the dewar position ipsilateral to the stimulation side, i.e. over C4. The peak latency of this response was 232.4 ms in this subject. Isomagnetic field maps at indicated peak latencies showed a clear dipolar field distribution. (b) Localisation and orientation of the ECDs resembling the two responses shown in (a). The response after cold stimulation on the right hand dorsum was allocated to the posterior part of the left insula (filled circle). The ECD of the ipsilateral response was estimated to be in the posterior part of the right insula (filled square). Symbols correspond to SEF waveforms in Fig. 3a
It is accepted that phasic innocuous cold stimuli applied to the skin activate cold-specific Ad nerve fibres (Dostrovsky and Craig, 1996; Campero et al., 2001). We found the first cortical activation following phasic cold stimulation after an average peak latency of 194 ms in the contralateral posterior insular cortex. This response was followed by a subsequent activation of the ipsilateral insular cortex with a delay of approximately 45 ms. We were unable to detect responses in brain regions other than the contra- and ipsilateral insula, particularly not in SI or SII. Also in averaged EEG recordings, only one single activation with a comparable peak latency was detected. This complements the MEG recordings, as purely radial current sources cannot be detected by MEG. Thus, we conclude that the primary sensory area for innocuous cold stimuli of the skin is the contralateral posterior part of the insular cortex. Both the spatial and the temporal patterns of cold-specific brain activation presented here corroborate previous findings from functional imaging and EEG studies. Following tonic-innocuous cold stimulation, bilateral activation of the middle and posterior parts of the insular regions has been reported in PET- and fMRI studies (Casey et al., 1996; Davis et al., 1998; Craig et al., 2000). Our MEG latencies of insular activations fit well with the previous EEG findings of Jamal et al. (1989). Following innocuous cold stimulation of the right hand, a positive deflection of the EEG from baseline could be recorded from scalp electrodes over the contralateral centro-parietal region after an average peak latency of 184 ^ 22.3 ms. Also, a cold evoked response over the ipsilateral centro-parietal region with a delay of approximately 45 ms was detected (Jamal et al., 1989).
C. Maiho¨ fner et al. / Pain 100 (2002) 281–290
287
Fig. 5. (a) Representative SEFs recorded at the C3 (contralateral) and C4 (ipsilateral) position following noxious cold stimulation of the dorsum of the right hand. The waveforms of all 37 channels of each sensor are superimposed. The stimulus onset is indicated by an arrow. Symbols correspond to the ECD locations in (b). Contralaterally, two polarity-reversal deflections were recorded, a third small deflection was not localisable (*). Ipsilaterally, three clear components can be distinguished.(b) Localisations of the ECDs corresponding to the respective deflections in (a). The ECDs were allocated to the ipsi- and contralateral insula (open and filled square; 4/4 subjects), the upper bank of the Sylvian fissure in both hemispheres (open and filled triangle; 3/4 and 4/4 subjects, respectively), and the cingulate cortex (filled circle; 2/4 subjects). Peak latencies are given in parenthesis.
This delay is longer than the known activations between contra- and ipsilateral SI and SII after tactile stimuli. The shortest trancallosal transmission between the two SI areas has been estimated as 6–7 ms on the basis of direct cortical recordings (Noachtar et al., 1997). Ipsilateral SII activation following painful heat stimuli to the skin (CO2 laser) was recorded intracortically 15 ms after activation of the contralateral SII area (Frot and Mauguiere, 1999). In contrast, long differences in peak latencies between the two insular cortices, which were comparable with those in the present study, have been reported in previous MEG studies following enteroceptive stimulation (Hecht et al., 1999; Aziz et al., 2000). A possible explanation for the differences in the interhemispheric time delays after activation of SI, SII and the insular cortex may be differences in the temporal summation or other mechanisms of central integration before the actual transcallosal or transcommisural transmission. Alternatively, the ipsilateral insula might also be accessed via an ipsilateral spino-thalamo-insular pathway with a neuronal relay in the brainstem or a slower conduction velocity than the spinothalamic tract. An ipsilateral tract mediating cold stimuli has been demonstrated in animals (Craig, 1991). However, definite proof is lacking in humans. In patients, after contralateral hemispherectomy, ipsilateral
activation of SII has been reported following various kinds of somatosensory stimuli (Bittar et al., 2000), suggesting a relevance of ipsilateral somatosensory pathways at least in pathological states. In the present study, we did not find an activation of SII with innocuous cold stimuli. This agrees with the fMRI study of Davis et al. (1998) and with the PET study of Casey et al. (1996). It does not agree with a PET study of Craig et al. (1996) who found a strong SII activation with innocuous cold stimulation. However, lesion studies indicate that SII plays no essential role in thermoperception. Dong et al. (1996) reported no change in innocuous thermosensitivity following experimental compression of the posterior parietal cortex (including SII). Moreover, Porter (1987) showed that SII is not essential for the discrimination of temperature differences. One possible explanation for SII activation in functional imaging studies could be mechanical co-stimulation (via slowly adapting mechanoreceptors) during application of the cooling stimuli itself. In an elegant study design, Craig et al. (2000) used regression correlation against the temperature of several graded cooling stimuli to convincingly demonstrate that the posterior insular cortex is responsible for discriminative thermal sensation. Therefore, the SII cortex seems to play no, or only a minor, role in temperature perception.
288
C. Maiho¨ fner et al. / Pain 100 (2002) 281–290
Based on the measurement of onset latencies of insular activations and the presumed conduction velocity of
spinothalamic tract fibres (Rossi et al., 2000), we also calculated the conduction velocity of peripheral nerve fibres mediating cold. The average conduction velocity of the fastest-excited cold-specific nerve fibres in our subjects fell within 7.4 ^ 0.8 m/s clearly in the range of Ad fibres. In contrast, the conduction velocity of tactile-excited nerve fibres was calculated to be 74.7 ^ 3 ms, consistent with the activation of A-b-fibres. Thereby, our results corroborate and extend indirect estimations of peripheral conduction velocities of cold-specific fibres based on the measurement of reaction times (Yarnitsky and Ochoa, 1991; Susser et al., 1999).
4.2. MEG patterns after noxious cold stimulation
Fig. 6. Synopsis of all activations. (a) Mean coordinates (A) for the ECDs (x; y; z) following innocuous cold stimulation (i.e. contra- and ipsilateral insula; n ¼ 7) and tactile stimulation (first response, i.e. SI cortex). SDs are indicated by the length of arrows in respective directions. The origin of the coordinate system is the midpoint between the two pre-auricular reference points (see inset). A positive x-value is towards the nasion, a positive yvalue is towards the left pre-auricular point, and a positive z-value is towards the vertex. B, C and D indicate mean peak latencies, durations and strengths of the magnetic fields (RMS), respectively. (b) Mean coordinates (A) for the ECDs (x; y; z) following noxious cold stimulation (n ¼ 4) (i.e. contra- and ipsilateral insula, contra- and ipsilateral SII and cingulate cortex). SDs are indicated by the length of arrows in respective directions. B, C and D indicate mean peak latencies, durations and strengths of the magnetic fields (RMS), respectively.
Noxious cold stimuli initially activated the posterior parts of the contralateral and ipsilateral insula. The peak latencies between non-noxious and noxious cold stimulations did not differ. This finding suggests that activation of the contralateral insula following noxious stimulation results from Ad nerve fibres rather than from nociceptive C-fibre afferents. However, as cold-specific and polymodal-nociceptive Ad nerve fibres are likely to be coactivated by noxious cold stimulation, insular activation cannot be solely explained by noxious input. Insular projections of these two different fibre systems might differ. In our experiments, no anatomical differences of insular activations following innocuous or noxious cold stimulation were detected (see Fig. 6a, b). Nevertheless, this has to be interpreted with caution, as subtle differences in neuronal activation patterns may be not detected by MEG. This is due to the spatial resolution of our MEG system which is in the range of 3–4 mm (Kakigi et al., 2000). Therefore, histological studies seem to be more appropriate to address this question. The quality of the pain was described as stinging or pricking and the sensation was perceived suddenly. Therefore, the stimulus differs markedly from the cold pain induced by tonic stimulation with cooling agents of less than 158C (Casey et al., 1996). The term ‘freeze-pain’ appears to be appropriate for this kind of stimulation. Based on the estimates of conduction velocity and on the blockade of freezepain-sensation during an A-fibre pressure block, freeze-pain is likely to be mediated by Ad-fibres (Beise et al., 1998). In the study of Beise et al. (1998), an abrupt and rapid jump in temperature was reported with the onset of the pain. These exothermic reactions were not observed in our study, most probably because of the short duration of the stimuli which was in the range of milliseconds. In humans, microneurography and intraneural microstimulation of identified cutaneous primary afferent fibres demonstrated that electrical stimulation of nociceptive Ad-fibres evokes a sensation of sharp pain. In contrast, stimulation of nociceptive C-fibres produced a sensation of dull or burning pain (Ochoa and Torebjork, 1989). These results fit well with comparable peak latencies of insular activations following innocuous
C. Maiho¨ fner et al. / Pain 100 (2002) 281–290
and noxious cold stimulation and in turn suggest an Ad fibre pathway. The MEG patterns that were found to be specific for painful cold stimuli to the skin were a bilateral activation of SII and, in two subjects, of the cingulate gyrus. Activation of the contralateral SII area following noxious cold was recorded, on average, 109 ms after activation of the contralateral insula. This corresponds to the average delay of 100 ms between activation of the SI and SII areas following tactile stimulation (Ploner et al., 1999). This could implicate a serial processing of noxious cold stimuli in the insula and SII which is comparable to the serial cortical processing of painful ballistic or electrical skin stimulation in SI and SII (Druschky et al., 2000; Kakigi et al., 2000). Involvement of SII in pain processing is suggested from electrophysiological and behavioural responses following noxious stimuli in monkeys (Robinson and Burton, 1980; Chudler et al., 1986; Dong et al., 1989) as well as from studies using functional imaging techniques (for review see Peyron et al., 2000). However, the present MEG study shows SII activations, after phasic noxious cold stimulation, more consistently than investigations using PET (Casey et al., 1996, 2000) or fMRI (Davis et al., 1998). PET studies showed no SII activation after induction of tonic-deep coldpain following immersion of one hand in ice water, (Casey et al., 1996) or only contralateral SII activation to the stimulation side (Casey et al., 2000). Similarly, in an fMRI study with tonic cold-pain (Davis et al., 1998), SII was activated in only 50% of subjects. As discussed by Davis et al. (1998), the missing activations in the fMRI study may be due to technical difficulties in observing the signal because of the small numbers of nociceptive neurons in SII and the spatial resolution of fMRI. This may also account for the variable activation of SII in PET studies (Casey et al., 1996, 2000). Possibly, the phasic noxious cold stimuli in the present MEG study may evoke stronger cortical responses than the tonic-noxious cold stimuli thereby improving the signal-to-noise ratio. Furthermore, phasic stimuli may have avoided possible habituation of afferent brain regions to the stimulus over time. Activation of the cingulate gyrus is regularly reported in the context of painful stimuli (Vogt et al., 1996; Rainville et al., 1997; Coghill et al., 1994; Hsieh et al., 1999) and is broadly considered to reflect the aversive component of pain (Rainville et al., 1997; Bromm and Lorenz, 1998). Cingulate gyrus activation following tonic-noxious cold stimulation was reported in studies using PET and fMRI (Craig et al., 1996; Davis et al., 1998; Casey et al., 1996; Kwan et al., 2000; Peyron et al., 2000). The present MEG results compare with these studies by detecting cingulate gyrus activation following phasic noxious cold stimulation in two of four subjects who felt stinging pain. Detection of sources in the cingulate gyrus by MEG has always been difficult due to the radial orientation of the sources and deep localisation (Kakigi et al., 2000). However, MEG allows insight into temporal activation patterns following
289
defined painful stimuli. Compared to cingulate gyrus activation after mechanical impact (Druschky et al., 2000), electrical or painful laser stimuli (reviewed by Bromm and Lorenz, 1998), the latencies of cingulate gyrus activation following phasic noxious cold stimuli in the present study were long (407 and 412 ms). This may be explained by the long stimulus duration of 70 ^ 12 ms, as this parameter was shown to be closely linked to the peak latency of activations associated with the cingulate gyrus (Bromm and Lorenz, 1998). Accordingly, using laser-evoked-potentials, Gibson et al. (1991) also found a peak latency of 400 ms for the P2 response with a stimulus duration of up to 66 ms. Therefore, the relatively long stimulus duration of 70 ms may contribute to the comparably late activation of the cingulate gyrus in this study. Taken together, the results of this study strongly corroborate a role of the insular cortex in the perception of cold. In contrast to tactile stimulation, no activation of SI was observed following cold stimulation. Following noxious cold stimulation, an additional activation of SII and the cingulate gyrus was observed. Therefore, this study hints at a differential processing of cold, freeze-pain and touch in the human brain. Acknowledgements We thank Dr Lothar Kohllo¨ ffel (Department of Physiology I, Erlangen-Nuremberg) and Dr Clive Brown (Department of Neurology) for helpful suggestions and corrections. We are grateful to Sieglinde Jakobi for her technical support during MEG recordings and data organisation. References Adams RW, Burke D. Deficits of thermal sensation in patients with unilateral cerebral lesions. Electroenceph clin Neurophysiol 1989;73:443– 452. Aziz Q, Schnitzler A, Enck P. Functional neuroimaging of visceral sensation. J Clin Neurophysiol 2000;17:604–612. Beise RD, Carstens E, Kohlloffel LU. Psychophysical study of stinging pain evoked by brief freezing of superficial skin and ensuing short-lasting changes in sensations of cool and cold pain. Pain 1998;74:275–286. Bittar RG, Ptito A, Reutens DC. Somatosensory representation in patients who have undergone hemispherectomy: a functional magnetic resonance imaging study. J Neurosurg 2000;92:45–51. Boivie J, Leijon G, Johansson I. Central post-stroke pain – a study of the mechanisms through analyses of the sensory abnormalities. Pain 1989;37:173–185. Bromm B, Lorenz J. Neurophysiological evaluation of pain. Electroenceph clin Neurophysiol 1998;107:227–253. Bushnell MC, Duncan GH, Tremblay N. ThalamicVPM nucleus in the behaving monkey. J Neurophysiol 1993;69:739–752. Campero M, Serra J, Bostock H, Ochoa JL. Slowly conducting afferents activated by innocuous low temperature in human skin. J Physiol 2001;535:855–865. Casey KL, Minoshima S, Morrow TJ, Koeppe RA. Comparison of human cerebral activation pattern during cutaneous warmth, heat pain, and deep cold pain. J Neurophysiol 1996;76:571–581. Casey KL, Svensson P, Morrow TJ, Raz J, Jone C, Minoshima S. Selective
290
C. Maiho¨ fner et al. / Pain 100 (2002) 281–290
opiate modulation of nociceptive processing in the human brain. J Neurophysiol 2000;84:525–533. Chudler EH, Dong WK, Kawakami Y. Cortical nociceptive responses and behavioral correlates in the monkey. Brain Res 1986;397:47–60. 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. Craig AD. Spinal distribution of ascending lamina I axons anterogradely labeled with Phaseolus vulgaris leucoagglutinin (PHA-L) in the cat. J Comp Neurol 1991;313:377–393. Craig AD, Chen K, Bandy D, Reiman EM. Thermosensory activation of insular cortex. Nat Neurosci 2000;3:184–190. Craig AD, Reiman EM, Evans A, Bushnell MC. Functional imaging of an illusion of pain. Nature 1996;384:258–260. Davis KD, Kwan CL, Crawley AP, Mikulis DJ. Functional MRI study of thalamic and cortical activations evoked by cutaneous heat, cold, and tactile stimuli. J Neurophysiol 1998;80:1533–1546. Davis KD, Lozano RM, Manduch M, Tasker RR, Kiss ZH, Dostrovsky JO. Thalamic relay site for cold perception in humans. J Neurophysiol 1999;81:1970–1973. Dong WK, Hayashi T, Roberts VJ, Fusco BM, Chudler EH. Behavioral outcome of posterior parietal cortex injury in the monkey. Pain 1996;64:579–587. Dong WK, Salonen LD, Kawakami Y, Shiwaku T, Kaukoranta EM, Martin RF. Nociceptive responses of trigeminal neurons in SII-7b cortex of awake monkeys. Brain Res 1989;484:314–324. Dostrovsky JO, Craig AD. Cooling-specific spinothalamic neurons in the monkey. J Neurophysiol 1996;76:3656–3665. Druschky K, Lang E, Hummel C, Kaltenhauser M, Kohlloffel LU, Neundorfer B, Stefan H. Pain-related somatosensory evoked magnetic fields induced by controlled ballistic mechanical impacts. J Clin Neurophysiol 2000;17:613–622. Frot M, Mauguiere F. Timing and spatial distribution of somatosensory responses recorded in the upper bank of the sylvian fissure (SII area) in humans. Cereb Cortex 1999;9:854–863. Gibson SJ, LeVasseur SA, Helme RD. Cerebral event-related responses induced by CO2 laser stimulation in subjects suffering from cervicobrachial syndrome. Pain 1991;47:173–182. Greenspan JD, Lee RR, Lenz FA. Pain sensitivity alterations as a function of lesion location in the parasylvian cortex. Pain 1999;81:273–282. Head H, Holmes G. Sensory disturbances from cerebral lesions. Brain 1911;34:102–254. Hecht M, Kober H, Claus D, Hilz M, Vieth J, Neundorfer B. The electrical and magnetical cerebral responses evoked by electrical stimulation of the esophagus and the location of their cerebral sources. Clin Neurophysiol 1999;110:1435–1444. Hsieh JC, Stone-Elander S, Ingvar M. Anticipatory coping of pain expressed in the human anterior cingulate cortex: a positron emission tomography study. Neurosci Lett 1999;262:61–64. Jamal GA, Hansen S, Weir AI, Ballantyne JP. Cerebral cortical potentials to pure non-painful temperature stimulation: an objective technique for the assessment of small fibre pathway in man. J Neurol Neurosurg Psychiatry 1989;52:99–105. Kakigi R, Hoshiyama M, Shimojo M, Naka D, Yamasaki H, Watanabe S,
Xiang J, Maeda K, Lam K, Itomi K, Nakamura A. The somatosensory evoked magnetic fields. Prog Neurobiol 2000;61:495–523. Kwan CL, Crawley AP, Mikulis DJ, Davis KD. An fMRI study of the anterior cingulate cortex and surrounding medial wall activations evoked by noxious cutaneous heat and cold stimuli. Pain 2000;85:359–374. Lamour Y, Willer JC, Guilbaud G. Rat somatosensory (SmI) cortex. Exp Brain Res 1983;49:35–45. Marquardt DW. An algorithm for least-squares estimation of non-linear parameters. J Soc Ind Appl Math 1963;11:431–441. Noachtar S, Luders HO, Dinner DS, Klem G. Ipsilateral median somatosensory evoked potentials recorded from human somatosensory cortex. Electroenceph clin Neurophysiol 1997;104:189–198. Norrsell U. Behavioural thermosensitivity after unilateral, partial lesions of the lateral funiculus in the cervical spinal cord of the cat. Exp Brain Res 1989;78:369–373. Norrsell U, Craig AD. Behavioral thermosensitivity after lesions of thalamic target areas of a thermosensory spinothalamic pathway in the cat. J Neurophysiol 1999;82:611–625. Ochoa J, Torebjork E. Sensations evoked by intraneural microstimulation of C nociceptor fibres in human skin nerves. J Physiol 1989;415:583– 599. Peyron R, Laurent B, Garcia-Larrea L. Functional imaging of brain responses to pain. A review and meta-analysis. Neurophysiol Clin 2000;30:263–288. Ploner M, Schmitz F, Freund HJ, Schnitzler A. Parallel activation of primary and secondary somatosensory cortices in human pain processing. J Neurophysiol 1999;81:3100–3104. Porter L, Semmes J. Preservation of cutaneous temperature sensitivity after ablation of sensory cortex in monkeys. Exp Neurol 1974;42:206–219. Porter LH. An experimental investigation of the parietal lobes and temperature discrimination in monkeys. Brain Res 1987;412:54–67. Rainville P, Duncan GH, Price DD, Carrier B, Bushnell MC. Pain affect encoded in human anterior cingulate but not somatosensory cortex. Science 1997;277:968–971. Robinson CJ, Burton H. Organization of somatosensory receptive fields in cortical areas 7b, retroinsula, postauditory and granular insula of M. fascicularis. J Comp Neurol 1980;192:69–92. Rossi P, Serrao M, Amabile G, Parisi L, Pierelli F, Pozzessere G. A simple method for estimating conduction velocity of the spinothalamic tract in healthy humans. Clin Neurophysiol 2000;111:1907–1915. Schmahmann JD, Leifer D. Parietal pseudothalamic pain syndrome. Clinical features and anatomic correlates. Arch Neurol 1992;49:1032–1037. Sikes RW, Vogt BA. Nociceptive neurons in area 24 of rabbit cingulate cortex. J Neurophysiol 1992;68:1720–1732. Susser E, Sprecher E, Yarnitsky D. Paradoxical heat sensation in healthy subjects: peripherally conducted by A delta or C fibres? Brain 1999;122:239–246. Vogt BA, Derbyshire S, Jones AK. Pain processing in our regions of human cingulate cortex localised with co-registered PET and MR imaging. Eur J Neurosci 1996;8:1461–1473. Yarnitsky D, Ochoa JL. Warm and cold-specific somatosensory systems. Psychophysical thresholds, reaction times and peripheral conduction velocities. Brain 1991;114:1819–1826.
Temporo-spatial analysis of cortical activation by phasic innocuous and noxious cold stimuli – a magnetoencephalographic study Christian Maiho¨fner*, Martin Kaltenha¨user, Bernhard Neundo¨rfer, Eberhard Lang Department of Neurology, University of Erlangen–Nuremberg, Schwabachanlage 6, Universitatsstrasse, D-91054 Erlangen, Germany Received 15 February 2002; accepted 1 August 2002
Abstract Clinical findings and recent non-invasive functional imaging studies pinpoint the insular cortex as the crucial brain area involved in cold sensation. By contrast, the role of primary (SI) and secondary (SII) somatosensory cortices in central processing of cold is controversial. So far, temporal activation patterns of cortical areas involved in cold processing have not been examined. Using magnetoencephalography, we studied, in seven healthy subjects, the temporo-spatial dynamics of brain processes evoked by innocuous and noxious cold stimulation as compared to tactile stimuli. For this purpose, a newly designed and magnetically silent cold-stimulator was employed. In separate runs, cold and painful cold stimuli were delivered to the dorsum of the right hand. Tactile afferents were stimulated by pneumatic tactile stimulation. Following innocuous cold stimulation (DT ¼ 5 ^ 0:38C in 50 ^ 2 ms), magnetic source imaging revealed an exclusive activation of the contra- and ipsilateral posterior insular cortex. The mean peak latencies were 194.3 ^ 38.1 and 241.0 ^ 31.7 ms for the response in the ipsiand contralateral insular cortex, respectively. Based on the measurement of onset latencies, the estimated conduction velocity of peripheral nerve fibres mediating cold fell in the range of Ad-fibres (7.4 ^ 0.8 m/s). Noxious cold stimulation (DT ¼ 35 ^ 58C in 70 ^ 12 ms) initially activated the contra- and ipsilateral insular cortices in the same latency ranges as innocuous cold stimuli. Additionally, we found an activation of the contra- and ipsilateral SII areas (peak latencies 304 ^ 22.7 and 310.1 ^ 19.4 ms, respectively) and a variable activation of the cingulate cortex. Notably, neither cold- nor painful cold stimulation produced an activation of SI. By contrast, the evoked cortical responses following tactile stimulation could be located to the contralateral SI cortex and bilateral SII. In conclusion, this study strongly corroborates the posterior insular cortex as the primary somatosensory area for cortical processing of cold sensation. Furthermore, it supports the role of SII and the cingulate cortex in mediating freeze-pain. Therefore, these results suggest different processing of cold, freeze-pain and touch in the human brain. q 2002 International Association for the Study of Pain. Published by Elsevier Science B.V. All rights reserved. Keywords: Magnetoencephalography; Somatosensory evoked magnetic fields; Cold; Freeze-pain; Pain
1. Introduction Cold and pain are phylogenetically old sensations that are necessary to sustain body integrity and survival of living beings. Therefore, pathways of cold and pain are intimately associated in the central nervous system (Craig et al., 2000). Despite a large number of animal studies, the processing of cold and cold-induced pain in humans is not clearly understood. Regarding cold, electrophysiological investigations in non-human primates, cats and rodents have revealed evidence for involvement of the medial and lateral thalamus * Corresponding author. Institute of Physiology and Experimental Pathophysiology, Universita¨tsstrasse 17, D-91054 Erlangen, Germany. Tel.: 149-91-3185-22498; fax: 149-91-3185-22497. E-mail address: [email protected] (C. Maiho¨fner).
(Bushnell et al., 1993), the insula and the primary (SI) somatosensory cortex (Lamour et al., 1983; Norrsell and Craig, 1999). The few reports in the literature on noxious thermal stimulation in animals are limited to electrophysiological measurements of neurons in SI (Lamour et al., 1983) and the cingulate cortex (Sikes and Vogt, 1992). Clinical findings from patients with cerebral lesions have pointed to the parieto-insular cortex as the brain area crucially associated with contralateral thermosensory loss (Boivie et al., 1989; Norrsell, 1989; Schmahmann and Leifer, 1992; Greenspan et al., 1999). In contrast, selective lesions of parietal somatosensory areas have little or no effect on thermal perception (Head and Holmes, 1911; Adams and Burke, 1989). These clinical findings are consistent with lesion studies in monkeys (Porter and Semmes, 1974; Porter, 1987)
0304-3959/02/$20.00 q 2002 International Association for the Study of Pain. Published by Elsevier Science B.V. All rights reserved. PII: S 0304-395 9(02)00276-2
282
C. Maiho¨ fner et al. / Pain 100 (2002) 281–290
Electrophysiological data on the processing of cold in the human brain arise from microelectrode recordings performed during neurosurgical interventions. Electrical stimulation of neurons in the thalamic ventromedial nucleus in awake humans evokes graded sensations of cold. These neurons were also shown to respond to innocuous cold stimulation of the skin (Davis et al., 1999). The progress in non-invasive exploration of human brain functions during the past decade mainly results from studies with positron emission tomography (PET) and functional magnetic resonance imaging (fMRI). However, the role of SI, secondary somatosensory cortex (SII) and the insula in the cortical processing of cold is still not fully understood. In PET studies, cold stimuli consistently activated the insula (Casey et al., 1996; Craig et al., 1996, 2000). However, Davis et al. (1998) found an insular activation for only half of the subjects in an fMRI study. Furthermore, the results regarding activation of SI and the secondary somatosensory cortex (SII) substantially differed between these studies. Magnetoencephalography (MEG) combines both high temporal (millisecond range) and spatial (approximately 3–4 mm) resolution (Kakigi et al., 2000). This technique portrays neuronal electrical activity more directly, rather than using intermediates such as changes in blood flow or metabolism. Recently, we reported on pain-related somatosensory evoked magnetic fields induced by controlled ballistic mechanical impact stimulation (Druschky et al., 2000). We were also able to localise insular activations following electrical stimulation of the oesophagus by MEG (Hecht et al., 1999). Therefore, we assumed that a possible insular activation following cold stimulation should be detected by MEG recordings. To investigate the temporo-spatial dynamics of brain processes evoked by innocuous and noxious cold stimulation, we used MEG recordings and applied a magnetically silent cold-specific stimulator of our own design. Data were compared with the known cortical activation patterns following tactile stimulation of the skin. 2. Methods Seven healthy right-handed male volunteers (mean age 29.4 years, range 24–35 years; mean height 175 cm, range 165–187 cm) participated in the experiment. Informed consent was obtained from all participants prior to the experiment and the study adhered to the tenets of the Declaration of Helsinki. All subjects were medical students or physicians and experienced in psychophysical studies. 2.1. Tactile stimulation An air-puff-derived tactile stimulator (9037-953, Biomagnetic Technologies, Inc., San Diego, CA), which provides a light superficial pressure stimulus to the skin surface, was used for tactile stimulation. The skin contact
area was a circular rubber-bladder, 10 mm in diameter, and the intensity of the mechanical stimulation was 40 g/cm 2. The rise time was 20 ms as measured from 10 to 90% of the intensity increment. No joint movement was observed in this stimulation. The stimulation device was fixed to the dorsum of the hand and the respective position was marked for the following cold stimulation. Two hundred stimuli were applied and the interstimulus interval was randomly jittered, from 980 to 1020 ms. 2.2. Cold stimulation The cold-stimulator was constructed by Christian Maiho¨ fner. A schematic drawing of the device is shown in Fig. 1a. A cylindrical copper thermode (0.5 mm thick, diameter 3.5 cm, baseline temperature 328C) was applied to the dorsum of the right hand in the skin territory of the superficial branch of the radial nerve. The temperature of the thermode was measured by a thermocouple. Coppermade gills rising vertical to the thermode were in contact with a copper chamber (20 cm diameter, 4 cm deep). The chamber was filled with tempered water (328C) to increase the warmth capacity of the system and accelerate re-warming of the thermode after cooling stimuli. A plastic tube connected to an aerosol was centred to the rear of the thermode. The aerosol had an aluminium container and a plastic valve, thus it was totally magnetically silent (kindly gifted by the Deutsche Pra¨ zisionsventil GmbH, Hattersheim, Germany). The aerosol was filled with a freezing agent normally used for electronic requirements (Ka¨ ltespray 75, Kontakt Chemie, Jena, Germany). The aerosol was mounted in a wooden framework and the valve could be triggered by a pneumatically driven piston. The feeding air tube was led through a hole in the shielding of the MEG room and connected outside to an electronic valve. This valve regulated the supply of compressed air (0.7 MPa) to the tubesystem and thereby the activation of the piston. The opening time of the valve could be controlled by the stimulation software of our MEG system (see Section 2.3). Initial psychophysical experiments showed that valve opening for 105 ms produced a reproducible cooling of 5 ^ 0.38C in 50 ^ 2 ms. This was associated with a welldefined local cold sensation (Fig. 1b). Valve opening for 150 ms produced a temperature fall of 35 ^ 58C in 70 ^ 12 ms into the sub-zero temperature range (12 to 288C) (Fig. 1b). In this case, the stimulus evoked a sudden, intense, sharp stinging pain sensation in four of the seven subjects. To minimise habituation, we used long inter-stimulus intervals (randomly jittered, from 25 to 35 s). The stimuli were not applied before thermode temperature had recovered to 328C. One hundred stimuli were applied. Audible noise associated with the stimulation was masked by binaural presentation of white noise applied through plastic tubes and ear moulds. Interference of auditory-evoked brain responses was ruled out by initial control experiments show-
C. Maiho¨fner et al. / Pain 100 (2002) 281–290
283
arrays were placed with the centre above C3 and C4, according to the international 10-20 electroencephalogram (EEG) system. Cerebral evoked magnetic fields were recorded in the time interval between 200 ms before and 3000 ms after the stimulus trigger with a bandwidth of 0–200 Hz. The sampling rate of the analogous signals was set to 1041 Hz. For off-line analysis of the signals, the data were filtered (1– 70 Hz, 50 Hz notch) and visually scanned for artefacts. Only epochs without obvious artefacts were averaged. To detect possible radially oriented current sources, to which MEG is blind, we recorded in three subjects, evoked potentials of the EEG from gold disk electrodes (1 cm in diameter), attached to the scalp with electrode jelly at T3 and T4, according to the international 10-20 EEG system. An electrode placed at Fpz was used as reference. Impedance was maintained less than 5 kV. EEG-data was simultaneously recorded with MEG data. Analysis of EEG data was done with the recording software of the MEG system (see above). A bandpass filter of 0.5–70 Hz was applied offline. Afterwards, data were averaged. 2.4. MEG source localisation
Fig. 1(a). A schematic drawing of the cold stimulator (for details see text).(b) Temperature profiles of the two different stimuli used. The arrow indicates the internal trigger of the MEG system. STIM 1 resembles a temperature fall of 5 ^ 0.38C within 50 ^ 2 ms from a baseline temperature of 328C. STIM 2 is a temperature fall of 35 ^ 58C in 70 ^ 12 ms into the sub-zero temperature range (12 to 288C).
ing absent MEG responses after triggering the cold-stimulator inside the shielded MEG room but without any contact with the subject’s skin.
Magnetic source imaging was performed as described previously (Druschky et al., 2000). Briefly, a sphere locally fitted to the head shape underneath the sensor was used as a volume conductor model. Afterwards, a mathematic process based on the Marquardt algorithm (Marquardt, 1963) was applied if the evoked magnetic fields showed an approximate dipolar distribution. Respective time intervals were analysed using the single equivalent current dipole (ECD) model for each side. Only stable clusters of dipoles of at least 10 ms duration were analysed. For each localisation, a correlation coefficient between the measured and the ideal magnetic dipole field was calculated. The pre-chosen criteria for the final dipole selection were a map correlation and a goodness of fit $0.96. Latencies were calculated relative to the time of stimulation. To visualise results with respect to brain anatomy, the dipole locations were superimposed on magnetic resonance (MR) images. A 1.5 T Magnetom (Siemens, Germany) was used and three skin markers were placed at fiducial points on the subject’s head. The location of the same fiducial points was also recorded relative to the neuromagnetometer position, thus establishing a common spatial reference for the transposition of three dimensional coordinates between MEG and MRI data. 2.5. Conduction velocity of cold excited nerve fibres
2.3. Data acquisition and analysis Cortical responses were recorded with a dual 37-channel neuromagnetometer (Magnes II w; 4 D Neuroimaging, Inc., San Diego, CA) in a magnetically shielded room. The detection coils of the neuromagnetometer were arranged in a uniformly distributed array in concentric circles over a spherically concave surface (144 mm diameter). The sensor
For estimation of the conduction velocity of the cold sensation-mediating peripheral nerve fibres, the distances for central (D1) and peripheral (D2) impulse conduction were measured (Fig. 2). The peripheral conduction distance D2 was measured between the stimulation site and the spinal processes of the sixth cervical vertebra with a tape measure. The central conduction distance D1 was calculated by addi-
284
C. Maiho¨ fner et al. / Pain 100 (2002) 281–290
tion of two distances (D1a and D1b). The spinal and medullar conduction distance (D1a) was approximated from the measured distance between the sixth cervical spinous processes and the projection of the midpoint between the two pre-auricular reference points on the dorsal skull surface by means of a tape measure (usually 1–2 cm below the external occipital protuberance). As confirmed previously in one subject by MRI scans of the head and cervical spine, this distance closely matches the real anatomical distances (11.6 versus 11.1 cm). The conduction distance to the posterior part of the contralateral insula (D1b) was measured from the midpoint of the line between the pre-auricular reference points from respective MRI scans of the subjects. The conduction velocity of the human spinothalamic tract fibres was taken as 9.87 m/s, the mean value reported by Rossi et al. (2000). The central conduction time is calculated as the ratio of D1 and 9.87 m/s. The peripheral conduction
time is calculated as the difference between the onset latency of contralateral insular activation and the central conduction time. The peripheral conduction velocity results from the ratio between D2 and the peripheral conduction time. A calculation for one subject is given in Fig. 2. 2.6. Psychophysical evaluation of the stimulation Subjects were instructed to report their sensations and scale them on a numeric rating scale (NRS). The NRS was graded in decimal steps from 0.0 (no cold sensation) to 0.9 (pronounced cold sensation), and in integer steps from 1 (just barely painful) to 10 (intolerably painful). 2.7. Statistics Comparisons of coordinates, latencies and magnetic field strength of the best fitting ECDs between both hemispheres as well as between innocuous and noxious cold stimulation were performed by Wilcoxon matched pairs sign rank test. All values are given as mean ^ standard deviation (SD). 3. Results 3.1. Psychophysics A rapid temperature drop from the baseline temperature of 328C (DT ¼ 5 ^ 0:38C within 50 ^ 2 ms) elicited a sudden sensation of cold at the stimulation site in all subjects. The intensity of the sensation was estimated as 0.82 ^ 0.06 (mean ^ SD) on the NRS. A temperature fall of 35 ^ 58C in 70 ^ 12 ms into the sub-zero temperature range (12 to 288C) elicited a sudden stinging pain sensation at the stimulation site in four of the seven subjects. Two of these subjects additionally reported a feeling of ‘pinprick’ or electric shock. The mean rating score on the NRS was 4.9 ^ 1.7. The other three subjects reported a sudden intense cold sensation but no feeling of pain. 3.2. Evoked magnetic fields following tactile stimulation
Fig. 2. Estimation of the conduction velocity of nerve fibres mediating cold. The central conduction time (cct) was calculated as the ratio of the central nerve distance D1 (sum of the distance between sixth cervical vertebra and the projection of the midpoint between the two pre-auricular points on the dorsal skull surface [D1a] and the distance from the latter point to the contralateral insula [D1b]) and the presumed conduction velocity of the spinothalamic tract (9.87 m/s, according to Rossi et al. (2000)). The peripheral conduction time (pct) is the difference between the onset latency of the cortical insular activation and cct. The peripheral conduction velocity results then with the ratio between D2 (distance between sixth cervical vertebra and stimulation site) and pct. An example is given for one subject. The table details the values of the seven subjects. In summary, the average conduction velocity of cold-specific nerve fibres is calculated as 7.4 ^ 0.8 m/s.
Tactile stimulation of the skin on the right hand dorsum evoked four clear deflections of the magnetic waveforms from baseline over the contralateral hemisphere (Fig. 3a). The peak latency of the first component (1M) was 35.4 ^ 2.9 ms. The duration and the strength of the magnetic field (root mean square [RMS], specified in femto Tesla [fT]) of 1M were 24.5 ^ 1.8 and 124.5 ^ 1.8 fT, respectively. The corresponding ECDs for the first three responses (1M–3M) were located close together in the hand area of the contralateral SI cortex (Fig. 3b). By contrast, long latency components were detected over both the contralateral (4M) and ipsilateral (4M(I)) hemispheres (Fig. 3a). The peak latency of the contralateral side was 121.4 ^ 2.9 ms (4 M) and of the ipsilateral 123.7 ^ 2,8 ms (4M(I)). In contrast to 1M–3M, the corresponding ECDs could be allocated to the superior bank
C. Maiho¨ fner et al. / Pain 100 (2002) 281–290
285
lapped with the respective MRI scans, they could be located in the left posterolateral insular region and the direction of the dipoles was from posteromedial to antero-latero-caudal (Fig. 4b). In five of the seven subjects, there was a clear signal deflection from baseline with the dewar positioned ipsilateral to the stimulation side, i.e. over C4. This response occurred approximately 45 ms later than the deflection detected over the C3-position (Fig. 4a). The peak latency, duration and strength of the evoked magnetic field (RMS) for the ipsilateral response were 241.0 ^ 31.7 ms, 121.3 ^ 10.1 ms and 97.2 ^ 12.4 fT, respectively. The corresponding field distribution was dipolar (Fig. 4a) and the ECDs could be located in the insular cortex ipsilateral to the stimulation side (Fig. 4b). The reported sensation after cold stimulation did not differ between the subjects with and without detectable ipsilateral insular responses. Average EEGs recorded at T3 (contralateral) and T4 (ipsilateral) showed prominent positive potentials clearly exceeding background noise that corresponded to the cold evoked SEF recorded by MEG. The mean peak latencies, durations and amplitudes for the potentials recorded at T3 were 195.5 ^ 12.7 ms, 137.5 ^ 17.7 ms, 9.9 ^ 2.4 mV and for the potentials at T4, they were 274.5 ^ 12.7 ms, 111.3 ^ 9.1 ms, 5.5 ^ 1.2 mV, respectively. 3.4. Conduction velocity of cold-excited nerve fibres
Fig. 3. (a) Chart showing SEF following tactile stimulation of the dorsum of the right hand. Waveforms recorded at 37 channels of each sensor are superimposed (right and left hemispheres, i.e. C4 and C3 position of the centres of the dewars). Stimulus onset is indicated by an arrow. Four components are identified over the contralateral (1M–4M) and one over the ipsilateral hemisphere (4M(I)).(b) MRIs showing the location and direction of ECDs of 1–4M and 4M(I). Symbols correspond to SEF waveforms in (a) 1M–3M are located close together in the hand area of SI, 4M and 4M(I) in the SII areas, contra- and ipsilaterally, respectively.
The mean onset latency of the contralateral insular activation was 158.4 ^ 13 ms. Using the individual values, we tried to calculate the conduction velocities of peripheral nerve fibres mediating cold (see Section 2.5 and Fig. 2). The average peripheral conduction velocity of the fastest excited cold-specific nerve fibres in our subjects was estimated to be 7.4 ^ 0.8 m/s. In contrast, using the same arithmetic operation, the conduction velocity of tactile excited nerve fibres was calculated to be 74.7 ^ 3 ms.
of the sylvian fissure, an area that is most likely to represent the SII cortex (Fig. 3b).
3.5. Evoked magnetic fields following noxious cold stimulation
3.3. Evoked magnetic fields following innocuous cold stimulation
Three subjects reported no feeling of stinging pain following a temperature fall of 35 ^ 58C. Consistently, the corresponding evoked magnetic fields were comparable to those following a temperature drop of 5 ^ 0.38C. The detected patterns of cerebral evoked responses in the four subjects who reported a sharp stinging feeling of pain were more complex (Fig. 5a). With the sensor position above the contralateral hemisphere, two main deflections were clearly identified in each subject, with peak latencies of 191.2 ^ 35.6 and 300.0 ^ 22.7 ms, respectively. The ECDs of the first response were located in the posterior part of the insula and those of the second in the superior bank of the Sylvian fissure, corresponding to the SII region. Ipsilateral to the stimulation side, in three of four subjects, two deflections were recorded with peak latencies of 235.0 ^ 20.5 and 310.1 ^ 19.4 ms, respectively (Fig. 5a).
Somatosensory evoked magnetic fields (SEF) following non-painful cold stimulation of the dorsum of the right hand with a temperature fall of 5 ^ 0.38C in 50 ^ 2 ms differed markedly from those obtained by tactile stimulation. In all subjects, a clear single deflection of the baseline was detected with the sensor above C3, i.e. contralateral to the stimulation site (Fig. 4a). The mean peak latency of this activation was 194.3 ^ 38.1 ms after onset of the cold stimulation. The duration of the signal was 154.3 ^ 33.1 ms and the strength of the magnetic field (RMS) was 111.1 ^ 12.9 fT. Isomagnetic field maps of this evoked response showed a clear dipolar field distribution (Fig. 4a). When the locations of their ECDs were over-
286
C. Maiho¨ fner et al. / Pain 100 (2002) 281–290
The ECDs of the first response were located in the posterior part of the insula and those of the second response in the SII area as well (Fig. 5b).
The strengths of the magnetic fields representing the contralateral insular activation after noxious cold stimulation (141 ^ 35.5 fT) tended to be greater compared to those obtained following innocuous (111.1 ^ 12.9 fT) cold stimulation. However, the difference was not statistically significant. In two of the subjects, a clear polarity reversal third deflection was observed occurring at 405 and 412 ms, respectively, after onset of the noxious cold stimulation (Fig. 5a). These responses could be allocated into the cingulate gyrus. A small third deflection was also seen in a third subject, but the ECDs did not fulfil the set criteria (goodness of fit , 92, map correlation , 90%). A synopsis of the mean values and SDs of coordinates (cm), peak latencies (ms), and magnetic field strength (RMS, fT) of the ECDs with the best fit to the Marquardt algorithm as well as the durations of the recorded deflections are summarised in Fig. 6a for innocuous cold and in Fig. 6b for noxious cold stimulation.
4. Discussion 4.1. MEG patterns of innocuous cold and touch sensation
Fig. 4. (a) Superimposed averaged waveforms (37 channels of each sensor) over the right and left hemispheres following innocuous cold stimulation of the dorsum of the right hand of a representative subject. Stimulus onset is indicated by an arrow. A clear deflection of the baseline was detected with the sensor above C3, i.e. contralateral to the stimulation site, the peak latency in this subject was 190.2 ms. A second signal deflection was recorded with the dewar position ipsilateral to the stimulation side, i.e. over C4. The peak latency of this response was 232.4 ms in this subject. Isomagnetic field maps at indicated peak latencies showed a clear dipolar field distribution. (b) Localisation and orientation of the ECDs resembling the two responses shown in (a). The response after cold stimulation on the right hand dorsum was allocated to the posterior part of the left insula (filled circle). The ECD of the ipsilateral response was estimated to be in the posterior part of the right insula (filled square). Symbols correspond to SEF waveforms in Fig. 3a
It is accepted that phasic innocuous cold stimuli applied to the skin activate cold-specific Ad nerve fibres (Dostrovsky and Craig, 1996; Campero et al., 2001). We found the first cortical activation following phasic cold stimulation after an average peak latency of 194 ms in the contralateral posterior insular cortex. This response was followed by a subsequent activation of the ipsilateral insular cortex with a delay of approximately 45 ms. We were unable to detect responses in brain regions other than the contra- and ipsilateral insula, particularly not in SI or SII. Also in averaged EEG recordings, only one single activation with a comparable peak latency was detected. This complements the MEG recordings, as purely radial current sources cannot be detected by MEG. Thus, we conclude that the primary sensory area for innocuous cold stimuli of the skin is the contralateral posterior part of the insular cortex. Both the spatial and the temporal patterns of cold-specific brain activation presented here corroborate previous findings from functional imaging and EEG studies. Following tonic-innocuous cold stimulation, bilateral activation of the middle and posterior parts of the insular regions has been reported in PET- and fMRI studies (Casey et al., 1996; Davis et al., 1998; Craig et al., 2000). Our MEG latencies of insular activations fit well with the previous EEG findings of Jamal et al. (1989). Following innocuous cold stimulation of the right hand, a positive deflection of the EEG from baseline could be recorded from scalp electrodes over the contralateral centro-parietal region after an average peak latency of 184 ^ 22.3 ms. Also, a cold evoked response over the ipsilateral centro-parietal region with a delay of approximately 45 ms was detected (Jamal et al., 1989).
C. Maiho¨ fner et al. / Pain 100 (2002) 281–290
287
Fig. 5. (a) Representative SEFs recorded at the C3 (contralateral) and C4 (ipsilateral) position following noxious cold stimulation of the dorsum of the right hand. The waveforms of all 37 channels of each sensor are superimposed. The stimulus onset is indicated by an arrow. Symbols correspond to the ECD locations in (b). Contralaterally, two polarity-reversal deflections were recorded, a third small deflection was not localisable (*). Ipsilaterally, three clear components can be distinguished.(b) Localisations of the ECDs corresponding to the respective deflections in (a). The ECDs were allocated to the ipsi- and contralateral insula (open and filled square; 4/4 subjects), the upper bank of the Sylvian fissure in both hemispheres (open and filled triangle; 3/4 and 4/4 subjects, respectively), and the cingulate cortex (filled circle; 2/4 subjects). Peak latencies are given in parenthesis.
This delay is longer than the known activations between contra- and ipsilateral SI and SII after tactile stimuli. The shortest trancallosal transmission between the two SI areas has been estimated as 6–7 ms on the basis of direct cortical recordings (Noachtar et al., 1997). Ipsilateral SII activation following painful heat stimuli to the skin (CO2 laser) was recorded intracortically 15 ms after activation of the contralateral SII area (Frot and Mauguiere, 1999). In contrast, long differences in peak latencies between the two insular cortices, which were comparable with those in the present study, have been reported in previous MEG studies following enteroceptive stimulation (Hecht et al., 1999; Aziz et al., 2000). A possible explanation for the differences in the interhemispheric time delays after activation of SI, SII and the insular cortex may be differences in the temporal summation or other mechanisms of central integration before the actual transcallosal or transcommisural transmission. Alternatively, the ipsilateral insula might also be accessed via an ipsilateral spino-thalamo-insular pathway with a neuronal relay in the brainstem or a slower conduction velocity than the spinothalamic tract. An ipsilateral tract mediating cold stimuli has been demonstrated in animals (Craig, 1991). However, definite proof is lacking in humans. In patients, after contralateral hemispherectomy, ipsilateral
activation of SII has been reported following various kinds of somatosensory stimuli (Bittar et al., 2000), suggesting a relevance of ipsilateral somatosensory pathways at least in pathological states. In the present study, we did not find an activation of SII with innocuous cold stimuli. This agrees with the fMRI study of Davis et al. (1998) and with the PET study of Casey et al. (1996). It does not agree with a PET study of Craig et al. (1996) who found a strong SII activation with innocuous cold stimulation. However, lesion studies indicate that SII plays no essential role in thermoperception. Dong et al. (1996) reported no change in innocuous thermosensitivity following experimental compression of the posterior parietal cortex (including SII). Moreover, Porter (1987) showed that SII is not essential for the discrimination of temperature differences. One possible explanation for SII activation in functional imaging studies could be mechanical co-stimulation (via slowly adapting mechanoreceptors) during application of the cooling stimuli itself. In an elegant study design, Craig et al. (2000) used regression correlation against the temperature of several graded cooling stimuli to convincingly demonstrate that the posterior insular cortex is responsible for discriminative thermal sensation. Therefore, the SII cortex seems to play no, or only a minor, role in temperature perception.
288
C. Maiho¨ fner et al. / Pain 100 (2002) 281–290
Based on the measurement of onset latencies of insular activations and the presumed conduction velocity of
spinothalamic tract fibres (Rossi et al., 2000), we also calculated the conduction velocity of peripheral nerve fibres mediating cold. The average conduction velocity of the fastest-excited cold-specific nerve fibres in our subjects fell within 7.4 ^ 0.8 m/s clearly in the range of Ad fibres. In contrast, the conduction velocity of tactile-excited nerve fibres was calculated to be 74.7 ^ 3 ms, consistent with the activation of A-b-fibres. Thereby, our results corroborate and extend indirect estimations of peripheral conduction velocities of cold-specific fibres based on the measurement of reaction times (Yarnitsky and Ochoa, 1991; Susser et al., 1999).
4.2. MEG patterns after noxious cold stimulation
Fig. 6. Synopsis of all activations. (a) Mean coordinates (A) for the ECDs (x; y; z) following innocuous cold stimulation (i.e. contra- and ipsilateral insula; n ¼ 7) and tactile stimulation (first response, i.e. SI cortex). SDs are indicated by the length of arrows in respective directions. The origin of the coordinate system is the midpoint between the two pre-auricular reference points (see inset). A positive x-value is towards the nasion, a positive yvalue is towards the left pre-auricular point, and a positive z-value is towards the vertex. B, C and D indicate mean peak latencies, durations and strengths of the magnetic fields (RMS), respectively. (b) Mean coordinates (A) for the ECDs (x; y; z) following noxious cold stimulation (n ¼ 4) (i.e. contra- and ipsilateral insula, contra- and ipsilateral SII and cingulate cortex). SDs are indicated by the length of arrows in respective directions. B, C and D indicate mean peak latencies, durations and strengths of the magnetic fields (RMS), respectively.
Noxious cold stimuli initially activated the posterior parts of the contralateral and ipsilateral insula. The peak latencies between non-noxious and noxious cold stimulations did not differ. This finding suggests that activation of the contralateral insula following noxious stimulation results from Ad nerve fibres rather than from nociceptive C-fibre afferents. However, as cold-specific and polymodal-nociceptive Ad nerve fibres are likely to be coactivated by noxious cold stimulation, insular activation cannot be solely explained by noxious input. Insular projections of these two different fibre systems might differ. In our experiments, no anatomical differences of insular activations following innocuous or noxious cold stimulation were detected (see Fig. 6a, b). Nevertheless, this has to be interpreted with caution, as subtle differences in neuronal activation patterns may be not detected by MEG. This is due to the spatial resolution of our MEG system which is in the range of 3–4 mm (Kakigi et al., 2000). Therefore, histological studies seem to be more appropriate to address this question. The quality of the pain was described as stinging or pricking and the sensation was perceived suddenly. Therefore, the stimulus differs markedly from the cold pain induced by tonic stimulation with cooling agents of less than 158C (Casey et al., 1996). The term ‘freeze-pain’ appears to be appropriate for this kind of stimulation. Based on the estimates of conduction velocity and on the blockade of freezepain-sensation during an A-fibre pressure block, freeze-pain is likely to be mediated by Ad-fibres (Beise et al., 1998). In the study of Beise et al. (1998), an abrupt and rapid jump in temperature was reported with the onset of the pain. These exothermic reactions were not observed in our study, most probably because of the short duration of the stimuli which was in the range of milliseconds. In humans, microneurography and intraneural microstimulation of identified cutaneous primary afferent fibres demonstrated that electrical stimulation of nociceptive Ad-fibres evokes a sensation of sharp pain. In contrast, stimulation of nociceptive C-fibres produced a sensation of dull or burning pain (Ochoa and Torebjork, 1989). These results fit well with comparable peak latencies of insular activations following innocuous
C. Maiho¨ fner et al. / Pain 100 (2002) 281–290
and noxious cold stimulation and in turn suggest an Ad fibre pathway. The MEG patterns that were found to be specific for painful cold stimuli to the skin were a bilateral activation of SII and, in two subjects, of the cingulate gyrus. Activation of the contralateral SII area following noxious cold was recorded, on average, 109 ms after activation of the contralateral insula. This corresponds to the average delay of 100 ms between activation of the SI and SII areas following tactile stimulation (Ploner et al., 1999). This could implicate a serial processing of noxious cold stimuli in the insula and SII which is comparable to the serial cortical processing of painful ballistic or electrical skin stimulation in SI and SII (Druschky et al., 2000; Kakigi et al., 2000). Involvement of SII in pain processing is suggested from electrophysiological and behavioural responses following noxious stimuli in monkeys (Robinson and Burton, 1980; Chudler et al., 1986; Dong et al., 1989) as well as from studies using functional imaging techniques (for review see Peyron et al., 2000). However, the present MEG study shows SII activations, after phasic noxious cold stimulation, more consistently than investigations using PET (Casey et al., 1996, 2000) or fMRI (Davis et al., 1998). PET studies showed no SII activation after induction of tonic-deep coldpain following immersion of one hand in ice water, (Casey et al., 1996) or only contralateral SII activation to the stimulation side (Casey et al., 2000). Similarly, in an fMRI study with tonic cold-pain (Davis et al., 1998), SII was activated in only 50% of subjects. As discussed by Davis et al. (1998), the missing activations in the fMRI study may be due to technical difficulties in observing the signal because of the small numbers of nociceptive neurons in SII and the spatial resolution of fMRI. This may also account for the variable activation of SII in PET studies (Casey et al., 1996, 2000). Possibly, the phasic noxious cold stimuli in the present MEG study may evoke stronger cortical responses than the tonic-noxious cold stimuli thereby improving the signal-to-noise ratio. Furthermore, phasic stimuli may have avoided possible habituation of afferent brain regions to the stimulus over time. Activation of the cingulate gyrus is regularly reported in the context of painful stimuli (Vogt et al., 1996; Rainville et al., 1997; Coghill et al., 1994; Hsieh et al., 1999) and is broadly considered to reflect the aversive component of pain (Rainville et al., 1997; Bromm and Lorenz, 1998). Cingulate gyrus activation following tonic-noxious cold stimulation was reported in studies using PET and fMRI (Craig et al., 1996; Davis et al., 1998; Casey et al., 1996; Kwan et al., 2000; Peyron et al., 2000). The present MEG results compare with these studies by detecting cingulate gyrus activation following phasic noxious cold stimulation in two of four subjects who felt stinging pain. Detection of sources in the cingulate gyrus by MEG has always been difficult due to the radial orientation of the sources and deep localisation (Kakigi et al., 2000). However, MEG allows insight into temporal activation patterns following
289
defined painful stimuli. Compared to cingulate gyrus activation after mechanical impact (Druschky et al., 2000), electrical or painful laser stimuli (reviewed by Bromm and Lorenz, 1998), the latencies of cingulate gyrus activation following phasic noxious cold stimuli in the present study were long (407 and 412 ms). This may be explained by the long stimulus duration of 70 ^ 12 ms, as this parameter was shown to be closely linked to the peak latency of activations associated with the cingulate gyrus (Bromm and Lorenz, 1998). Accordingly, using laser-evoked-potentials, Gibson et al. (1991) also found a peak latency of 400 ms for the P2 response with a stimulus duration of up to 66 ms. Therefore, the relatively long stimulus duration of 70 ms may contribute to the comparably late activation of the cingulate gyrus in this study. Taken together, the results of this study strongly corroborate a role of the insular cortex in the perception of cold. In contrast to tactile stimulation, no activation of SI was observed following cold stimulation. Following noxious cold stimulation, an additional activation of SII and the cingulate gyrus was observed. Therefore, this study hints at a differential processing of cold, freeze-pain and touch in the human brain. Acknowledgements We thank Dr Lothar Kohllo¨ ffel (Department of Physiology I, Erlangen-Nuremberg) and Dr Clive Brown (Department of Neurology) for helpful suggestions and corrections. We are grateful to Sieglinde Jakobi for her technical support during MEG recordings and data organisation. References Adams RW, Burke D. Deficits of thermal sensation in patients with unilateral cerebral lesions. Electroenceph clin Neurophysiol 1989;73:443– 452. Aziz Q, Schnitzler A, Enck P. Functional neuroimaging of visceral sensation. J Clin Neurophysiol 2000;17:604–612. Beise RD, Carstens E, Kohlloffel LU. Psychophysical study of stinging pain evoked by brief freezing of superficial skin and ensuing short-lasting changes in sensations of cool and cold pain. Pain 1998;74:275–286. Bittar RG, Ptito A, Reutens DC. Somatosensory representation in patients who have undergone hemispherectomy: a functional magnetic resonance imaging study. J Neurosurg 2000;92:45–51. Boivie J, Leijon G, Johansson I. Central post-stroke pain – a study of the mechanisms through analyses of the sensory abnormalities. Pain 1989;37:173–185. Bromm B, Lorenz J. Neurophysiological evaluation of pain. Electroenceph clin Neurophysiol 1998;107:227–253. Bushnell MC, Duncan GH, Tremblay N. ThalamicVPM nucleus in the behaving monkey. J Neurophysiol 1993;69:739–752. Campero M, Serra J, Bostock H, Ochoa JL. Slowly conducting afferents activated by innocuous low temperature in human skin. J Physiol 2001;535:855–865. Casey KL, Minoshima S, Morrow TJ, Koeppe RA. Comparison of human cerebral activation pattern during cutaneous warmth, heat pain, and deep cold pain. J Neurophysiol 1996;76:571–581. Casey KL, Svensson P, Morrow TJ, Raz J, Jone C, Minoshima S. Selective
290
C. Maiho¨ fner et al. / Pain 100 (2002) 281–290
opiate modulation of nociceptive processing in the human brain. J Neurophysiol 2000;84:525–533. Chudler EH, Dong WK, Kawakami Y. Cortical nociceptive responses and behavioral correlates in the monkey. Brain Res 1986;397:47–60. 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. Craig AD. Spinal distribution of ascending lamina I axons anterogradely labeled with Phaseolus vulgaris leucoagglutinin (PHA-L) in the cat. J Comp Neurol 1991;313:377–393. Craig AD, Chen K, Bandy D, Reiman EM. Thermosensory activation of insular cortex. Nat Neurosci 2000;3:184–190. Craig AD, Reiman EM, Evans A, Bushnell MC. Functional imaging of an illusion of pain. Nature 1996;384:258–260. Davis KD, Kwan CL, Crawley AP, Mikulis DJ. Functional MRI study of thalamic and cortical activations evoked by cutaneous heat, cold, and tactile stimuli. J Neurophysiol 1998;80:1533–1546. Davis KD, Lozano RM, Manduch M, Tasker RR, Kiss ZH, Dostrovsky JO. Thalamic relay site for cold perception in humans. J Neurophysiol 1999;81:1970–1973. Dong WK, Hayashi T, Roberts VJ, Fusco BM, Chudler EH. Behavioral outcome of posterior parietal cortex injury in the monkey. Pain 1996;64:579–587. Dong WK, Salonen LD, Kawakami Y, Shiwaku T, Kaukoranta EM, Martin RF. Nociceptive responses of trigeminal neurons in SII-7b cortex of awake monkeys. Brain Res 1989;484:314–324. Dostrovsky JO, Craig AD. Cooling-specific spinothalamic neurons in the monkey. J Neurophysiol 1996;76:3656–3665. Druschky K, Lang E, Hummel C, Kaltenhauser M, Kohlloffel LU, Neundorfer B, Stefan H. Pain-related somatosensory evoked magnetic fields induced by controlled ballistic mechanical impacts. J Clin Neurophysiol 2000;17:613–622. Frot M, Mauguiere F. Timing and spatial distribution of somatosensory responses recorded in the upper bank of the sylvian fissure (SII area) in humans. Cereb Cortex 1999;9:854–863. Gibson SJ, LeVasseur SA, Helme RD. Cerebral event-related responses induced by CO2 laser stimulation in subjects suffering from cervicobrachial syndrome. Pain 1991;47:173–182. Greenspan JD, Lee RR, Lenz FA. Pain sensitivity alterations as a function of lesion location in the parasylvian cortex. Pain 1999;81:273–282. Head H, Holmes G. Sensory disturbances from cerebral lesions. Brain 1911;34:102–254. Hecht M, Kober H, Claus D, Hilz M, Vieth J, Neundorfer B. The electrical and magnetical cerebral responses evoked by electrical stimulation of the esophagus and the location of their cerebral sources. Clin Neurophysiol 1999;110:1435–1444. Hsieh JC, Stone-Elander S, Ingvar M. Anticipatory coping of pain expressed in the human anterior cingulate cortex: a positron emission tomography study. Neurosci Lett 1999;262:61–64. Jamal GA, Hansen S, Weir AI, Ballantyne JP. Cerebral cortical potentials to pure non-painful temperature stimulation: an objective technique for the assessment of small fibre pathway in man. J Neurol Neurosurg Psychiatry 1989;52:99–105. Kakigi R, Hoshiyama M, Shimojo M, Naka D, Yamasaki H, Watanabe S,
Xiang J, Maeda K, Lam K, Itomi K, Nakamura A. The somatosensory evoked magnetic fields. Prog Neurobiol 2000;61:495–523. Kwan CL, Crawley AP, Mikulis DJ, Davis KD. An fMRI study of the anterior cingulate cortex and surrounding medial wall activations evoked by noxious cutaneous heat and cold stimuli. Pain 2000;85:359–374. Lamour Y, Willer JC, Guilbaud G. Rat somatosensory (SmI) cortex. Exp Brain Res 1983;49:35–45. Marquardt DW. An algorithm for least-squares estimation of non-linear parameters. J Soc Ind Appl Math 1963;11:431–441. Noachtar S, Luders HO, Dinner DS, Klem G. Ipsilateral median somatosensory evoked potentials recorded from human somatosensory cortex. Electroenceph clin Neurophysiol 1997;104:189–198. Norrsell U. Behavioural thermosensitivity after unilateral, partial lesions of the lateral funiculus in the cervical spinal cord of the cat. Exp Brain Res 1989;78:369–373. Norrsell U, Craig AD. Behavioral thermosensitivity after lesions of thalamic target areas of a thermosensory spinothalamic pathway in the cat. J Neurophysiol 1999;82:611–625. Ochoa J, Torebjork E. Sensations evoked by intraneural microstimulation of C nociceptor fibres in human skin nerves. J Physiol 1989;415:583– 599. Peyron R, Laurent B, Garcia-Larrea L. Functional imaging of brain responses to pain. A review and meta-analysis. Neurophysiol Clin 2000;30:263–288. Ploner M, Schmitz F, Freund HJ, Schnitzler A. Parallel activation of primary and secondary somatosensory cortices in human pain processing. J Neurophysiol 1999;81:3100–3104. Porter L, Semmes J. Preservation of cutaneous temperature sensitivity after ablation of sensory cortex in monkeys. Exp Neurol 1974;42:206–219. Porter LH. An experimental investigation of the parietal lobes and temperature discrimination in monkeys. Brain Res 1987;412:54–67. Rainville P, Duncan GH, Price DD, Carrier B, Bushnell MC. Pain affect encoded in human anterior cingulate but not somatosensory cortex. Science 1997;277:968–971. Robinson CJ, Burton H. Organization of somatosensory receptive fields in cortical areas 7b, retroinsula, postauditory and granular insula of M. fascicularis. J Comp Neurol 1980;192:69–92. Rossi P, Serrao M, Amabile G, Parisi L, Pierelli F, Pozzessere G. A simple method for estimating conduction velocity of the spinothalamic tract in healthy humans. Clin Neurophysiol 2000;111:1907–1915. Schmahmann JD, Leifer D. Parietal pseudothalamic pain syndrome. Clinical features and anatomic correlates. Arch Neurol 1992;49:1032–1037. Sikes RW, Vogt BA. Nociceptive neurons in area 24 of rabbit cingulate cortex. J Neurophysiol 1992;68:1720–1732. Susser E, Sprecher E, Yarnitsky D. Paradoxical heat sensation in healthy subjects: peripherally conducted by A delta or C fibres? Brain 1999;122:239–246. Vogt BA, Derbyshire S, Jones AK. Pain processing in our regions of human cingulate cortex localised with co-registered PET and MR imaging. Eur J Neurosci 1996;8:1461–1473. Yarnitsky D, Ochoa JL. Warm and cold-specific somatosensory systems. Psychophysical thresholds, reaction times and peripheral conduction velocities. Brain 1991;114:1819–1826.