85
Pain, 40 (1990) 85-91 Elsevier
PAIN 01528
Basic Section Studies of heat pain sensation in man: perception thresholds, rate of stimulus rise and reaction time David Department
of Neurology, Good Samaritan (Received
Yarnitsky
1 and Jose L. Ochoa
Hospital and Medical Center, and Oregon Health Sciences University, Portland,
7 December
1988, revision received 28 July 1989, accepted
17 August
OR (U.S.A.)
1989)
Afferent impulse frequency, one of the determinants of subjective magnitude of sensation, varies with the rate of rise sunlmaly of stimulus intensity: the faster the increase in stimulus energy, the higher the frequency of firing for a given amount of energy. This predicts that the steeper the stimulus ramp the lower will be the threshold for perception. While such inverse relation holds for myelinated fibre mediated cold sensation and mechanical pressure sensation, the opposite has been reported for unmyelinated fibre mediated heat pain and cold pain sensations. These paradoxical results intuitively suggest possible reaction time artefact. Indeed, a fixed time interval that includes conduction of the impulses to the brain, central processing and efferent conduction, intervenes between sufficient peripheral stimulus and the voluntary signal in reaction to subjective experience. As stimulus temperature continues to rise along this time, an artefactually high threshold reading results: the steeper the temperature rise, the larger will be the artefact, particularly for submodalities with longer reaction time. The present study compared heat pain threshold, obtained through a method that involves reaction time participation, with heat pain thresholds obtained bypassing reaction time. It was found in 16 volunteers that: (a) Heat pain thresholds decreased as the rate of temperature rise increased when reaction time was not a factor (P < 0.001). (b) Heat pain thresholds determined through the method involving reaction time participation were significantly higher than those obtained bypassing reaction time (P < 0.01). Such difference increased with increasing rates of temperature rise. (c) Peripheral conduction velocity calculated from average reaction time was found to be approximately 0.6 m/set. This indicates that at the rates of temperature rise utilized for the present study, heat pain thresholds reflect activity in unmyelinated C nociceptors rather than in small myelinated A6 nociceptors. Key words:
Reaction
time; C nociceptors;
Psychophysics;
Heat pain thresholds;
Introduction Increasing stimulus intensity elevates subjective sensory magnitude through the collaborative contributions of (a) recruitment of larger numbers of afferent units into discharge (spatial summation), and (b) increased impulse frequency in individual
’ Present address: rology, Rambam
David Medical
Yarnitsky, Department Center, Haifa, Israel.
Correspondence to: Dr. D. Yarnitsky, rology, Rambam Medical Center, Haifa, 0304-3959/90/%03.50
Department Israel.
0 1990 Elsevier Science
of Neu-
of Neu-
Publishers
Method
of limits;
Method
of levels
units (temporal summation). In the case of pain induced by heat in man, direct relations have been demonstrated for stimulus intensity, pain magnitude and firing frequency of the responsible C polymodal nociceptors recorded microneurographically [15,31]. Over and above these relations, it has been shown for warm and cold specific skin afferents in man 1211 and monkey [3,10,19,29] that the faster the increase in stimulus energy, the higher the frequency of firing for a given amount of energy. If this observation applies to human C nociceptors, it could be predicted that, for a given thermal energy level, subjective magnitude should be higher
B.V. (Biomedical
Division)
X6
the faster the stimulus reaches that level. Correspondingly, threshold for perception should be lower the steeper the stimulus ramp, as actually demonstrated for the sensations of warmth and cold (at lower rates of temperature rise) [16,20], and pressure [14]. Paradoxically. either no change or an increase in thresholds has been reported in man for heat pain [8,27], for cold pain [8] and for warmth sensation [7,27,30]. Analysis of the methodology used in those reports reveals that. when the method relied on reaction time, threshold increased with steeper stimulus ramps [F&27.30]. The opposite was found when threshold measurement bypassed reaction time [16,20]. No studies seem to be available to compare heat pain thresholds determined with and without reaction time participation in the same individual. Our study compared, at various rates of temperature rise, heat pain thresholds determined by a method of limits that includes reaction time, and a method of levels that excludes it. Reaction time artefact was found to explain a significant difference between the threshold levels obtained by the two methods in the same individuals. When bypassing reaction time, thresholds were unquestionably lower for steeper stimulus ramps. By taking advantage of the difference in results between the methods, peripheral conduction velocity for the neural message induced by a heat pain stimulus was calculated. Under the conditions of the present study the calculated velocities reflected utilization of unmyelinated afferents. Since functional evaluation of small caliber sensory afferents draws progressive clinical attention and since heat pain is one of the critical submodalities served by a subgroup of such fibers [l&25,26,33], it seems timely to incorporate in pertinent studies the methodological considerations discussed here.
Materials and methods Experiments were performed on 16 healthy volunteers, 10 females and 6 males, aged 18-42 years (average 27.8). Controlled, focal, thermal stimulation of the skin was delivered through a rectangular 12.5 cm2 contact Peltier-type ther-
mode, manufactured by Somedic AB (Stockholm, Sweden), as described elsewhere [13]. The thermode included a small thermocouple for instantaneous temperature measurement, connected to a digital thermometer (sensitivity of O.l”C). An Omniscribe D-5000 strip chart recorder (Houston Instruments, Austin, TX) graphically registered the measured temperature along time. The stimulating probe was attached to hairy skin of the forearm, on the dorsal aspect, with its distal edge 5 cm proximal to the wrist joint. Distance to the spinous process of C7 was 63.3 cm on the average. The probe was located on the right side in 7 subjects and on the left in 9. Baseline temperature of the skin ranged from 28 to 32.8” C, and adapting temperature was kept at 35 a C. Each test session started with a training series of 6 presentations of heat stimuli (three for each method. as described below), the results of which were discarded. The purpose of this exercise was to familiarize the subjects with the methodology. The actual experimental session included 12 series of stimulations, each consisting of combinations of 2 different methods and 3 different rates of stimulus rise, in random order. Intervals of at least 30 set were observed between successive presentations, in order to avoid possible sensitization of cutaneous receptors. Two methods of threshold determination were applied: (a) Method of limits. The subjects were asked to signal by pressing a switch when first feeling pain, threshold being determined from the average reading of 4 consecutive presentations. (b) Method of levels. The temperature of the probe was increased to a certain target level and then decreased automatically to baseline. The subjects were retrospectively asked whether or not they felt pain after each presentation. The examiner then adjusted the maximal target temperature in search for a threshold, as follows. Starting at a subthreshold level, target temperature was elevated in steps of 0.8’ C until threshold was exceeded and the subject first reported pain. The subsequent target temperature presented was 0.4”C lower. If this was still suprathreshold, a further final down step of 0.2” C was taken. Alternatively, a final step up of 0.2”C was given. Temperature midway between the highest negative response and the lowest
87
positive response, which were always 0.2’ C apart, was taken as the threshold. On the average, 4.7 (range 3-9) stimulations were needed to define a threshold. A similar automated method is applied in the determination of warm and cold sensory thresholds through the Glasgow thermal system [17], based on the Mayo Clinic method [ll]. Three different rates of temperature increase were used, 0.5, 1.2 and 2’ C/set. Despite constant current input to the thermal probe, a slight interindividual difference in rate of actual temperature rise was found. This is thought to derive from uncontrolable variables such as the pressure at which the probe is attached to the skin, the heat capacity of the adjacent tissues, the efficiency of the local circulation in dissipating heat, etc. Intraindividual variability was minimized by keeping the probe location and pressure of application unchanged throughout the test session.
Results Thresholds for heat pain measured through the 2 different methods and each of the 3 rates are summarized in Table I. A clear inverse relationship prevailed between heat pain threshold and rate of temperature rise when measured by the method of levels. Average threshold fell from 42.17”C at O.S”C/sec to 41.63”C at 1.2”C/sec and 41.00 o C at the highest rate used of 2” C/set. In contrast, the thresholds obtained by the method of limits, which includes the reaction time artefact,
TABLE
I
HEAT PAIN “C+S..D.,N=16)
THRESHOLDS
Method
Method Method
a b c d
were not affected by similarly changing the rate of stimulation (43.00, 42.76 and 42.99OC for the lower, medium and higher rates, respectively). The thresholds obtained by the method which includes reaction time were significantly higher than those measured through the method of levels. The faster the rate of temperature increase, the greater was the observed difference between thresholds (Table II). This happened at the expense of progressive decrease in the method of levels thresholds (Table I). The duration of the reaction time interval was calculated by dividing the temperature difference by the rate of temperature rise. The average reaction time in this study was 1.20 sec. Three- by 2-way repeated measures ANOVA was used to test significance of the results. The overall sign interaction effect was F (2, 30) = 9.06, P = 0.0008. For post hoc analysis the Bonferroni t test was applied when testing for the effect of rate of temperature change on heat pain thresholds, and Tukey’s test when testing the effect of the method of measurement on the thresholds. Peripheral conduction velocity of the neural message conveying the sensation of heat pain at near threshold intensities was calculated on the basis of reaction time. Average peripheral conduction distance of 0.63 m was measured between stimulus site and spinal process of C,. Time for central and efferent components of the overall reaction time was arbitrated as 200 msec based on the generally accepted 130-190 msec for central processing [4-6,281 and estimated efferent time of
MEASURED
BY 2 METHODS
Rate (O C/set)
AT
3 RATES
OF
TEMPERATURE
RISE
(RESULTS
0.5 vs. 1.2
1.2 vs. 2
0.5 vs. 2
AT(“C)
0.5
1.2
2
of levels
42.17 +
41.63 f
41.00 f
2.48 42.76 f
2.48 42.99 +
0.54 =
0.63 b
1.17 c
of limits
2.36 43.00 + 1.74
1.96
2.02
0.24 d
0.23 d
0.01 d
P < 0.02. P
Pain, 40 (1990) 85-91 Elsevier
PAIN 01528
Basic Section Studies of heat pain sensation in man: perception thresholds, rate of stimulus rise and reaction time David Department
of Neurology, Good Samaritan (Received
Yarnitsky
1 and Jose L. Ochoa
Hospital and Medical Center, and Oregon Health Sciences University, Portland,
7 December
1988, revision received 28 July 1989, accepted
17 August
OR (U.S.A.)
1989)
Afferent impulse frequency, one of the determinants of subjective magnitude of sensation, varies with the rate of rise sunlmaly of stimulus intensity: the faster the increase in stimulus energy, the higher the frequency of firing for a given amount of energy. This predicts that the steeper the stimulus ramp the lower will be the threshold for perception. While such inverse relation holds for myelinated fibre mediated cold sensation and mechanical pressure sensation, the opposite has been reported for unmyelinated fibre mediated heat pain and cold pain sensations. These paradoxical results intuitively suggest possible reaction time artefact. Indeed, a fixed time interval that includes conduction of the impulses to the brain, central processing and efferent conduction, intervenes between sufficient peripheral stimulus and the voluntary signal in reaction to subjective experience. As stimulus temperature continues to rise along this time, an artefactually high threshold reading results: the steeper the temperature rise, the larger will be the artefact, particularly for submodalities with longer reaction time. The present study compared heat pain threshold, obtained through a method that involves reaction time participation, with heat pain thresholds obtained bypassing reaction time. It was found in 16 volunteers that: (a) Heat pain thresholds decreased as the rate of temperature rise increased when reaction time was not a factor (P < 0.001). (b) Heat pain thresholds determined through the method involving reaction time participation were significantly higher than those obtained bypassing reaction time (P < 0.01). Such difference increased with increasing rates of temperature rise. (c) Peripheral conduction velocity calculated from average reaction time was found to be approximately 0.6 m/set. This indicates that at the rates of temperature rise utilized for the present study, heat pain thresholds reflect activity in unmyelinated C nociceptors rather than in small myelinated A6 nociceptors. Key words:
Reaction
time; C nociceptors;
Psychophysics;
Heat pain thresholds;
Introduction Increasing stimulus intensity elevates subjective sensory magnitude through the collaborative contributions of (a) recruitment of larger numbers of afferent units into discharge (spatial summation), and (b) increased impulse frequency in individual
’ Present address: rology, Rambam
David Medical
Yarnitsky, Department Center, Haifa, Israel.
Correspondence to: Dr. D. Yarnitsky, rology, Rambam Medical Center, Haifa, 0304-3959/90/%03.50
Department Israel.
0 1990 Elsevier Science
of Neu-
of Neu-
Publishers
Method
of limits;
Method
of levels
units (temporal summation). In the case of pain induced by heat in man, direct relations have been demonstrated for stimulus intensity, pain magnitude and firing frequency of the responsible C polymodal nociceptors recorded microneurographically [15,31]. Over and above these relations, it has been shown for warm and cold specific skin afferents in man 1211 and monkey [3,10,19,29] that the faster the increase in stimulus energy, the higher the frequency of firing for a given amount of energy. If this observation applies to human C nociceptors, it could be predicted that, for a given thermal energy level, subjective magnitude should be higher
B.V. (Biomedical
Division)
X6
the faster the stimulus reaches that level. Correspondingly, threshold for perception should be lower the steeper the stimulus ramp, as actually demonstrated for the sensations of warmth and cold (at lower rates of temperature rise) [16,20], and pressure [14]. Paradoxically. either no change or an increase in thresholds has been reported in man for heat pain [8,27], for cold pain [8] and for warmth sensation [7,27,30]. Analysis of the methodology used in those reports reveals that. when the method relied on reaction time, threshold increased with steeper stimulus ramps [F&27.30]. The opposite was found when threshold measurement bypassed reaction time [16,20]. No studies seem to be available to compare heat pain thresholds determined with and without reaction time participation in the same individual. Our study compared, at various rates of temperature rise, heat pain thresholds determined by a method of limits that includes reaction time, and a method of levels that excludes it. Reaction time artefact was found to explain a significant difference between the threshold levels obtained by the two methods in the same individuals. When bypassing reaction time, thresholds were unquestionably lower for steeper stimulus ramps. By taking advantage of the difference in results between the methods, peripheral conduction velocity for the neural message induced by a heat pain stimulus was calculated. Under the conditions of the present study the calculated velocities reflected utilization of unmyelinated afferents. Since functional evaluation of small caliber sensory afferents draws progressive clinical attention and since heat pain is one of the critical submodalities served by a subgroup of such fibers [l&25,26,33], it seems timely to incorporate in pertinent studies the methodological considerations discussed here.
Materials and methods Experiments were performed on 16 healthy volunteers, 10 females and 6 males, aged 18-42 years (average 27.8). Controlled, focal, thermal stimulation of the skin was delivered through a rectangular 12.5 cm2 contact Peltier-type ther-
mode, manufactured by Somedic AB (Stockholm, Sweden), as described elsewhere [13]. The thermode included a small thermocouple for instantaneous temperature measurement, connected to a digital thermometer (sensitivity of O.l”C). An Omniscribe D-5000 strip chart recorder (Houston Instruments, Austin, TX) graphically registered the measured temperature along time. The stimulating probe was attached to hairy skin of the forearm, on the dorsal aspect, with its distal edge 5 cm proximal to the wrist joint. Distance to the spinous process of C7 was 63.3 cm on the average. The probe was located on the right side in 7 subjects and on the left in 9. Baseline temperature of the skin ranged from 28 to 32.8” C, and adapting temperature was kept at 35 a C. Each test session started with a training series of 6 presentations of heat stimuli (three for each method. as described below), the results of which were discarded. The purpose of this exercise was to familiarize the subjects with the methodology. The actual experimental session included 12 series of stimulations, each consisting of combinations of 2 different methods and 3 different rates of stimulus rise, in random order. Intervals of at least 30 set were observed between successive presentations, in order to avoid possible sensitization of cutaneous receptors. Two methods of threshold determination were applied: (a) Method of limits. The subjects were asked to signal by pressing a switch when first feeling pain, threshold being determined from the average reading of 4 consecutive presentations. (b) Method of levels. The temperature of the probe was increased to a certain target level and then decreased automatically to baseline. The subjects were retrospectively asked whether or not they felt pain after each presentation. The examiner then adjusted the maximal target temperature in search for a threshold, as follows. Starting at a subthreshold level, target temperature was elevated in steps of 0.8’ C until threshold was exceeded and the subject first reported pain. The subsequent target temperature presented was 0.4”C lower. If this was still suprathreshold, a further final down step of 0.2” C was taken. Alternatively, a final step up of 0.2”C was given. Temperature midway between the highest negative response and the lowest
87
positive response, which were always 0.2’ C apart, was taken as the threshold. On the average, 4.7 (range 3-9) stimulations were needed to define a threshold. A similar automated method is applied in the determination of warm and cold sensory thresholds through the Glasgow thermal system [17], based on the Mayo Clinic method [ll]. Three different rates of temperature increase were used, 0.5, 1.2 and 2’ C/set. Despite constant current input to the thermal probe, a slight interindividual difference in rate of actual temperature rise was found. This is thought to derive from uncontrolable variables such as the pressure at which the probe is attached to the skin, the heat capacity of the adjacent tissues, the efficiency of the local circulation in dissipating heat, etc. Intraindividual variability was minimized by keeping the probe location and pressure of application unchanged throughout the test session.
Results Thresholds for heat pain measured through the 2 different methods and each of the 3 rates are summarized in Table I. A clear inverse relationship prevailed between heat pain threshold and rate of temperature rise when measured by the method of levels. Average threshold fell from 42.17”C at O.S”C/sec to 41.63”C at 1.2”C/sec and 41.00 o C at the highest rate used of 2” C/set. In contrast, the thresholds obtained by the method of limits, which includes the reaction time artefact,
TABLE
I
HEAT PAIN “C+S..D.,N=16)
THRESHOLDS
Method
Method Method
a b c d
were not affected by similarly changing the rate of stimulation (43.00, 42.76 and 42.99OC for the lower, medium and higher rates, respectively). The thresholds obtained by the method which includes reaction time were significantly higher than those measured through the method of levels. The faster the rate of temperature increase, the greater was the observed difference between thresholds (Table II). This happened at the expense of progressive decrease in the method of levels thresholds (Table I). The duration of the reaction time interval was calculated by dividing the temperature difference by the rate of temperature rise. The average reaction time in this study was 1.20 sec. Three- by 2-way repeated measures ANOVA was used to test significance of the results. The overall sign interaction effect was F (2, 30) = 9.06, P = 0.0008. For post hoc analysis the Bonferroni t test was applied when testing for the effect of rate of temperature change on heat pain thresholds, and Tukey’s test when testing the effect of the method of measurement on the thresholds. Peripheral conduction velocity of the neural message conveying the sensation of heat pain at near threshold intensities was calculated on the basis of reaction time. Average peripheral conduction distance of 0.63 m was measured between stimulus site and spinal process of C,. Time for central and efferent components of the overall reaction time was arbitrated as 200 msec based on the generally accepted 130-190 msec for central processing [4-6,281 and estimated efferent time of
MEASURED
BY 2 METHODS
Rate (O C/set)
AT
3 RATES
OF
TEMPERATURE
RISE
(RESULTS
0.5 vs. 1.2
1.2 vs. 2
0.5 vs. 2
AT(“C)
0.5
1.2
2
of levels
42.17 +
41.63 f
41.00 f
2.48 42.76 f
2.48 42.99 +
0.54 =
0.63 b
1.17 c
of limits
2.36 43.00 + 1.74
1.96
2.02
0.24 d
0.23 d
0.01 d
P < 0.02. P
