INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In humans, painful sensations can be evoked by application of thermal, mechanical, and chemical stimuli to the veins (Arndt and Klement 1991; Fruhstorfer and Lindblom 1983). These painful sensations have been suggested to be signaled by polymodal nociceptors of the vein wall (Arndt and Klement 1991). Correspondingly, anatomical studies in animals showed a sensory innervation of blood vessels (Hinsey 1928; Lim et al. 1962; Polley 1955; Truex 1936; Woollard 1926) presumably involved in vascular nociception (Bazett and McGlone 1928; Glaser 1926; Moore and Moore 1933; Moore and Singleton 1933; Odermatt 1922). Physiological studies confirmed that thinly myelinated and unmyelinated afferents of the veins respond to thermal (Minut-Sorokhtina and Glebova 1976), mechanical (Davenport and Thompson 1987; Göder et al. 1993; Michaelis et al. 1994), and chemical (Göder et al. 1993; Michaelis et al. 1994) stimuli of high, presumably noxious intensities. Functionally, these afferents may be particularly relevant under pathophysiological conditions (Göder et al. 1993; Michaelis et al. 1994). In cats, stimulation of these afferents elicits responses in the sensory-motor cortex (Thompson et al. 1980). However, in humans the cortical representation of venous nociception is unknown.

Painful stimulation of cutaneous nociceptive afferents has been shown to activate the contralateral primary (SI) and bilateral secondary (SII) somatosensory cortices that are included in a broad network associated with nocieption (for review see Schnitzler and Ploner 2000). Magnetoencephalographic (MEG) investigations revealed nearly simultaneous activation of these cortical areas peaking at 160-180 ms after cutaneous application of selective nociceptive laser stimuli to the hand (Ploner et al. 1999, 2000; Timmermann et al. 2001). This activation pattern suggests parallel thalamocortical distribution of cutaneous nociceptive information to SI and SII.

Here, by using whole-head magnetoencephalography we directly compared cortical responses to cutaneous and intravenous painful laser stimuli applied to the same site on the dorsum of the hand in healthy human subjects. We aimed to investigate whether venous nociception shares the cortical representation of cutaneous nociception in humans.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Six healthy right-handed volunteers aged between 31 and 38 yr (mean 34 yr) participated in the study. None of the subjects had a history of neurological or psychiatric disorders. Informed consent was obtained from all subjects before participation. The study was approved by the local ethics committee and conducted in conformity with the declaration of Helsinki.

Stimulation

Forty painful cutaneous laser stimuli (Bromm et al. 1984) were delivered to the skin of the dorsum of the hand. Depending on the individual anatomy of the veins and its accessability for venous puncture, in four subjects the right hand and in two subjects the left hand was chosen for stimulation. The laser device was a Tm:YAG-laser (Baasel Lasertech) with a wavelength of 2,000 nm and a pulse duration of 1 ms. The optical fiber leading the laser beam into the recording room was connected to a handpiece resulting in a spot diameter of 6 mm. Stimulation site was sligthly changed within an area of 4 × 3 cm after each stimulus. Interstimulus intervals were randomly varied between 10 and 14 s. Applied stimulus intensity was 600 mJ evoking moderately painful sensations.

In a separate run, 40 painful laser stimuli were intravenously applied to the dorsum of the same hand as in the cutaneous stimulation condition. The same laser stimulator with a wavelength of 2,000 nm and a pulse duration of 1 ms was used. In this condition, the bare end of the optical fiber with a diameter of 0.6 mm was led through a cannula of 1.4 mm OD into the veins. A puncture site distal to a vein crossing was chosen so that the tip of the fiber was targeted to the vein wall. The distance between the bare end of the optical fiber in the intravenous stimulation condition and the center of the stimulated area in the cutaneous stimulation condition was not allowed to exceed 5 cm in proximal-distal direction. The intravenous position of the optical fiber was continuously controlled and if necessary corrected by one of the investigators who was present in the magnetically shielded room throughout the experiment. Interstimulus intervals were randomly varied between 30 and 60 s. To obtain a maximum signal-to-noise ratio, stimulus intensity was adjusted to evoke moderately to severely painful sensations corresponding to intensities between 6 and 8 on a visual analog scale between 0 (no pain) and 10 (maximally tolerable pain) resulting in stimulus intensities between 500 and 800 mJ.

In both conditions, any stimulus-related noise was masked by white noise applied to subject's ears through plastic tubes.

Data acquisition and analysis

Cortical activity was recorded with a Neuromag-122 whole-head neuromagnetometer (Ahonen et al. 1993) in a magnetically shielded room. The helmet-shaped sensor array contains 122 planar gradiometers that detect the largest signals just above the local cortical current sources. Signals were recorded with a 0.03-Hz high-pass filter and digitally low-pass filtered at 40 Hz. Cortical responses were averaged time locked to stimulus application. Vertical electrooculogram was used to reject epochs contaminated with blink artifacts. Analysis of evoked responses was focused on an epoch comprising 100-ms prestimulus baseline and 250 ms after stimulation. Sources of evoked responses were modeled as equivalent current dipoles identified during clearly dipolar field patterns. Only sources with 95% confidence limits of source localization <10 mm were accepted. Dipole location, orientation, and strength were calculated within a spherical conductor model of each subject's head determined from the individual magnetic resonance images (MRI) acquired on a 1.5 T Siemens-Magnetom. Dipoles were introduced into a spatiotemporal source model where locations and orientations were fixed and source strengths were allowed to vary over time to provide the best fit for the recorded data (for further details concerning data acquisition and analysis see Hämäläinen et al. 1993). Resulting source strengths as a function of time were used for determination of peak latencies defined as first peak after stimulus delivery of at least 5 nAm.

Based on fiducial point markers, MRI and MEG coordinate systems were aligned, and sources were superposed on the individual MRI scans. Locations of sources were determined in a coordinate system with the x-axis passing through the preauricular points, the y-axis passing through the nasion normal to the x-axis, and the z-axis pointing up normal to the xy-plane.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Laser stimuli applied to the skin evoked moderately painful well-localized "pinprick-like" sensations. Likewise, intravenously applied laser stimuli evoked moderately to severely painful well-localized sharp sensations. In both conditions, quality of sensations were stable both within and between subjects. In the intravenous stimulation condition there was a tendency toward habituation that was counterbalanced by changing the position of the optical fiber after 5-15 stimuli. Thus intensity of evoked sensations was slightly more variable in the intravenous laser stimulation condition than in the cutaneous laser stimulation condition.

Figure 1 compares timing and location of cortical responses to cutaneous and intravenous laser stimuli in a representative subject. Cutaneous stimuli evoked responses similar to previous investigations (Ploner et al. 1999, 2000; Timmermann et al. 2001). Field patterns of neuromagnetic responses to cutaneous laser stimuli at 147 ms after stimulus application indicated simultaneous activation of a contralateral parietal source with anterior-posterior current direction and of sources with inferior-superior current directions in the temporoparietal cortex of both hemispheres (Fig. 1A). Sensors located above these sources detected maximum signals at 147 and 145 ms in the contralateral and at 161 ms in the ipsilateral hemisphere, respectively (Fig. 1A). No earlier responses were detected. Dipole modeling and superposition of dipole location on individual MRI scans showed location of sources in the contralateral postcentral gyrus and bilaterally in the upper banks of the Sylvian fissures, corresponding to contralateral SI and bilateral SII (Fig. 1B). Source strengths as a function of time showed activation peaks at 150 ms (contralateral SI), 146 ms (contralateral SII), and 155 ms (ipsilateral SII), respectively (Fig. 1C).

View larger version (33K): [in this window] [in a new window]   Fig. 1. Exemplary results. Cortical responses to painful cutaneous and venous stimulation of the left hand. A: magnetic field patterns at 147 and 97 ms after cutaneous (left) and venous (right) stimulus application, respectively. The magnetic field patterns are displayed over the helmet-shaped sensor array viewed from the right, top, and the left. Squares show locations of 61 sensor pairs. Isocontour lines are separated by 100 fT. Dashed lines indicate fields directed into the head. Arrows represent location and direction of dipoles. The middle panel shows responses to cutaneous and venous stimuli recorded from the same sensors marked by shaded squares on the left and right. B: location of sources superposed on axial and coronal magnetic resonance imaging (MRI) scans of the subject's brain, slice thickness 15 mm. C: source strengths as a function of time. SI, primary somatosensory cortex; SII, secondary somatosensory cortex; cl, contralateral to stimulus application; il, ipsilateral to stimulus application.

Intravenously applied laser stimuli yielded magnetic field patterns similar to cutaneous stimulation suggesting activation of a contralateral SI-source with anterior-posterior current direction and bilateral SII-sources with inferior-superior current directions (Fig. 1A). However, in the venous condition this field pattern was observed at 97 ms, i.e., 50 ms earlier than in the cutaneous stimulation condition. Correspondingly, sensors located above the cortical sources detected maximum signals at 87, 89, and 99 ms, respectively (Fig. 1A). Source modeling and superposition of source locations on MRI scans confirmed location of sources in the contralateral postcentral gyrus (SI) and in the upper banks of the Sylvian fissures, bilaterally (SII; Fig. 1B). Comparison of source strengths as a function of time between both conditions corroborated an earlier activation of sources to intravenously as compared with cutaneously applied stimuli (Fig. 1C).

Table 1 shows peak latencies of activations in contralateral SI and bilateral SII to both stimuli for all subjects. Two-way repeated measures ANOVA revealed a significant effect of condition (venous/cutaneous; P = 0.003) and of source (SI contralateral/SII contralateral/SII ipsilateral; P = 0.03) on the observed latencies. Latencies to venous stimuli were 55 ± 6 ms (mean ± SE) shorter than to cutaneous stimuli. Scheffé's post hoc test showed that latencies of contralateral SII activations were significantly shorter than latencies of ipsilateral SII activations (P = 0.03). There was no significant interaction between condition and source (P = 0.6). Figure 2 illustrates this latency differences by showing grand averages across source waveforms of contralateral SI and bilateral SII in all subjects. Grand averages of source activations to cutaneous and venous stimuli peaked at 160 and 121 ms (contralateral SI), 151 and 129 ms (contralateral SII), and 167 and 134 ms (ipsilateral SII), respectively.

View larger version (25K): [in this window] [in a new window]   Fig. 2. Group results. Mean source strengths as a function of time across all subjects for painful cutaneous and venous stimulations. Shaded areas depict ±SE. SI, primary somatosensory cortex; SII, secondary somatosensory cortex; cl, contralateral to stimulus application; il, ipsilateral to stimulus application.

Peak amplitudes of source activations to cutaneous and venous stimuli were 28 ± 7 and 14 ± 3 nAm (contralateral SI), 41 ± 7 and 43 ± 10 nAm (contralateral SII), and 47 ± 9 and 29 ± 9 nAm (ipsilateral SII), respectively (mean ± SE). Two-way repeated measures ANOVA revealed a significant effect of condition (venous/cutaneous; P = 0.03) but not of source (P = 0.2) on the observed amplitudes. There was no significant interaction between condition and source (P = 0.1).

Group analysis of source locations did not show a significant effect of condition on source location in x-, y-, or z-direction (repeated measures ANOVA, P > 0.1).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In the present study we compared cortical responses to cutaneous and intravenous painful laser stimuli applied to the dorsum of the hand in healthy human subjects. For the first time our results show the cortical representation of venous nociception in humans. Similar to the here and previously (Ploner et al. 1999, 2000; Timmermann et al. 2001) studied responses to cutaneous stimuli venous stimulation evoked nearly simultaneous activation of contralateral SI and bilateral SII. However, comparison of response latencies between stimulations revealed that venous stimuli activated all three areas about 50 ms earlier than cutaneous stimuli.

These results demonstrate involvement of the somatotopically appropriate region of SI and of SII in processing of nociceptive information from the veins. Locations of SI- and SII-responses did not differ between venous and cutaneous nociceptive stimulations. Previous investigations indicated generation of SI-responses to cutaneous nociceptive stimuli in cytoarchitectonical area 1 of SI (Ploner et al. 2000). Thus venous SI-responses most probably originate from cytoarchitectonical area 1 of SI, too. Locations of SII-responses in the present and previous studies (Ploner et al. 1999, 2000; Timmermann et al. 2001) and inferior-superior source orientations to both stimuli indicate that these responses predominantly originate from SII-cortex in the parietal operculum. However, a contribution of more medially located insular cortex expected to yield mainly radially oriented sources not detected by MEG cannot be ruled out.

Our finding of involvement of SI is in accordance with investigations in the cat showing that stimulation of venous afferents evokes responses in the sensory-motor cortex (Thompson et al. 1980). No further evidence on the cortical representation of vascular nociception has been presented so far, neither in humans nor in animals.

Apart from the general latency difference between cortical responses to cutaneous and venous stimuli, the temporal relationship between activation of sources does not differ between cutaneous and venous stimulation. Thus the organization of nociceptive venous processing appears to share the parallel organization of cutaneous nociceptive processing in human contralateral SI and SII (Ploner et al. 1999, 2000; Timmermann et al. 2001). The significant latency difference between contralateral and ipsilateral SII in both conditions suggests transcallosal transfer of nociceptive information from contralateral to ipsilateral SII. However, there was a significant effect of condition on source amplitudes. Here, differences in stimulus parameters do not allow to distinguish between a physiological and a methodological effect.

Considering the mesodermal origin of the veins, our results might be compared with investigations of the cortical representation of other tissues of mesodermal origin. In a recent functional imaging study, cerebral activations to painful stimulation of the muscle and of the skin were compared (Svensson et al. 1997). In this study, similar cerebral activation patterns including SI and SII were observed in both conditions. This suggests that cortical nociceptive processing from tissues of mesodermal origin, e.g., blood vessels and muscle, is similar to nociceptive processing from tissues of ectodermal origin, e.g., the skin.

Our findings indicate that the afferent volley from the veins reaches the cerebral cortex significantly earlier than that from the skin. It is unlikely that the observed latency difference is due to a difference in stimulation site between venous and cutaneous stimulation. Differences in stimulation site between conditions were kept below 5 cm in distal-proximal direction. Consequently, considering conduction velocities of A-fibers of about 10 m/s (Raja et al. 1999), a systematic difference in conduction distance cannot adequately explain a latency difference of about 50 ms, but at most 5 ms. Likewise, a difference in pain intensity between conditions is unlikely to yield a latency difference of 50 ms since previous evoked potential studies investigating intensity dependence of cortical responses did not show a dependence of latencies on pain intensity (Carmon et al. 1978; Chen et al. 1979; Harkins and Chapman 1978; Spiegel et al. 2000). Nor is the observed latency difference likely to be due to a difference in receptor activation latencies. In comparison to the CO₂-laser with a receptor activation time for A-nociceptive afferents of about 40 ms (Bromm and Treede 1984), physical stimulus properties of the Tm:YAG-laser indicate a considerably shorter activation latency (Spiegel et al. 2000). This is confirmed by shorter latencies of cortical responses to Tm:YAG-laser stimuli than to CO₂-laser stimuli (Spiegel et al. 2000). Thus most probably, the latency difference between cortical responses to cutaneous and venous stimulation is due to a difference in peripheral or central conduction velocity. Based on the present results we cannot decide between these possibilities. However, except for the latency difference our results reveal a cortical representation of venous nociception similar to cutaneous nociception. Thus a difference in conduction velocity between peripheral venous and cutaneous nociceptive afferents appears most likely to account for the latency difference. However, peripheral conduction velocities of venous and cutaneous afferents have not been compared yet. In monkeys two types of cutaneous nociceptive A-fibers with different response characteristics and a difference in conduction velocity of about 10 m/s, which could well explain a 50-ms latency difference from hand to cerebral cortex have been described (Treede et al. 1995, 1998). It is unclear whether the observed latency difference in our study corresponds to these findings. In addition, a higher intravenous than cutaneous temperature might contribute to faster conduction of venous nociceptive afferents.

In conclusion, the present results show that vascular nociception shares the cortical representation of cutaneous nociception in human somatosensory cortices. However, the afferent volley from the veins reaches the cerebral cortex significantly earlier than from the skin. This might be due to differences in peripheral conduction velocity. Our findings suggest that in human somatosensory cortices nociceptive processing from tissues of mesodermal origin, e.g., the blood vessels, is similar to nociceptive processing from tissues of ectodermal origin, e.g., the skin.