Thermoreceptive innervation of human glabrous and hairy
skin: a contact heat evoked potential analysis
Yelena Granovskya,b,c, Dagfinn Matrea,b,d, Alexander Sokolika,b,
Ju ¨rgen Lorenza,b,e, Kenneth L. Caseya,b,*
aDepartment of Neurology, University of Michigan, 2215 Fuller Road, Ann Arbor, MI 48105, USA
bNeurology Research Laboratories, Veterans Affairs Medical Center, University of Michigian, 2215 Fuller Rd., Ann Arbor, MI 48105, USA
cDepartment of Neurology, Rambam Medical Center and Technion Faculty of Medicine, Haifa, Israel
dDepartment of Physiology, National Institute of Occupational Health, N-0033 Oslo, Norway
eCenter for Experimental Medicine, Institute of Neurophysiology and Pathophysiology, University Clinic Hamburg Eppendorf Martinistr. 52,
D-20246 Hamurg, Germany
Received 2 November 2004; received in revised form 7 January 2005; accepted 14 February 2005
The human palm has a lower heat detection threshold and a higher heat pain threshold than hairy skin. Neurophysiological studies of
monkeys suggest that glabrous skin has fewer low threshold heat nociceptors (AMH type 2) than hairy skin. Accordingly, we used a
temperature-controlled contact heat evoked potential (CHEP) stimulator to excite selectively heat receptors with C fibers or Ad-innervated
AMH type 2 receptors in humans. On the dorsal hand, 51 8C stimulation produced painful pinprick sensations and 41 8C stimuli evoked
warmth. On the glabrous thenar, 41 8C stimulation produced mild warmth and 51 8C evoked strong but painless heat sensations. We used
CHEP responses to estimate the conduction velocities (CV) of peripheral fibers mediating these sensations. On hairy skin, 41 8C stimuli
evoked an ultra-late potential (mean, SD; N wave latency: 455 (118) ms) mediated by C fibers (CV by regression analysis: 1.28 m/s, NZ15)
whereas 51 8C stimuli evoked a late potential (N latency: 267(33) ms) mediated by Ad afferents (CV by within-subject analysis: 12.9 m/s,
NZ6). In contrast, thenar responses to 41 and 51 8C were mediated by C fibers (average N wave latencies 485(100) and 433(73) ms,
respectively; CVs 0.95–1.35 m/s by regression analysis, NZ15; average CVZ1.7 (0.41) m/s calculated from distal glabrous and proximal
hairy skin stimulation, NZ6). The exploratory range of the human and monkey palm is enhanced by the abundance of low threshold,
C-innervated heat receptors and the paucity of low threshold AMH type 2 heat nociceptors.
q 2005 Published by Elsevier B.V. on behalf of International Association for the Study of Pain.
Keywords: Somatic pain; Heat pain; Human; Skin; Warm; Evoked potentials; Thermal sensation
Psychophysical studies have shown that the glabrous
skin of the human hand has a lower heat detection threshold
than any cutaneous surface other than the lips (Stevens and
Choo, 1998). Yet, glabrous hand skin has a much higher
heat pain threshold than hairy skin (Taylor et al., 1993).
Neurophysiological studies of the thermal innervation of
monkey skin have provided a possible explanation for these
differences by showing that glabrous skin, while richly
innervated by heat-sensitive receptors with unmyelinated C
fibers and high threshold nociceptors with finely myelinated
Ad fibers (AMH type I), is innervated sparsely, if at all, by
lower threshold heat nociceptors (AMH type II) (Treede
et al., 1995). In experiments relevant to this differential
innervation hypothesis, Towell and colleagues (1996)
examined the cerebral vertex potentials evoked by infrared
laser stimulation of human hairy and glabrous (palm) skin.
Pain xx (2005) 1–10
0304-3959/$20.00 q 2005 Published by Elsevier B.V. on behalf of International Association for the Study of Pain.
*Corresponding author. Address: Neurology Research Laboratories,
Veterans Affairs Medical Center, University of Michigian, 2215 Fuller Rd.,
Ann Arbor, MI 48105, USA. Tel.: C734 769 7100x5870; fax: C734 769
E-mail address: email@example.com (K.L. Casey).
ARTICLE IN PRESS
These investigators evoked potentials mediated by both Ad
and C fibers from both glabrous and hairy skin. They could
not, however, determine the type of Ad fiber receptor that
was excited by the laser stimulation. Based on reaction time
experiments, they suggested that either Ab or AMH type 1
fibers may have responded to laser stimulation of the palm
(Towell et al., 1996).
In this paper, we present evidence that Ad and C fibers
innervating the human glabrous and hairy skin have
different intensity dependencies and relative innervation
densities. We used a newly developed contact heat evoked
potential (CHEP) stimulator (Medoc Ltd, Ramat Yishai,
Israel) to deliver rapidly-ramped contact heat stimuli that
evoke late and ultra-late cerebral potentials reliably. CHEP
stimulators have been developed and used by others to
demonstrate the feasibility of this method in human
neurophysiological studies (Arendt-Nielsen and Chen,
2003; Chen et al., 2001). An advantage of the device we
used is that it delivers heat pulses with peak temperatures
that are adjustable, monitored on-line and controlled by
rapid feed-back, thus taking advantage of the differential
heat thresholds of receptors innervated by C and Ad fibers
(Treede et al., 1995). To estimate the CVs of the fibers
mediating these potentials, we measured evoked potential
latencies among subjects with different arm lengths and
within subjects following stimulation of proximal and distal
arm locations; we also recorded the potentials evoked within
subjects following glabrous and hairy skin stimulation of the
hand and foot. Because glabrous skin occurs only at distal
sites of extremities, we recruited a subject population with a
wide range of arm lengths.
Sixteen paid healthy volunteers (8 males and 8 females, ages
18–35 years) participated in this study The data from one female
subject was excluded due to a history of ongoing cervical pain. The
local institutional review board of the Ann Arbor Veteran’s Affairs
Medical Center approved the study protocol. Each subject signed a
consent form after receiving a complete explanation about the
purpose and design of the study.
We used a contact heat evoked potential stimulator (CHEPs,
Medoc Ltd, Ramat Yishai, Israel), with a thermode that contacts a
cutaneous area of 572.5 mm2. The thermode is comprised of a
heating thermofoil (Minco Products, Inc.,Minneapolis, MN) that is
covered with a 25 mm layer of thermoconductive plastic (Kaptonw,
thermal conductivity at 23 8C of 0.1–0.35 W/m/K). Two thermo-
couples are embedded 10 mm within this conductive coating,
which contacts the skin directly, thus providing an estimate of the
skin temperature at the thermode surface. We also confirmed
the baseline temperature of the thermode surface and of the
thermode-skin interface by measurements with an independent
digital thermometer. Nonetheless, this estimate of skin temperature
is necessarily an approximation because we did not insert
intracutaneous thermocouples or thermistors in our subjects.
Accordingly, the stimulus temperatures given throughout this
report refer to the temperature of the thermofoil as applied to the
skin surface. The thermofoil permitted a heating rate of up to
70 8C/s and the Peltier device allowed a cooling rate of 40 8C/s.
Cooling began immediately following attainment of the target heat
pulse temperature, which was set by the investigator using software
provided by the manufacturer. The thermofoil-skin temperature is
measured and software-controlled 150 times per second or just
over 52 times during a 350 ms heat pulse (FWHM).
Subjects sat in an armchair in a quiet room with an ambient
temperature near 22 8C. We applied heat stimuli at two peak
intensities, one potentially noxious (51 8C) and one innocuous, to
five body sites: the right thenar eminence, the dorsum of the right
hand, the proximal volar forearm, and the dorsum and sole of the
foot. The innocuous temperature was 41 8C for all sites except the
sole of the foot where the thick layer of the skin required a higher
intensity stimulus (46–48 8C) to obtain sensations similar to those
from the hand. The baseline temperature was 35 8C for all stimuli.
The average time from onset to peak temperature was 190G24 ms
for 41 8C and 250G8 ms for 51 8C.
Each stimulus block consisted of 20 constant-intensity stimuli
applied to the same site at pseudorandom inter-stimulus intervals
of approximately 15 s. The thermode remained at the same site
during each block. We stimulated body sites in a pseudorandom
order. To avoid sensitization and desensitization, low intensity
stimuli preceded high intensity stimuli at the beginning of each
block. To avoid expectation effects and to reduce the novelty effect
on heat evoked potentials we applied several stimuli before
beginning evoked potential recording. The subjects were given a
3–5 min break after each stimulation block.
The subjects rated the perception of each stimulus 3 s after
stimulus onset. The ratings were based on a 0–10 level numerical
ranking scale. The extremes of the scale were ‘no sensation’ at 0
and ‘unbearable burning sensation’ at 10. A level of 4 was the
threshold for a pinprick-like pain sensation.
2.4. Contact heat evoked potential recording
Contact heat evoked potentials (CHEPs) were recorded from
six midline electrodes (Fz, FCz, Cz, CPz, Pz, POz) using an
electrode cap that contains 59 electrode positions according to the
10%-system, an extended montage of the standard 10–20 system.
Here, we report only data recorded from the Cz (vertex) position.
Linked earlobe electrodes served as reference. The electroenceph-
alogram (EEG) was recorded within a 0.15 and 100 Hz bandpass
and digitized at a sampling rate of 500 Hz. For artifact control, we
monitored the electrooculogram (EOG) from supra- and infra-
orbital electrodes. The impedance from all electrodes was below
5 kU. EEG data were stored on disk and analyzed off-line (Scan
4.2, Compumedics USA, El Paso, TX, USA). Each recording
epoch of 2600 ms included a period of 200 ms baseline before
stimulus onset, which was initiated by a TTL pulse at the beginning
of the temperature increase.
Y. Granovsky et al. / Pain xx (2005) 1–102
ARTICLE IN PRESS
2.5. Data analysis
Peri-stimulus epochs contaminated by ocular artifacts were
discarded and not included in signal averaging. We averaged the
remaining sweeps separately for stimulus intensity and stimulus
site. Within each individual’s average waveform, we identified the
peak latency and amplitude of the major negative and subsequent
positive peaks of the evoked potentials.
To estimate the conduction velocity (CV) of peripheral nerve
fibers mediating evoked potentials, we used a tape measure on the
fully abducted arm to estimate the distance from the middle of
hand to the C7 vertebral spinous process (straight arm abducted 908
from the trunk); we also measured the distance between the sites of
thenar and proximal forearm stimulation in each subject. Because
glabrous skin is located distally only, we used a regression analysis
to estimate the conduction velocity of peripheral nerve fibers from
latency differences in the potentials evoked from all subjects. We
plotted the latencies of negative and positive peaks in milliseconds
analysis. The regression coefficients define the slope of the
regression curve and its intercept at the ordinate; the reciprocal of
the slope parameter is an estimate of the CV in meters per second.
This method of analysis obviates the need for applying corrections
for differences between nerve and arm length over long distances
analysis was applied to data obtained from a subpopulation of our
subjects (NZ11) from whom both late and ultra-late potentials
(hairy versus glabrous skin) and intensity (low versus high) on
intensity ratings, and on the latency and amplitude of evoked
potential components. Sources of any significant main and
interaction effects were explored using two-tailed paired t-tests.
In a subpopulation of 6 subjects, we were able to evoke
potentials following 41 8C stimulation of the proximal forearm and
following 51 8C stimulation of the dorsal hand and proximal
forearm. In these subjects, we performed conduction velocity
estimates directly from the intra-individual differences in latency
and conduction distance.
Because of the variance introduced by individual differences in
peripheral CV and central delays,we wished toconfirm the validity
of the inter-subject CV estimates and the intra-subject estimates
obtained from the small sub-samples of subjects. Accordingly, we
measured the difference between the latency of the late potential
evoked by noxious stimulation of hairy skin and the ultra-late
potential evoked by high or low intensity stimulation of glabrous
skin in each subject. Assuming negligible differences in intra-
individual fiber conduction velocities and central conduction
delays, this analysis allows an estimate of the changes in the
latency difference that would be expected from conduction
velocity differences between peripheral fiber populations. Thus,
if DZthe increased latency difference (ms) between the earliest
(late) and latest (ultra-late) evoked potentials in two individuals
with different arm lengths, m1Om2, then
where CVuZconduction velocity (m/s) of fibers mediating the
latest (ultra-late) potential and CVlZconduction velocity of fibers
mediating the earlier (late) potential. Thus, for fixed arm length
differences, the latency difference is determined primarily by the
relative conduction velocities of the peripheral fibers. Since the
latency difference is positively related to arm length, the latency
difference observed across the range of arm lengths should be
within the range of values computed in the above equation from
CV estimates obtained within and across individuals. We will
present the comparison between hand and foot data only
descriptively, because it involves an additional difference in spinal
3.1. Intensity ratings
Stimulation of glabrous skin at the thenar produced
sensations of mild and strong painless warmth at peak
temperatures of 41 and 51 8C, respectively. In contrast,
51 8C stimulation of the hairy skin of the dorsum of the hand
always produced a moderately painful pinprick-like sen-
sation and a warm sensation only following stimuli at 41 8C.
The mean intensity ratings for the two peak temperatures at
either skin site are represented in Fig. 1. Two-way repeated
measures ANOVA with site (glabrous and hairy) and
intensity (41 and 51 8C) as variables yielded significant
intensity (F1.14Z125.7; P!0.001) and site (F1.14Z44.7;
P!0.001) main effects and a significant intensity-by-site
interaction (F1.14Z20.2; PZ0.001). This result is due to the
greater intensity rating following 51 than 41 8C (tZ11.2;
P!0.001), a greater intensity rating at hairy than glabrous
skin (tZ6.7; P!0.001), and a greater intensity difference at
hairy than at glabrous skin (tZ4.5; PZ0.001).
3.2. Contact heat evoked potentials (CHEP)
Thenar stimulation at peak pulse temperatures of 41 and
51 8C evoked well-defined potentials in all 15 participants.
Fig. 1. Perceived intensity ratings to heat stimuli applied on the thenar
(open bars) and on the dorsal hand (filled bars). Pain was evoked only by
stimulation of the dorsal hand at 51 8C.
Y. Granovsky et al. / Pain xx (2005) 1–103
ARTICLE IN PRESS
Fig. 2a shows the waveforms evoked in five representative
subjects following stimulation of the glabrous skin at 41 8C.
Subjects are ordered from top to bottom row according to
increasing arm lengths indicated at the left of the tracings.
Broad negative-to-positive ultra-late components are dis-
cernible between 400 and 900 ms poststimulus; these
responses shift to longer latencies with increasing arm
length. Fig. 2b shows the waveforms evoked in the same
subjects following the 51 8C stimulus applied to the hairy
skin of the hand dorsum. These shorter duration negative-to-
positive late components appear between 250 and 500 ms
and vary little with arm length.
Table 1 shows the average latency and amplitude values
of both negative and positive components following 41 and
51 8C stimuli applied to glabrous and hairy skin sites of the
hand. Repeated-measures ANOVA of the peak-to-peak
amplitudes revealed a significant site-by-intensity inter-
action (F1.10Z11.1; PZ0.008) and no main effects of either
site (F1.10Z1.9; PZ0.2) or intensity (F1.10Z0.5; PZ0.5).
This result was due to the higher effects of intensity when
stimulating hairy skin (tZ2.9; PZ0.016). The latency of
the negative wave showed significant main effects for both
site (F1.10Z28.9; P!0.001) and intensity (F1.10Z22.4;
P!0.001) and significant site-by intensity interactions
(F1.10Z38.1; P!0.001). This result is because high
intensity stimuli caused shorter evoked potential latencies
than low intensity stimuli (tZ4.73; PZ0.001), an effect
which was greater on hairy skin (tZ6.17; P!0.001). The
generally shorter latencies following hairy skin stimulation
(tZ5.37, P!0.001) explain the significant main effect of
site. Notably, entering arm length as a covariate in the
ANOVA eliminated all site and intensity main and
interaction effects, and revealed the expected significant
main effect of arm length on latency (F1.9Z11.1; PZ0.009)
that also showed an interaction with site (F1.9Z5.6;
PZ0.04). This result suggests that the variance of the
ultra-late N-wave latency is largely due to a slower
peripheral conduction time following low compared to
high intensity stimulation of glabrous compared to hairy
skin. The P-wave latency was more variable than the
N-wave latency, but yielded a similar result.
Thus far summarized, the results indicate higher pain
sensitivity of hairy compared to glabrous skin; warm
sensitivity showed little difference between these sites.
Consistent with the subjective ratings, CHEPs have greater
amplitudes and shorter latencies when evoked from hairy
Fig. 2. Contact heat cerebral evoked potentials in 5 subjects. Arm length
(cm) is shown beside each record. (A) The ultra-late potential evoked by
41 8C stimulation of thenar. (B) The late potential in the same subjects
evoked by 51 8C stimulation of the dorsal hand.
Average (SD) peak latencies and amplitudes of contact heat potentials evoked by stimulationof glabrous and hairy skin of the hand (nZ15; average arm length
737.3 (70.6) mm)
Stimulus and siteN latency (ms)P latency (ms)N amplitude (mV)P amplitude (mV) Ratings
Thenar, 41 8C
Thenar, 51 8C
Dors. Hand, 41 8C
Dors. Hand, 51 8C
N, negative; P, positive. Ratings are given on a 0–10 cm scale.
Y. Granovsky et al. / Pain xx (2005) 1–104
ARTICLE IN PRESS
compared to glabrous skin, an effect which was greater for
high intensity stimuli. Next, we address the hypothesis that
the latency difference is due to activities originating from
different peripheral fiber spectra.
3.3. Conduction velocity estimates
Figure 3 shows the data points and linear regression
curves with upper and lower confidence levels (95%) for
latencies of the first major negative wave following low and
high intensity stimuli applied to the glabrous (top) and hairy
(bottom) skin sites of the hand plotted as a function of arm
length. Low intensity stimulation of hairy skin failed to
evoke any responses in 4 subjects; therefore, we show data
from only 11 subjects for this condition. CVs derived from
the slope of the regression curves yield 0.95 and 1.35 m/s
for glabrous skin following 41 and 51 8C stimuli, respect-
ively, and 1.28 m/s for hairy skin following 41 8C stimuli. In
contrast, 51 8C stimulation of hairy skin evokes potentials
with significantly shorter latencies and a flatter regression
curve. The CV calculated from the slope of the regression
following high intensity stimulation of hairy skin is 3.91 m/
s. However, these estimates of CV are compromised by
several variables, including arm length measurements and
possible inter-subject differences in the CV of myelinated
and unmyelinated fibers. The greatest error is in estimating
the CV of fibers mediating the shorter latency responses
because the effect of small measurement errors is magnified.
Moreover, the statistical significance level for regressions
obtained from the hairy skin is smaller for both high
intensity (r2Z0.29; PZ0.03) and low intensity stimulation
(r2Z0.23; PZ0.13) than those obtained for both intensities
from the glabrous skin (41 8C: r2Z0.61; PZ0.001; 51 8C:
r2Z0.52; PZ0.002). Accordingly, we obtained additional
estimates of the CV of fibers mediating both the late and
ultra-late potentials evoked from hairy skin within a
subgroup of our subjects as presented below.
Most high intensity stimuli applied to hairy skin evoked
only late, but no ultra-late responses; low intensity stimuli
often failed to evoke any responses. However, we were able
to obtain ultra-late responses after stimulating the proximal
volar forearm with 41 8C in 6 subjects, which allowed a
direct within-subject comparison with ultra-late responses
following a 41 8C stimulus to the thenar. Because the 51 8C
stimuli evoked late components following stimulation of
both the dorsal hand and the proximal forearm, we could
estimate and compare directly the CV mediating late and
ultra-late responses in these individuals. The results are
presented in Table 2A and B and sample responses from
three of these subjects are shown in Fig. 4. The average
conduction velocity is 1.7G0.4 m/s for fibers mediating the
ultra-late potential in these 6 subjects and 12.9G7.5 m/s for
fibers mediating the late potential. Note that the estimated
average CV derived from the ultra-late potential applies
only to the applied stimulation of 41 8C and is based on
stimuli applied to different skin types (glabrous and hairy).
Given the relatively short distance used for all of these
estimates, they are not affected significantly when corrected
for estimations of nerve length in relation to arm length
(Kakigi and Shibasaki, 1991).
To confirm further the CV estimates obtained both within
and across subjects, we computed the latency difference
between the late and ultra-late negative and positive
potentials evoked from hairy and glabrous skin for each
individual. This within-subject latency difference will
Fig. 3. Peak latencies of the late and ultra-late heat potentials evoked from
the glabrous (top) and hairy (bottom) skin of the hand from 15 subjects with
arms of different length. Peak latencies were measured from the negative
phase of each heat potential. There is a significant regression of ultra-late
potential latency on arm length following stimulation of glabrous skin at
41 8C (open triangles; r2Z0.61; PZ0.001) and at 51 8C (filled squares;
r2Z0.52; PZ0.002). For hairy skin, the regression is marginally significant
for 51 8C stimulation (late potential; r2Z0.29; PZ0.03), and shows a trend
only following 41 8C stimulation (r2Z0.23; PZ0.13), which failed to
evoke an ultra-late potential in 4 subjects.
Y. Granovsky et al. / Pain xx (2005) 1–105
ARTICLE IN PRESS
increase with arm length according to the equation
presented in Section 2. This measurement reduces the
variance introduced by inter-subject differences in periph-
eral conduction velocities and central conduction times. As
shown in Fig. 5, negative (filled squares) and positive (open
triangles) ultra-late components following all glabrous heat
stimuli had increasing latency differences relative to the late
components following heat stimulation as arm length
increased across subjects. The intercept of the regression
line predicts a latency difference of approximately 100 ms
for an arm length of 600 mm, which is close to the value
obtained from the shortest subject in our sample (130 ms
over 640 mm). The direct estimate of the average CV of
fibers mediating the late and ultra-late potentials obtained
from the subgroup of six subjects (12.9 and 1.7 m/s,
respectively) predicts an increase in the latency difference
between these potentials of 107 ms over the maximum arm
length difference of 210 mm (850–640 mm) (see equation in
Section 2). The estimates obtained from the regression
analysis of latencies on arm length across 15 subjects
predict latency differences ranging from 139 to 204 ms over
this same conduction distance. All of these predicted latency
differences are within the 100–300 ms range of the latency
difference increases obtained from the regression across the
210 mm range of arm lengths as shown in Fig. 5. Thus, the
within-subject late and ultra-late latency difference
observed among 15 subjects is consistent with the CVs
estimated by the regression analysis of group differences in
ultra-late latency and by a sub-sample of direct within-
subject measurements. The discrepancy between the within-
subject and across-subject estimates for the CV mediating
the late response is attributable to the additional variance
introduced by inter-subject differences in conduction
distance and both peripheral and central conduction times.
The results provide strong evidence that: (1) there is a
difference in the abundance of slowly and rapidly conduct-
ing thermo-nociceptive afferents innervating the glabrous
and hairy skin of the human hand, and (2) that the receptors
innervated by these different fiber spectra have different
Conduction velocity estimation in six subjects based on peak ultra-late (A, 41 8C distal glabrous and proximal hairy skin) and peak late evoked potentials (B,
51 8C distal and proximal hairy skin)
Distance (mm)N latency, ms (distal,
N latency, ms (proximal,
Subjects Arm length (mm)Distance (in mm)N latency, ms (distal, late
N latency, ms (proximal,
12.9 (7.5) 268,3 (27.1)243 (22.5)
The conduction distance is the difference between the proximal and distal stimulation sites.
Fig. 4. Ultra-late potentials evoked (41 8C) from 3 of the 6 subjects who
showed these responses at both distal (thenar, thick traces) and proximal
(forearm, thin traces) sites.
Y. Granovsky et al. / Pain xx (2005) 1–106
ARTICLE IN PRESS
To evaluate further whether these functional differences
also apply to the foot, we applied similar heat stimuli to
the glabrous skin of the sole and the hairy skin of the dorsum
of the foot. We found that in some individuals it was
difficult to get responses from the glabrous skin of the foot
probably because of skin thickness. However, when
responses appeared, they were markedly delayed compared
to the responses from the dorsal hairy skin. Thus,
stimulation of the glabrous sole at either temperature
(46–51 8C) evoked an ultra-late heat potential in six
subjects, which was associated with a sensation of warmth.
The latency difference between the potential evoked by
stimulation of glabrous skin of the sole and the hand ranged
from 600–900 ms. In contrast, stimulation of the dorsum of
the foot with 51 8C pulses evoked late potentials with much
shorter latencies (ranging between 120–150 ms to the first
negative peak) associated with a strong, sharp pain.
Our data show that (1) the ultra-late potentials evoked
from glabrous skin by 41 and 51 8C stimuli and from hairy
skin by 41 8C stimuli are mediated by unmyelinated (C)
fibers and (2), the late potential evoked from hairy skin by
51 8C stimuli is mediated by finely myelinated Ad fibers.
These interpretations are consistent with the late potential
being associated with pinprick sensations and the ultra-late
potential with mild to strong warmth. The findings
and conclusions are consistent with those obtained by
recordings from fibers innervating monkey skin (Treede
et al., 1995) and with human psychophysical experiments
(Stevens and Choo, 1998; Taylor et al., 1993).
Zaslansky and colleagues (1996) have presented evi-
dence that the pain-related laser-evoked late potential
contains cognitive components. It is unlikely that our
results are confounded by P300-like potentials associated
with attention and cognition because of the strong positive
relationship between evoked potential latency and arm
length (Lorenz and Garcia-Larrea, 2003). Furthermore,
cerebral potentials mediated by C and Ad fibers are
physiologically distinct from P300 cognitive potentials
(Towell and Boyd, 1993).
4.1. Peripheral and central timing
The average latency of the C-fiber mediated potentials
we recorded is shorter than that obtained by brief laser
stimulation of very small skin areas (Bragard et al., 1996;
Kakigi et al., 2003; Opsommer et al., 1999; Tran et al.,
2001, 2002). In addition, the average latencies for the Ad-
mediated contact heat potentials are slightly longer than
those reported for the Ad-mediated potential evoked by
brief laser stimulation of a larger skin area (Bromm and
Lorenz, 1998; Kakigi and Shibasaki, 1991; Kakigi et al.,
2000; Lorenz and Garcia-Larrea, 2003).
The different results we obtained are probably due to
several critical variables, including the location, tempera-
ture, duration, and surface area of the stimuli. These
variables affect the composition, dispersion, and magnitude
of the afferent volley, which affects central temporal and
spatial summation and, therefore, the timing of activity
ascending to the cerebral cortex. We applied temperature-
controlled pulses that are approximately 10 to 300 times
longer over a surface area 30 to 3000 times greater than
those typically used in laser evoked potential studies.
Furthermore, we applied stimuli that were physiologically
selective because the lowest intensity exceeded the
threshold for heat receptors innervated by C fibers but was
Fig. 5. Differences in the peak latency of the of late and ultra-late heat
potentials evoked from the glabrous and hairy skin of the hand from 15
subjects with arms of different length. Late potentials are evoked by 51 8C
stimulation of hairy (dorsal) skin, ultra-late potentials by glabrous
stimulation at either temperature. Peak latencies were measured from the
negative phase (filled circles) or positive phase (open circles) of each heat
potential. (A) Stimuli applied to the glabrousand hairy skinat 41 and 51 8C,
respectively. (B) 51 8C stimulus applied to glabrous and hairy skin. The
regression slope is less in B than in A because the ultra-late heat potential
occurred at a shorter latency when evoked from glabrous skin by a stimulus
of 51, compared to 41 8C (see text).
Y. Granovsky et al. / Pain xx (2005) 1–107
ARTICLE IN PRESS
below the threshold for all receptors innervated by Ad
fibers; the highest intensity was below the threshold for type
1 AMH nociceptors. Brief laser stimulation of very small
skin areas must excite many fewer C fiber-innervated
receptors than our contact heat stimuli, which may explain
the shorter latencies of our C fiber-mediated potentials. As
estimated from the within-subject studies, the CV of
the fibers mediating the ultra-late potential is 1.7 m/s
(Table 1A), which slightly exceeds some previously
reported values (see, for example Kakigi et al., 2003; Tran
et al., 2001, 2002), but is less than the estimates of others
(Magerl et al., 1999). The CV of these fibers ranged from
0.95 to 1.35 as estimated by the regression method from all
subject participants and is within the lower ranges reported
by others (Kakigi et al., 2003; Tran et al., 2001, 2002).
To validate our results, the average peripheral conduc-
tion time should be less than the average evoked potential
latency. The best estimate is provided by using the average
CV obtained from our within-subject analysis because small
measurement errors may give individual CV estimates that
yield negative values for individual central conduction time
(for example, subjects 4 and 6, Table 2A). Using the average
CV of 1.7 m/s, the average peripheral conduction time is
434 ms over the average arm length of 737 mm (Table 1) for
15 subjects. The average N wave latency for 41 8C
stimulation of glabrous skin is 485 ms (Table 1), leaving
51 ms for the estimated average central conduction time.
The average conduction velocity of heat-responsive primate
spinothalamic tract neurons ranges from 17 to 51.6 m/s
(Ferrington et al., 1987; Willis et al., 1974; Zhang et al.,
2000a,b), so a 51 ms conduction time is adequate for the
transmission from the cervical spinal cord to the cerebral
cortex. However, this estimate is significantly shorter than
the central conduction time obtained following laser
stimulation (Cruccu et al., 2000; Iannetti et al., 2003; Tran
et al., 2002), suggesting that, compared with laser
stimulation, the more prolonged and spatially extensive
CHEP excitation of C-innervated receptors favors a more
rapid central transmission, possibly because of differences
in the physiological, spatial, and temporal characteristics of
the afferent volley.
The much higher intracutaneous temperatures achieved
by infra-red laser stimuli should excite many more Ad high
threshold afferents than our low intensity contact heat
(12.9 m/s; Table 2B) predict average peripheral and central
conduction times that are within the ranges of earlier
estimates (Kakigi and Shibasaki, 1991).
We observed that potentials evoked by 51 8C thenar
It is possible that an increased rate of intracutaneous heat
transfer from the higher stimulus temperature resulted in an
earlier activation of heat receptors. However, this result also
suggests the stimulation of a faster-conducting population
with a higher heat threshold. This latter interpretation agrees
with the findings of Opsommer et al. (1999), who used a
reaction time paradigm to estimate a CV of 0.5 m/s in
C-warm afferents and laser-evoked potential latency
measurements to estimate a CV of 1.3 m/s in C-nociceptive
afferents. Our findings are also consistent with the laser
found that, comparedwith painless laser stimulation,painful
stimulation of either hairy or glabrous skin evoked an ultra-
late potential at shorter latencies.
4.2. Other methods selective for C fiber-mediated potentials
Earlier experiments suggested that the excitation of Ad-
mediated potentials suppressed the excitation of central
cells by C fiber afferents (Bragard et al., 1996; Bromm et al.,
1983; Bromm and Treede, 1988, 1991; Opsommer et al.,
1999). These observations were supported by neurophysio-
logical evidence of A fiber inhibition of C-mediated spinal
responses (Chung et al., 1984a,b) and led to the develop-
ment of special techniques to avoid the excitation of
receptors with Ad afferents by restricting the area of skin
stimulation (Bragard et al., 1996; Kakigi et al., 2003;
Opsommer et al., 1999; Tran et al., 2001).
Special stimulating, recording, and selective averaging
methods have also been used to detect the C fiber-mediated
ultra-late responses (Arendt-Nielsen, 1990; Bragard et al.,
1996; Towell et al., 1996). By selecting a relatively large
stimulus area, an intensity below the threshold for exciting
Ad afferents, and by focusing on a skin area richly supplied
with C-innervated heat receptors, other investigators have
evoked C fiber-mediated ultra-late potentials selectively
with laser stimulation (Cruccu et al., 2003; Iannetti et al.,
2003; Magerl et al., 1999; Valeriani et al., 2002a,b).
Other CHEP methods have been used recently to evoke
cerebral potentials mediated by Ad afferents, but the
selective excitation of C fibers could not be established
(Arendt-Nielsen and Chen, 2003; Chen et al., 2001; Harkins
et al., 2000; Itskovich et al., 2000; Valeriani et al., 2002).
Compared to most laser stimulation techniques, the
advantages of the CHEP method include the control of
peak temperature and the possibility of stimulating more
receptors with a relatively long duration pulse over a much
larger surface area. The Ad-mediated laser evoked potential
is abnormal in early diabetic polyneuropathy (Agostino
et al., 2000a,b), so the ability to assess C fiber function by
both laser and contact heat may facilitate the detection of
small fiber polyneuropathies (Krarup, 2003).
4.3. Comparison of hairy and glabrous skin
Towell and colleagues (1996) recorded cerebral poten-
tials mediated by Ad fibers following both painful and
painless laser stimulation of the glabrous skin of the palm.
The most intense contact heat stimulus we applied to
glabrous skin (51 8C) failed to evoke an Ad-mediated
Y. Granovsky et al. / Pain xx (2005) 1–108
ARTICLE IN PRESS
response. Because 51 8C is below the average threshold of
monkey AMH type 1 receptors but above that of AMH type
2 receptors, the combined evidence supports the interpret-
ation that the human palm has few, if any, AMH type 2
receptors. The possibility that AMH type 1 receptors in
human glabrous skin may be excited by innocuous infrared
laser stimulation is consistent with estimates of receptor
depth and laser heat distribution in the skin (Haimi-Cohen
et al., 1983; Tillman et al., 1995).
Our findings support the hypothesis of Treede and
colleagues (1995) about the differential innervation of
monkey hairy and glabrous skin and extend the hypothesis
to include the human hand. The combined neurophysiolo-
gical and psychophysical data (Gescheider et al., 1994;
Greenspan et al., 1993; Stevens and Choo, 1998; Schlereth
et al., 2001; Taylor et al., 1993), suggest that the glabrous
surface of the primate hand has a wide range of thermal and
mechanical sensitivity which, together with the relative
paucity of low threshold heat nociceptors, enhances its
function as an exploratory surface.
Supported by NIH/NIAMS grant AR46045, and the
Department of Veteran’s Affairs (KLC). Dr Matre was
supported by the National Institute of Occupational Health,
Norway. The availability of the CHEP stimulator was made
possible through the Pfizer Company (Pfizer Global
Research and Development).
Agostino R, Cruccu G, Iannetti GD, Innocenti P, Romaniello A, Truini A,
Manfredi M. Trigeminal small-fibre dysfunction in patients with
diabetes mellitus: a study with laser evoked potentials and corneal
reflex. Clin Neurophysiol 2000a;111:2264–7.
Agostino R, Cruccu G, Romaniello A, Innocenti P, Inghilleri M,
Manfredi M. Dysfunction of small myelinated afferents in diabetic
polyneuropathy, as assessed by laser evoked potentials. Clin Neuro-
Arendt-Nielsen L. Second pain event related potentials to argon laser
stimuli: Recording and quantification. J Neurol Neurosurg Psychiat
Arendt-Nielsen L, Chen ACN. Lasers and other thermal stimulators for
activation of skin nociceptors in humans. Neurophysiol Clin 2003;33:
Bragard D, Chen ACN, Plaghki L. Direct isolation of ultra-late (C-fibre)
evoked brain potentials by CO2laser stimulation of tiny cutaneous
surface areas in man. Neurosci Lett 1996;209:81–4.
Bromm B, Lorenz J. Neurophysiological evaluation of pain. Electro-
encephalogr Clin Neurophysiol 1998;107:227–53.
Bromm B, Treede RD. Pain related cerebral potentials: Late and ultra-late
components. Int J Neurosci 1988;33:15–23.
Bromm B, Treede RD. Laser-evoked cerebral potentials in the assessment
of cutaneous pain sensitivity in normal subjects and patients. Rev
Neurol (Paris) 1991;147:625–43.
Bromm B, Neitzel H, Tecklenburg A, Treede RD. Evoked cerebral
potential correlates of C-fibre activity in man. Neurosci Lett 1983;43:
Chen ACN, Niddam DM, Arendt-Nielsen L. Contactheat evoked potentials
as a valid means to study nociceptive pathways in human subjects.
Neurosci Lett 2001;316:79–82.
Chung JM, Fang ZR, Hori Y, Lee KH, Willis WD. Prolonged inhibition of
primate spinothalamic tract cells by peripheral nerve stimulation. Pain
Chung JM, Lee KH, Hori Y, Endo K, Willis WD. Factors influencing
peripheral nerve stimulation produced inhibition of primate spinotha-
lamic tract cells. Pain 1984b;19:277–93.
Cruccu G, Iannetti GD, Agostino R, Romaniello A, Truini A, Manfredi M.
Conduction velocity of the human spinothalamic tract as assessed by
laser evoked potentials. Neur Rep 2000;11:3029–32.
Cruccu G, Pennisi E, Truini A, Iannetti GD, Romaniello A, Le Pera D, De
Armas L, Leandri M, Manfredi M, Valeriani M. Unmyelinated
trigeminal pathways as assessed by laser stimuli in humans. Brain
Ferrington DG, Sorkin LS, Willis Jr WD. Responses of spinothalamic tract
cells in the superficial dorsal horn of the primate lumbar spinal cord.
J Physiol 1987;388:681–703.
Gescheider GA, Bolanowski SJ, Hall KL, Hoffman KE, Verrillo RT. The
effects of aging on information-processing channels in the sense of
touch: I. Absolute sensitivity. Somatosens Mot Res 1994;11:345–57.
Greenspan JD, Taylor DJ, McGillis SL. Body site variation of cool
perception thresholds, with observations on paradoxical heat. Somato-
sens Mot Res 1993;10:467–74.
Haimi-Cohen R, Cohen A, Carmon A. A model for the temperature
distribution in skin noxiously stimulated by a brief pulse of carbon
dioxide laser radiation. J Neurosci Methods 1983;8:127–37.
Harkins SW, Price DD, Roy A, Itskovich VV, Fei DY. Somatosensory
evoked potentials associated with thermal activation of type II Adelta
mechanoheat nociceptive afferents. Int J Neurosci 2000;104:93–111.
Iannetti GD, Truini A, Romaniello A, Galeotti F, Rizzo C, Manfredi M,
Cruccu G. Evidence of a specific spinal pathway for the sense of
warmth in humans. J Neurophysiol 2003;89:562–70.
Itskovich VV, Fei DY, Harkins SW. Psychophysiological and psycho-
physical responses to experimental pain induced by two types of
cutaneous thermal stimuli. Int J Neurosci 2000;105:63–75.
Kakigi R, Shibasaki H. Estimation of conduction velocity of the spino-
thalamic tract in man. Electroencephalogr Clin Neurophysiol 1991;80:
Kakigi R, Watanabe S, Yamasaki H. Pain-Related somatosensory evoked
potentials. J Clin Neurophysiol 2000;17:295–308.
Kakigi R, Tran TD, Qiu Y, Wang X, Nguyen TB, Inui K, Watanabe S,
Hoshiyama M. Cerebral responses following stimulation of unmyeli-
nated C-fibers in humans: electro- and magneto-encephalographic
study. Neurosci Res 2003;45:255–75.
Krarup C. An update on electrophysiological studies in neuropathy. Curr
Opin Neurol 2003;16:603–12.
Lorenz J, Garcia-Larrea L. Modulation attentionnelle et cognitive des
reponses evoquees par laser. Neurophysiol Clin 2003;33:293–301.
Magerl W, Ali Z, Ellrich J, Meyer RA, Treede RD. C- and A delta-fiber
components of heat-evoked cerebral potentials in healthy human
subjects. Pain 1999;82:127–37.
Opsommer E, Masquelier E, Plaghki L. Determination of nerve conduction
velocity of C-fibres in humans from thermal thresholds to contact heat
(thermode) and from evoked brain potentials to radiant heat (CO2
laser). Neurophysiol Clin 1999;29:411–22.
SchlerethT, MagerlW,TreedeR. Spatialdiscriminationthresholdsforpain
and touch in human hairy skin. Pain 2001;92:187–94.
Stevens JC, Choo KK. Temperature sensitivity of the body surface over the
life span. Somatosens Mot Res 1998;15:13–28.
Taylor DJ, McGillis SLB, Greenspan JD. Body site variation of heat pain
sensitivity. Somatosens Mot Res 1993;10:455–65.
Y. Granovsky et al. / Pain xx (2005) 1–109
ARTICLE IN PRESS
Tran TD, Lam K, Hoshiyama M, Kakigi R. A new method for measuring the Download full-text
conduction velocities of Abeta-, Adelta- and C-fibers following electric
and CO(2) laser stimulation in humans. Neurosci Lett 2001;301:187–90.
Tran TD, Inui K, Hoshiyama M, Lam K, Kakigi R. Conduction velocity of
the spinothalamic tract following CO2laser stimulation of C-fibers in
humans. Pain 2002;95:125–31.
Tillman DB, Treede RD, Meyer RA, Campbell JN. Response of C fibre
nociceptors in the anaesthetized monkey to heat stimuli: estimates of
receptor depth and threshold. J Physiol 1995;485:753–65.
Towell AD, Boyd SG. Sensory and cognitive components of the CO2 laser
evoked cerebral potential. Electroencephalogr Clin Neurophysiol 1993;
Towell AD, Purves AM, Boyd SG. CO2 laser activation of nociceptive and
non-nociceptive thermal afferents from hairy and glabrous skin. Pain
Treede RD, Meyer RA, Raja SN, Campbell JN. Evidence for two different
heat transduction mechanisms in nociceptive primary afferents
innervating monkey skin. J Physiol 1995;483:747–58.
Valeriani M, Le Pera D, Niddam D, Chen ACN, Arendt-Nielsen L. Dipolar
modelling of the scalp evoked potentials to painful contact heat
stimulation of the human skin. Neurosci Lett 2002a;318:44–8.
Valeriani M, Restuccia D, Le Pera D, De Armas L, Maiese T, Tonali P.
Attention-related modifications of ultra-late CO2 laser evoked
potentials to human trigeminal nerve stimulation. Neurosci Lett
Willis WD, Trevino DL, Coulter JD, Maunz RA. Responses of primate
spinothalamic tract neurons to natural stimulation of hindlimb.
J Neurophysiol 1974;37:358–72.
Zaslansky R, Sprecher E, Tenke CE, Hemli JA, Yarnitsky D. The P300 in
pain evoked potentials. Pain 1996;66:39–49.
Zhang X, Honda CN, Giesler Jr GJ. Position of spinothalamic tract axons in
upper cervical spinal cord of monkeys. J Neurophysiol 2000a;84:
Zhang X, Wenk HN, Honda CN, Giesler Jr GJ. Locations of spinothalamic
tract axons in cervical and thoracic spinal cord white matter in
monkeys. J Neurophysiol 2000b;83:2869–80.
Y. Granovsky et al. / Pain xx (2005) 1–1010
ARTICLE IN PRESS