- Email: [email protected]
Pain-related and cognitive components of somatosensory evoked potentials following CO2 laser stimulation in man
ELSEVIER
Electroencephalography and clinical Neurophysiology 100 (1996) 105- I 14
Pain-related and cognitive components of somatosensory evoked potentials following CO 2 laser stimulation in man Masutaro Kanda ~,c, Naohito Fujiwara a, Xiaoping Xu ~, Kazuo Shindo ~,b, Takashi Nagamine a, Akio Ikeda a, Hiroshi Shibasaki a,* a Department ofBrain Pathophysiology, Kyoto University School ofMedicine, Shogoin, Sakyo-ku, Kyoto 606-01. Japan b Department ofAnesthesiology, Kyoto University School ofMedicine, Kyoto, Japan c Division ofNeurology, Takeda General Ho.~pital, Kyoto, Japan Accepted for publication: 17 August 1995
Abstract
We recorded cortical potentials evoked by painful CO 2 laser stimulation (pain SEP), employing an oddball paradigm in an effort to demonstrate event-related potentials (ERP) associated with pain. In 12 healthy subjects, frequent (standard) pain stimuli (probability 0.8) were delivered to one side of the dorsum of the left hand while rare (target) pain stimuli (probability 0.2) were delivered to the other side of the same hand. Subjects were instructed to perform either a mental count or button press in response to the target stimulio Two early components (N2 and P2) of the pain SEP demonstrated a Cz maximal distribution, and showed no difference in latency, amplitude or scalp topography between the oddball conditions or between response tasks. In addition, another positive component (P3) following the P2 was recorded maximally at Pz only in response to the target stimuli with a peak latency of 593 msec for the count task and 560 msec for the button press task. lts scalp topography was the saine as that for electric and auditory P3. The longer latency of pain P3 can be explained not only by its slower impulse conduction but also by the effects of task difficulty in the oddball paradigm employing the pain stimulus compared with electric and auditory stimulus paradigms. It is concluded that the P3 for the pain modality is mainly related to a cognitive proeess and corresponds to the P3 of electric and auditory evoked responses, whereas both N2 and P2 are mainly pain-related components. Keywords: CO 2 laser stimulation; Pain SEP; Oddball paradigm; Event-related potential; Pain-related component; Cognitive component; Scalp topography
1. Introduction
Somatosensory evoked potentials (SEPs) following CO 2 laser stimulation are mainly composed of 3 components (NI, N2 and P2) (Miyazaki et al., 1994; Xu et al., 1995). One of the most important factors for clearly recording the initial component (N1) is the intensity of the stimulus (Treede et al., 1988; Xu et al., 1995). Although the NI elicited by hand stimulation is localized to the contralateral central a n d / o r midtemporal area, that in response to foot stimulation is located in the midline central area, suggesting that the N1 is an exogenous component generated in the primary somatosensory cortex (Miyazaki et al., 1994; Xu et ai., 1995).
* Corresponding author. Tel.: + 81 75 7513601 ; Fax: + 81 75 7513202.
With regard to N2 and P2 of the pain SEPs, however, both components are maximal at Cz. The peak latency of P2 is between 300 and 400 msec, which is similar to the latency of the P3 or P300 component of the event-related potential (ERP) recorded in the oddball paradigm using auditory as well as electric stimuli (Barrett et al., 1987). Miyazaki et al. (1994) reported that the amplitude of P2 was significantly larger with a preceding waming signal than without. Furthermore, Beydoun et al. (1993) showed variability in the peak-to-peak amplitude of negative (N2) and positive (P2) components of pain SEPs by modulating various conditions. In particular, amplitude decreased significantly during both a distraction task and drowsiness and was absent during stage II sleep. In the oddball paradigm employing CO 2 laser stimuli, Towell and Boyd (1993) first demonstrated another positive potential following the P2, which appeared only in the
0168-5597/96/$15.00 © 1996 Elsevier Science Ireland Ltd. AIl rights reserved SSDI 00 1 3 - 4 6 9 4 ( 9 5 ) 0 0 1 90-5
EEP 94181
106
M.Kandaetal./Electroencephalography andclinicalNeurophysiology100(1996)105-114
oddball target task. Latency and amplitude of the preceding negative (N2) and positive (P2) peaks in response to target stimuli were not significantly different from those to the standard stimuli. They adopted mental counting for the oddball target task, and the potentials were recorded from only 1 electrode placed at Cz. Therefore, the field distribution of the potential, effects of different response tasks, and differences from other stimulus modalities remained unclear. The aim of the present study, therefore, was to elucidate the characteristics of the endogenous, cognitive component of pain SEPs following CO 2 laser stimulation by employing an oddball paradigm.
2. Methods 2.1. Subjects
Twelve normal subjects (9 men and 3 women; ail right handed) volunteered for this study. Each subject gave written informed consent before the experiment. Their ages ranged from 21 to 39 years (mean 28.2 years), and their height ranged from 158 to 180 cm (mean 167.2 cm). No particular medication was given for the present study. The subject was seated in a reclining armchair in a quiet and electrically shielded room, with the ambient temperature controlled at about 24°C, while their eyes were kept open. The electroencephalogram (EEG) was monitored to check the vigilance level of the subject, and, when it was reduced, as judged by the monitored EEG, the subject was instructed to keep the same level. 2.2. Stimulus 2.2.1. Pain stimulus
Pain stimuli were provided by a special CO 2 laser stimulator (Nippon lnfrared Industries Co. Ltd., Kawasaki, Japan). Its maximum power was 23 W, but an attenuator limited the output to approximately 12.5 W for safety reasons. The laser wave length was 10.6 /xm and the strength of power was 6 W with an irradiation duration of 60 msec, while the irradiation beam was adjusted to about 6 mm in diameter on the skin with the aid of a helium-neon laser (Kakigi et al., 1989). The caiculated stimulus intensity was approximately 12.8 m J / m m ~ which elicited a painful sensation like a "pinprick" in normal subjects. The use of the machine was approved by the Ethics Committee of Kyoto University School of Medicine. Stimuli were presented to either the radial or the ulnar side of the dorsum of the left hand, avoiding the central part (see Section 2.3) (Fig. lA), while the stimulated side was masked so that the subject could not see it. The stimulus spot on each side of the dorsum was slightly shifted for each stimulus in order to avoid sensitization, habituation and receptor fatigue. The interstimulus interval varied at random between 3 and 7 sec.
A. co2 Laser Stirn. Freq. stim. p=o.s
Rate stim. -_. 2
B. E]ectric Stirn. Freq. stim.
Rare stim.
p=0.8~.2
Fig. 1. Schematic diagram illustrating examples of the pain (A) and electdc stimulation (B) for the oddball paradigm. In A pain stimuli for targets (rare) were applied to the radial side of the left hand dorsum at a probability of 0.2, whereas those for standards (freq.) were applied to the ulnar side at a probability of 0.8. In B electric stimuli for targets were applied to the second digit of the left hand at a probability of 0.2, whereas those for standards were applied to the fourth digit at a probability of 0.8. For both modalities, 1 session consisted of 20 frequent (standard) and 5 rare (targe0 stimuli.
2.2.2. Electric stimulus
Percutaneous electric stimuli were presented to either the second or the fourth digit of the left hand with ring electrodes (Fig. lB). The cathode was placed at the proximal interphalangeal joint and the anode at the distal interphalangeal joint of the corresponding digit. The stimulus puise width was 0.2 msec, while its voltage was adjusted to 20% above the sensory threshold, which ranged from 80 to 140 V. The interstimulus intervai for electric stimulation also varied randomly between 3 and 7 sec. 2.2.3. Auditory stimulus
Tone bursts (100 msec duration, 10 msec rise/fall rime, 45 dB SPL above threshold) were presented binaurally through headphones. The frequency of the tone was either 1000 or 2000 Hz (see below). The interstimulus interval for auditory stimulation varied randomly between 1.5 and 2 sec. 2.3. Oddball paradigm
Pain stimuli for target (rare) trials were applied to either the ulnar or the radial side of the left hand dorsum, depending on the session, at a probability of 0.2, whereas those for standard (frequen0 triais were applied to the other side of the same hand at a probability of 0.8 (Fig. lA). Likewise, electric stimuli for target (rare) trials were applied to either the second or the fourth digit of the left hand, depending on the session, at a probability of 0.2, whereas those for standard (frequent) trials were applied to the other digit with a probability of 0.8 (Fig. lB). For auditory stimulation, the 2000 Hz tone was presented binaurally with a probability of 0.2 for target stimuli, whereas the 1000 Hz tone was given at a probability of 0.8 for standard stimuli.
M. Kanda et al./Electroencephalography
and clinical Neurophysiology 100 (1996) 105-114
The response tasks to the target stimuli were the same for each stimulus modality. Subjects were asked to either mentally count the number of infrequently presented target stimuli or to pressa button with the right hand in response to each target promptly and as correctly as possible. The response tasks were varied over sessions. One session consisted of 20 frequent (standard) and 5 rare (target) stimuli, irrespective of either the stimulus modality or the kind of response task. Recordings were performed using auditory, electric and pain stimulus modalities in this order. For each stimulus modality, 4 successive sessions using each of the mental count and button press tasks were given in this order. Consequently every subject was given 24 sessions in all. For the button press task for ail stimulus modalities, reaction time was measured from the onset of the target stimulus to the onset of the button press. The target side and the target digit for the first session for pain and electric stimulation, respectively, were determined randomly and counterbalanced across subjects, and were changed alternatively from session to session.
2.4. Recording In ail subjects, EEGs were recorded with 15 silver/silver chloride shallow cup electrodes (1 cm in diameter) placed on the scalp at Fpz, F3, Fz, F4, T3, C3, Cz, C4, T4, P3, Pz, P4, O1, Oz and 0 2 according to the International 10-20 System, using a Pathfinder with 32 recording channels (Nicolet Biomedical Inc., Wisconsin, USA). Each scalp electrode was referenced to linked earlobes. An electrode placed 1 cm just below the right eye was referenced to another electrode placed just above that eye for monitoring artifacts arising from blinks and eye movements. The ground electrode for pain and auditory stimulation was placed midway between Fpz and Fz, and that for electric stimulation was a stainless steel plate of 3.0 cm × 3.0 cm placed on the skin at the left antero-lateral part of the neck. Ail electrodes were filled with electroconductive jelly, and the impedance was maintained at less than 5 k J2. The bandpass filter was set to 0.05-30 Hz ( - 3 dB) for pain and electric stimulation and 0.05-60 Hz ( - 3 dB) for auditory stimulation.
107
2.6. Data analysis After confirming the reproducibility of wave forms among like sessions with respect to stimulus modality (pain, electric or auditory), oddball condition (target or standard) and response task (mental count or button press) for each subject, group average responses for the 12 modality/condition/response combinations were obtained by averaging data with respect to stimulus onset. The amplitude of each identifiable peak in the group averaged responses was measured from the baseline, which was determined by averaging the first 10% of the epoch for each electrode. For analyzing the latency and amplitude of N2 and P2 components of the pain SEPs, an ANOVA with repeated measurements was performed with factors of oddball condition and response task. For analysis of latency and amplitude of each of the 3 oddball components (P3) obtained using the different stimulus modalities, ANOVAs with repeated measurements were performed with factors of stimulus modality and response task. For evaluating the reaction time for the 3 kinds of stimulus modality, an ANOVA with repeated measurements was performed with stimulus modality as the single factor. Although a multiplicative dipole model is assumed for ERP components, the ANOVA model is based on an additive assumption and can consequently lead to misinterpretation of amplitude distribution (McCarthy and Wood, 1985; Naumann et al., 1992). In an attempt to reduce this problem for the topographic analyses of N2, P2 and P3, amplitudes were normalized by scaling the data relative to the maximum grand mean of the ERP amplitude, before performing ANOVA with the same designs as used for the analysis of the raw amplitude data. The significance level was set at P < 0.01. In addition, the P3 for each stimulus modality for each response task was displayed as topographic maps by using the Nicolet Pathfinder TMAP software, in which purely linear rectangular interpolation between electrode positions was performed on the normalized data.
3. Results
2.5. Averaging One epoch for pain and electric ERPs was 2 sec including a prestimulus period of 200 msec, and that for the auditory ERP was 1 sec including a prestimulus period of 100 msec. The sampling rate for the analog-to-digital conversion was 125 Hz for the former 2 ERPs and 250 Hz for the latter. For each session, responses to 20 frequent stimuli and those to 5 rare (target) stimuli were averaged separately time-locked to stimulus onset.
In the count task, it is impossible to detect target responses in which subjects failed to perform the task correctly. However, as the number of target stimuli reported by the subject after each session was correct, it was considered that there was no inaccurate performance. In the button press task, there were only a few target trials with incorrect responses, and these were excluded from the averaged response for each stimulus modality as were trials contaminated by artifact.
108
M. Kanda et al. / Electroencephalography and clinical Neurophysiolo gy 100 (1996) 105-114
Frequent
Rare
Button press
Count EOG"
EOG'
Count
PAIN
Fz
N2
Cz
P2
Cz
P3
PZ i
10pV +l
P2
i i i i , k l l l
, stim.
I 1800 mec
i
P3 i
i
i
i
i
[
I
1800 ms
stim.
I'''''',LI stim.
i
i
i
i
i
i
i
i
i
I
~' stirn,
1800 ms
1800 fl~ec
EOG -,
Fz
.
- ......
ELECTRIC E0G
Button F z press cz I
P3
P3 I
laoo stim,
PZ 10pV +l
I' ' stil.
''
'
I'
'''1 1800
mec
stim.
' ' ' ' ' '
'
t
3.1. W a v e f o r m s
I n the averaged wave forrns of pain SEPs o b t a i n e d in the oddball p a r a d i g m u s i n g C O 2 laser stimulation, early negative-positive potentials (N2-P2) were recorded regardless o f oddball conditions (frequent or rare) and response tasks (mental count or b u t t o n press) (Fig. 2). F o l l o w i n g these potentials, a positive deflection at a latency o f 572 msec for the count response and 508 msec for the b u t t o n press response was observed at Pz and Cz o n l y in the rare (target) conditions. In the grand average w a v e forms of pain SEPs across ail subjects obtained in the oddball p a r a d i g m (Fig. 3), a negative c o m p o n e n t was followed b y a positivity corres p o n d i n g to N 2 and P2, respectively, in both sets of w a v e forms. The latency of N 2 as well as that of P2 was similar b e t w e e n the oddball conditions and b e t w e e n the response tasks. Both N 2 and P2 were m a x i m a l at Cz in ail of the 4 sets o f c o m b i n a t i o n s (oddball c o n d i t i o n and response task) (Fig. 4). A n o t h e r positive c o m p o n e n t was seen after the P2, exclusively in the wave forms of target responses, which was considered to be an oddball c o m p o n e n t and was n a m e d P3 (Fig. 3). The peak latency of P3 was 572 msec
L
L
i
~
i
i
i
[
~soo
stim.
ms
EOG"
1800 msec
Fig. 2. Averaged wave forms of SEPs obtained in an oddball paradigm using CO z laser stimuli in a single subject. Each trace of superimposed wave forms is the result of averaging 2 like sessions in which either the ulnar or the radial side of the left hand dorsum was stimulated as oddball target. In this and the following figures, 3 selected EEG channels (Fz, Cz and Pz) and the electrooculogram (EOG) are shown. "Stim" denotes the rime of stimulus onset. Relative negativity in the exploring electrode is shown upward in this and the following figures. A positive potential (arrow) is seen only in the rare (target) conditions, at a latency of 572 msec for the count response and 508 msec for the button press response.
ms
i
ç
Fz
AUD. Cz Pz
,o.v~i
P3 I
i
i
h
i
P3 i
h
i
i
I
900 stim.
ms
.
.
.
.
.
.
i
L
i
I
900
stim,
ms
Fig. 3. Grand average wave forms obtained in the oddball paradigm across ail subjects. Solid lines show responses to the rare (target) stimuli, whereas interrupted lines show those to the frequent (standard) stimuli. In pain SEPs, a negative peak (N2) and the following positive peak (P2) are observed in the target as well as standard conditions for both response tasks (count and button press). Another positive potential (pain P3) can be seen following P2 only in the target condition. Electric P3 and auditory P3 also appear only in the target condition for both response tasks.
for the count response and 508 msec for the b u t t o n press response. The amplitude of P3 was m a x i m a l at Pz regardless of the response task (Fig. 5).
RARE
FREQUENT
~~
N~
-1
-1 Fz
Cz ELECTRODE
Pz
Fz
Cz
Pz
ELECTRODE
Fig. 4. Amplitudes of N2 and P2 of pain SEP at 3 electrodes (Fz, Cz and Pz). In each figure, amplitudes for the count response are plotted with solid lines, whereas those for the button press response are plotted with interrupted lines. Both N2 and P2 are maximal at the Cz electrode, regardless of the oddball condition or response task.
M. Kanda et al./Electroencephalography and clinical Neurophysiolo gy 100 (1996) 105-114
In the grand average wave forms of SEPs obtained in the oddball paradigm using electric stimulation (Fig. 3), a large positive component appeared maximally at Pz only in response to the target stimuli for both response tasks. This was called the electric oddball P3, The peak latency of this P3 was 428 msec for the count response and 404 msec for the button press response, and its amplitude was maximal at Pz regardless of the kind of response task (Fig. 5). Grand average wave forms of auditory evoked potentials (AEPs) across ail subjects obtained in the oddball paradigm showed a large positive comportent corresponding to the conventional P300 or P3 which appeared only in the target condition for both response tasks (Fig. 3). The peak latency of P3 was 352 msec for the count response and 364 msec for the button press response, and its amplitude was also maximal at Pz irrespective of the kind of response task (Fig. 5).
109
12> =-
/
COUNT BUTTON --O.-- ..-o-ELECTRIC --42-- ---oe-AUD. -..,b-- ...~..
.. ..-" .,~ ..." . , ~
PAIN
10-
III
oO~
~
"°
~,.....I
8-
..e
~
........•
13.
< 13.
6-
4
i Fz
i
[
Cz
Pz
ELECTRODE Fig. 5. P3 amplitude of 3 different kinds of ERPs at 3 electrodes (Fz, Cz and Pz). P3 is maximal at the Pz electrode regardless of the stimulus modality or response task.
Table 1 Mean latency and amplitude of N2 and P2 of pain SEPs and P3 components of 3 different kinds of evoked potentials obtained in the oddball paradigm in 12 subjects (mean + S.D.) Comp/RT
Electrode
Mod
Task
Cond
Mean dz S.D.
Cz Cz Cz Cz Cz Cz Cz Cz Pz Pz Pz Pz Pz Pz
Pz Pz Pz
p p p p p p p p p p e e a a p e a p e a
c c b b c c b b c b c b c b b b b b b b
freq rare freq rare freq rare freq rare rare rare rare rare rare rare rare rare rare rare rare rare
211 214 218 217 318 315 323 318 593 560 425 417 351 336 668 467 372 108 49 19
Cz Cz Cz Cz Cz Cz Cz Cz Pz Pz Pz Pz Pz Pz
p p p p p p p p p p e e a a
c c b b c c b b c b c b c b
freq rare freq rare freq rare freq rare rare rare rare rare rare rare
- 5.3 ___ -6.2+ - 6.6 i -8.0+ 8.1 + 10.8 + 8.7 _ 9.0_+ 10.6+ 10.5 + 11.2 dz 12.2 dz 11.7 + 13.4_+
Latency (msec) N2
P2
P3
RT
RT-P3
_+ 22 _+ 23 dz 19 dz 18 dz 22 dz 26 + 21 dz 29 + 3l + 54 dz 69 dz 58 dz 31 dz 38 dz 91 dz 93 dz 90 dz 94 _+105 _+ 82
Amplitude ( btV) N2
P2
P3
3.1 4.3 3.9 5.0 1.8 4.3 3.4 3.0 3.8 1.8 3.7 5.5 4.8 5.9
Comp, component; RT, reaction time; RT-P3, time difference from the P3 peak to RT; Mod, stimulus modality; p, e and a, pain, electric and auditory, respectively; c and b, count and button press response task, respectively; Cond, oddball conditions; freq and rare, frequent (standard) and rare (target) condition, respectively.
M. Kanda et al./ Electroencephalography and clinical Neurophysiology 100 (1996) 105-114
110
Table 2 A N O V A s for latency and amplitude of N2, P2 and P3, and for R T and RT-P3 Factor ( d f )
Latency N2
Response time P2
P3
RT
Amplitude RT-P3
N2
P2
P3
c(1, ll) t ( l , 11) c t ( l , 11) s (2, 22) t ( 1 , 11) st (2, 22) s (2, 22) N2 and P2: measured at Cz; P3: measured at Pz. Abbreviations of A N O V A terms: c = oddball condition; s = stimulus modality; t = response task. Significance values are calculated using Greenhouse-Geisser correction: - P > 0.01; * P < 0.01; * * P < 0.005; * * " P < 0.001.
Table 3 A N O V A s for scalp topography of N2, P2 and P3 with application of m a x i m u m normalization Condition
Modality
Factor (df)
N2
c(1, ll) t(l, Il) e ( 1 4 , 154) c t ( l , 11) et (14, 154) ce (14, 154) cet ( 14, 154)
. . . . . . -
P2
-
Factor (df)
where its amplitude was maximal among the 15 electrodes. An ANOVA on normalized P2 also showed no effect for the ANOVA terms described above, apart from electrode site (e) (Table 3).
P3
s ( 2 , 11) t ( 1 , 11) e ( 1 4 , 154) s t ( 2 , 22) et (14, 154) es (28, 308) est ( 2 8 , 3 0 8 )
Abbreviations of A N O V A terms: c = oddball condition; e = electrode; s = stimulus modality; t = response task. Significance values are obtained using Greenhouse-Geisser correction: P > 0.01; *** P < 0 . 0 0 1 .
3.2. N2 and P2 of pain SEPs 3.2.1. Latency and amplitude : Mean values of latency and amplitude for N2 and P2 of pain SEPs obtained in the oddball paradigm and measured at Cz for each combination of oddball condition and response task are shown in Table 1, and Table 2 shows the ANOVA results. For the 3 terms for ANOVA involving oddball condition (c), response task (t) and condition x response task interaction (ct), neither N2 nor P2 was significantly different in latency (Table 2). Similarly, the raw amplitude of the 2 components did not show statistically significant differences for the ANOVA terres. 3.2.2. Scalp topography An ANOVA on the N2 amplitudes normalized by scaling as a percentage of the Cz amplitude showed no effect of oddball condition (c), response task (t), condition X response task interaction (ct), electrode site X response task interaction (et), or condition X electrode site × response task interaction (cet), and only electrode site (e) showed significant differences (Table 3). Similarly, P2 was normalized as a percentage of the value at the Cz electrode
3.3. Oddball component (P3) 3.3.1. Latency and amplitude Mean values of latency and amplitude of the oddball component (P3) evoked in the 3 different tests are shown in Table 1. The iatency of P3 was significantly different between stimulus modalities (s), although it showed no difference between the response tasks (t) or for the stimulus modality x response task interaction (st) (Table 2). An ANOVA on the raw amplitude of P3 did not show differences between the stimulus modalities (s), response tasks (t) and stimulus modality X response task interaction (st) (Table 2). ANOVAs on reaction time (RT) obtained from the button press responses in the target conditions and the time difference between the corresponding P3 peak and the reaction time (RT-P3) showed significant differences between stimulus modalities (s) (Table 2). 3.3.2. Scalp topography An ANOVA on the normalized P3 amplitude (scaled with respect to amplitude at Pz) showed no effect of stimulus modality (s), response task (t), stimulus modality X response task interaction (st), electrode site X response task interaction (et), electrode site x stimulus modality interaction (es) and electrode site X stimulus modality X response task interaction (est), although electrode site (e) showed a significant difference (Table 3). Fig. 6 demonstrates topographic maps of the oddball component P3, comparing pain, electric and auditory modalities for each of the count and button press responses. In ail 6 maps, the P3 component was maximal at Pz, extended symmetrically to the bilateral parietal electrodes. The field distribution of each P3 on the left hemisphere was similar to the corresponding one on the right, and, therefore, further statistical
M. Kanda et a l . / Electroencephalography and clinical Neurophysiology 100 (1996) 105-114
111
P3 Count
Button press
Pain
Electric
Auditory
-1.0 ~
1.0
Fig. 6. Scaip topography of the oddball component P3 of pain, electric and auditory ERPs ail obtained in the oddball paradigm for count and button press responses. Note similar distribution pattem in all maps, which is maximal at Pz and extends symmetrically fo the bilateral parietal electrodes.
analyses comparing interhemispheric differences were not performed. 4. Discussion
The wave form of pain SEPs obtained in the present study demonstmted N2 and P2 components, both of which were present regardless of the oddball condition (rare, target or frequent, standard) and of the kind of response task (mental count or button press). Miyazaki et al. (1994) reponed that P2 increased in amplitude by applying a predictive waming signal, whereas N2 did not. Beydoun et al. (1993) reponed that the amplitude difference between
negative and positive peaks of CO 2 laser SEPs, which correspond to N2 and P2, respectively, diminished by distraction and during lowered arousal state. Moreover, even early cortical components of the electric SEPs have been suggested to be modulated by attention (Desmedt et al., 1983; Garda-Larrea et al., 1991). In the present study, however, neither the peak latency nor the amplitude of N2 and P2 was significantly different according to the oddball condition or response task and there was no condition × response interaction. This observation also confirmed and extended the finding by Towell and Boyd (1993) which used the count task and showed non-significant changes in the peak-to-peak amplitude (N2-P2) between the oddbali
112
M. KarMa et aL / Electroencephalography and clinieal Neurophysiology I00 (1996) 105-114
conditions. Furthermore, we found that the scãlp distribution of N2 and P2 was not different between the response tasks or between the oddball conditions. These findings suggest that the oddball type of manipulation does not significantly alter either N2 or P2. In addition to N2 and P2, another positive component (P3) following the P2 was recognized exclusively in the oddball target conditions for both count and button press responses, with a mean peak latency of 593 msec for the count response and 560 msec for the button press response, and a mean amplitude of 10.6/zV for the former and 10.5 /zV for the latter. The late positive component reported by Towell and Boyd (1993) had a mean peak latency of 621 msec and a peak-to-peak amplitude (N2-P3) of 22.5 /zV. Although these values were slightly different from those of the P3 observed in the present study, this component is considered to be a cognitive component because it appeared exclusively in the oddball target condition irrespective of the response tasks. While they used only 1 electrode placed at Cz (Towell and Boyd, 1993), we used 15 electrodes to delineate scalp distribution of the potential. As a result, we showed that the P3 potential was maximal at Pz, in contrast with the P2 which was maximal at Cz, suggesting different generator sources for these 2 positive components. Becker et al. (1993), by studying SEPs produced by intracutaneous electrical stimulation of a fingertip, reported that the pain and intensity components were larger at Cz than at Pz, and vice versa for the cognitive component. According to Donchin et al. (1978), endogenous components of ERPs for any stimulus modality are characterized by: (1) the component must be non-obligatory responses to stimuli; (2) the amplitude, latency, and scalp distribution of the potential are often invariant to changes in the physical parameters of the eliciting stimulus; (3) the variance of the endogenous component is normally accounted for by variation in the tasks assigned to the subject. Therefore, it is conceivable that, although P2 might be mainly related to pain perception depending on physical parameters of the pain stimulus a n d / o r the subject's attention directed to it, P3 is an endogenous component related to the "cognition," which means "categorization processes" through which a subject makes a decision to perform the task in an oddball paradigm. There was a significant difference in the P3 latency between the 3 stimulus modalities (pain, electric and auditory) (Table 2). The P3 latency of the electric ERP was longer than that of the auditory ERP, confirming previous reports (Pieton et al., 1984; Barrett et al., 1987). P3 latency of the pain ERP was much longer than that for electric or auditory ERPs. It has been suggested that the initial cortical response evoked by CO 2 laser stimulation of the hand occurs approximately 150 msec after the stimulus onset due to the slow impulse conduction through A 3 fibers. This is estimated as 19.2 +_ 7.2 m / s e c (Adriaensen et al., 1983), and 9 - 1 4 m / s e c based on pain SEPs (Bromm and Treede, 1987; Kakigi et al., 1991). Assuming tbat the time
for the impulse arrival at the cerebral cortex for electric and auditory stimuli is about 20 msec after stimulation, the latency of pain P3 is estimated to be about 130 msec longer than that of electric or auditory P3. According to the present results (Table 1), the latency difference of P3 between pain and eiectric ERPs and between pain and auditory ERPs was 168 msec and 242 msec, respectively, for the count response, and 143 msec and 224 msec, respectively, for the button press response. These findings suggest that, although the difference in conduction time may be an important factor for causing the latency difference, other factors might also play a part. Our experimental design for recording electric and auditory P3 was similar to that ernployed by Barrett et al. (1987), who explained that the latency difference between electric and auditory P3 was due to the possible effects of differences in task difficulty or in the number of channels to which the subject had to pay attention; single channel for auditory stimulation and 2 channels for electric stimulation (2 fingers; one for the target and the other for standard). Furthermore, since in the present study the irradiation site for CO 2 laser stimulation was changed for each stimulus to avoid possible receptor sensitization and habituation, the subject had to pay attention hOt only to the side of stimulation on tbe left hand dorsum but also to the stimulus spot within each side of the hand dorsum. It is therefore most likely that all these effects of task difficulty in the oddball paradigm employing a pain stimulus compared with electric and auditory stimuli explain why the latency of pain P3 is longer than estimated theoretically. There was also a significant difference in the reaction time (RT) between the 3 stimulus modalities (Table 1), which can be largely explained, as in case of P3 latency, by differences in timing of the impulse arrival at the cerebral cortex as well as the effects of task difficulty. The present results also showed that the time difference from P3 to RT (RT-P3) was shortest for the auditory ERP and longest for the pain ERP, with that for tbe electric ERP in between (Table 2). It bas been reported that the association between P3 latency and RT is affected by various factors including evaluation of the stimulus, instructions about speed vs. accuracy and factors of task difficulty (Hillyard and Kutas, 1983; Pfefferbaum et al., 1983; Magliero et al., 1984). Therefore, it is likely that these factors, especially evaluation of the stimulus, affected response processes which produced differences in RT-P3 between the 3 stimulus modalities in the present study. Donchin et al. (1986) pointed out in their review that, while many different processes determine RT, only a subset of these processes detelxnines the P3 latency. I t i s possible that the initial recognition of pain sensation might have actually occurred earlier than the P3 and served as a reference for response processes. Thus, i t i s conceivable that, while N2 and/or P2 may be the reference onset, P3 of the pain ERP is related to categorization processes corresponding to those of auditory P3 and possibly of electric P3 as well.
M. Kanda et al./Electroencephalography and clinical Neurophysiology 100 (1996) 105-114
With regard to the effects of response task, the latency of pain P3 in the count response was not significantly different from that in the button press response. This result is in conformity with previous reports that the latency of P3 is determined by the duration of categorization processes and is relatively independent of the time required for response selection and execution processes (Kutas et al., 1977; McCarthy and Donchin, 1981; Pfefferbaum et al., 1986). According to Johnson (1988), one of the experimental variables that influences P3 amplitude is stimulus meaning in terms of task complexity, stimulus complexity and stimulus value. Although our experimental design differed in stimulus meaning between the 3 stimulus modalities in a manner described above, the amplitude of P3 was not significanfly different between the 3 ERPs. As the amplitude of P3 in the averaged wave form depends entirely on trial-by-trial latency stability (Johnson, 1988), it is likely that poorly synchronized averaging with respect to stimulus onset, which might be associated with longer response latency, might reduce the amplitude of pain P3, and possibly of electric P3 as well. Another possible explanation might be that the P3 amplitude is inversely related to the influence of task difficulty on central processing operations (Polich, 1987). The scalp distribution of P3 was maximum at Pz regardless of the type of response task for ail 3 ERPs, and extended anteriorly to Cz and symmetrically to P3 and P4. Therefore it is suggested that the brain structures participating in cognitive processes are almost the same for the 2 different response tasks and the 3 different stimulus modalities studied in the present experiment. Some investigators have found that the topography of P3 depends on stimulus modality (Barrett et al., 1987; Johnson, 1989a, b), but modality independence has aiso been reported (Snyder et al., 1980; Polich et al., 1991). The latter finding is in conformity with the present result which showed neither significant interaction between modality and electrode site nor a significant main effect of modality on P3 topography (Table 3). In a positron emission tomography (PET) study, Talbot et al. (1991) demonstrated that painful heat caused significant activation in the contralateral primary as well as secondary somatosensory cortices, and anterior cingulate. It is likely that functions of all these cerebral areas serve as candidates for sources of P3-evoking volleys such as those reported for the somatosensory P3 (Bruyant et al., 1993). According to Naumann et al. (1992), a modality-dependent difference in the scalp topography of P3 depends on overlapping by the preceding vertex potential. However, in the present study the latency difference between P2 and P3 of pain ERPs was so long irrespective of the response task (Table 1), that the effects of P2 possibly terminated before development of the P3, and it is unlikely that the P3 is overlapped by P2. Therefore, an experimental design to study pain ERPs such as the present pain P3 may provide a
I 13
future key to elucidate the cortical areas that take part in P3-evoking processes irrespective of the kind of stimulus modality.
Acknowledgements This study was partly supported by Grants-in-Aid for Scientific Research 06404031, for Priority Scientific Research 06260225, and for New Program 06NP0101 from the Japan Ministry of Education, Science and Culture.
References Adriaensen, H., Gybels, J., Handwerker, H.O, and Van Hees, J. Response properties of thin myelinated (A-8) fibres in human skin nerves. J. Neurophysiol., 1983, 49:111-122. Barrett, G., Neshige, R. and Shibasaki, H. Human auditory and somatosensory event-related potentials: effects of oesponse condition and age. Electroenceph. clin. Neurophysiol., 1987, 66: 409-419. Becker, D.E., Yingling, C.D. and Fein, G. Identification of pain, intensity and P300 components in the pain evoked potential. Electroenceph. clin. Neurophysiol., 1993, 88: 290-301. Beydoun, A., Morrow, T.J., Shen, J.F. and Casey, K.L. Variability of laser-evoked potentials: attention, arousal and lateralized differences. Electroenceph. clin. Neurophysiol., 1993, 88:173-181. Bromm, B. and Treede, R.-D. Pain related cerebral potentials: late and ultralate components. Int. J. Neurosci., 1987, 33: 15-23. Bruyant, P., Garcla-Larrea, L. and Mauguière, F. Target side and scalp topography of the somatosensory P300. Electroenceph. clin. Neurophysiol., 1993, 88: 468-477. Desmedt, J.E., Nguyen, T.H. and Bourget, M. The cognitive P40, N60 and PI00 components of somatosensory evoked potentials and the earliest electrical signs of sensory processing in man. Electroenceph. clin. Neurophysiol., 1983, 56: 272-282. Donchin, E., Ritter, W. and McCallum, W.C. Cognitive psychophysiology: the endogenous component of the ERP. In: E. Callaway, P. Tueting and S.H. Koslow (Eds.), Event-Related Brain Potentials in Man. Academic Press, New York, 1978:349-411. Donchin, E., Karis, D., Bashore, T.R., Coles, M.G.H. and Gratton, G. Cognitive psychophysiology and human information processing. In: M.G.H. Coles, E. Donchin and S.W. Porges (Eds.), Psychophysiology: Systems, Processes and Applications. Guilford Press, New York, 1986: 244-267. Garcfa-Larrea, L., Bastuji, H. and Mauguière, F~ Mapping study of the somatosensory evoked potentials during selective spatial attention. Electroenceph. clin. Neurophysiol., 1991, 80: 201-214. Hillyard, S.A. and Kutas, M. Electrophysiology of cognitive processing. Annu. Rev. Psychol., 1983, 34: 33-61. Johnson, Jr., R. The amplitude of the P300 component of the event-related potential: review and synthesis. In: P.K. Ackles, J.R. Jennings and M.G.H. Coles (Eds.), Advances in Psychophysiology, Vol. I1. JAI Press, Greenwich, CT, 1988: 69-138. Johnson, Jr., R. Auditory and visual P300s in temporal lobectomy patient: evidence for modality-dependent generators. Psychophysiology, 1989a, 26: 633-650. Johnson, Jr., R. Developmental evidence for modality-dependent P300 generators: a normative study. Psychophysiology, 1989b, 26:651-667. Kakigi, R., Shibasaki, H. and Ikeda, A. Paiu-related somatosensory evoked potentials following CO 2 laser stimulation in man. Electroenceph. clin. Neurophysiol., 1989, 74: 139-146. Kakigi, R., Endo, C., Neshige, R., Kuroda, Y. and Shibasaki, H. Estimation of conduction velocity of A 6 fibers in humans. Muscle Nerve, 1991, 14: 1193-1196.
114
M. Kanda et al./Electroencephalography
and clinical Neurophysiology 100 (1996) 105-114
Kutas, M., McCarthy, G. and Donchin, E. Augmenting mental chronometry: the P300 as a measure of stimulus evaluation time. Science, 1977, 197: 792-795. Magliero, A., Bashore, T.R., Coles, M.G.H. and Donchin, E. On the dependence of P300 on stimulus evaluation process. Psychophysiology, 1984, 21: 171-186. McCarthy, G. and Donchin, E. A metric for thought: a comparison of P300 latency and reaction rime. Science, 1981, 211: 77-80. McCarthy, G. and Wood, C.C. Scalp distributions of event-related potentials: an ambiguity associated with analysis of variance models. Electroenceph. clin. Neurophysiol., 1985, 62: 203-208. Miyazaki, M., Shibasaki, H., Kanda, M., Xu, X., Shindo, K., Honda, M., Ikeda, A., Nagamine, T., Kaji, R. and Kimura, J. Generator mechanisms of pain-related evoked potentials following CO 2 laser stimulation of the hand: scalp topography and effect of predictive warning signal. J. Clin. Neurophysiol., 1994, 11: 242-254. Naumann, E., Huber, C., Maier, S., Plihal, W., Wustmans, A., Diedrich, O. and Bartussek, D. The scalp topography of P300 in the visual and auditory modalities: a comparison of three normalization methods and the control of statistical type II error. Electroenceph. clin. Neurophysiol., 1992, 83: 254-264. Pfefferbaum, A., Ford, J., Johnson, R., Wenegrat, B. and Kopell, B.S. Manipulation of P3 latency; speed vs accuracy instructions. Electroenceph. clin. Neurophysiol., 1983, 55: 188-197. Pfefferbaum, A., Christensen, C., Ford, J.M. and Kopell, B.S. Apparent response incompatibility effects on P3 latency depend on the task. Electroenceph. clin. Neurophysiol., 1986, 64: 424-437.
Picton, T.W., Stuss, D.T., Champagne, S.C. and Nelson, R.F. The effects of age on human event-related potentials. Psychophysiology, 1984, 21: 312-325. Polich, J. Task difficulty, probability and inter-stimulus interval as determinants of P300 from anditory stimuli. Electroenceph. clin. Neurophysiol., 1987, 68: 311-320. Polich, J., Brock, T. and Geisler, M.W. P300 from auditory and somatosensory stimuli: probability and inter-stimulus interval. Int. J. Psychophysiol., 1991, I1: 219-223. Snyder, E., HiUyard, S.A. and Galambos, R. Similarities and differences among the P3 waves to detected signais in three modalities. Psychophysiology, 1980, 17: 112-122. Talbot, J.D., Marrett, S., Evans, A.C., Meyer, E., Bushnell, M.C. and Duncan, G.H. Multiple representations of pain in human cerebral cortex. Science, 1991, 251: 1355-1358. Towell, A.D. and Boyd, S.G. Sensory and cognitive components of the CO 2 laser evoked cerebral potential. Electroenceph. clin. Neurophysiol., 1993, 88: 237-239. Treede, R.-D., Kief, S., Hölzer, T. and Bromm, B. Late somatosensory evoked cerebral potentials in response to cutaneous heat stimuli. Electroenceph. clin. NeurophysioL, 1988, 70: 429-441. Xu, X., Kanda, M., Shindo, K., Fujiwara, N., Nagamine, T., Ikeda, A., Honda, M., Tachibana, N., Barrett, G., Kaji, R., Kimura, J. and Shibasaki, H. Pain-related somatosensory evoked potentials following CO e laser stimulation of foot in man. Electroenceph. clin, Neurophysiol., 1995, 96: 12-23.
Electroencephalography and clinical Neurophysiology 100 (1996) 105- I 14
Pain-related and cognitive components of somatosensory evoked potentials following CO 2 laser stimulation in man Masutaro Kanda ~,c, Naohito Fujiwara a, Xiaoping Xu ~, Kazuo Shindo ~,b, Takashi Nagamine a, Akio Ikeda a, Hiroshi Shibasaki a,* a Department ofBrain Pathophysiology, Kyoto University School ofMedicine, Shogoin, Sakyo-ku, Kyoto 606-01. Japan b Department ofAnesthesiology, Kyoto University School ofMedicine, Kyoto, Japan c Division ofNeurology, Takeda General Ho.~pital, Kyoto, Japan Accepted for publication: 17 August 1995
Abstract
We recorded cortical potentials evoked by painful CO 2 laser stimulation (pain SEP), employing an oddball paradigm in an effort to demonstrate event-related potentials (ERP) associated with pain. In 12 healthy subjects, frequent (standard) pain stimuli (probability 0.8) were delivered to one side of the dorsum of the left hand while rare (target) pain stimuli (probability 0.2) were delivered to the other side of the same hand. Subjects were instructed to perform either a mental count or button press in response to the target stimulio Two early components (N2 and P2) of the pain SEP demonstrated a Cz maximal distribution, and showed no difference in latency, amplitude or scalp topography between the oddball conditions or between response tasks. In addition, another positive component (P3) following the P2 was recorded maximally at Pz only in response to the target stimuli with a peak latency of 593 msec for the count task and 560 msec for the button press task. lts scalp topography was the saine as that for electric and auditory P3. The longer latency of pain P3 can be explained not only by its slower impulse conduction but also by the effects of task difficulty in the oddball paradigm employing the pain stimulus compared with electric and auditory stimulus paradigms. It is concluded that the P3 for the pain modality is mainly related to a cognitive proeess and corresponds to the P3 of electric and auditory evoked responses, whereas both N2 and P2 are mainly pain-related components. Keywords: CO 2 laser stimulation; Pain SEP; Oddball paradigm; Event-related potential; Pain-related component; Cognitive component; Scalp topography
1. Introduction
Somatosensory evoked potentials (SEPs) following CO 2 laser stimulation are mainly composed of 3 components (NI, N2 and P2) (Miyazaki et al., 1994; Xu et al., 1995). One of the most important factors for clearly recording the initial component (N1) is the intensity of the stimulus (Treede et al., 1988; Xu et al., 1995). Although the NI elicited by hand stimulation is localized to the contralateral central a n d / o r midtemporal area, that in response to foot stimulation is located in the midline central area, suggesting that the N1 is an exogenous component generated in the primary somatosensory cortex (Miyazaki et al., 1994; Xu et ai., 1995).
* Corresponding author. Tel.: + 81 75 7513601 ; Fax: + 81 75 7513202.
With regard to N2 and P2 of the pain SEPs, however, both components are maximal at Cz. The peak latency of P2 is between 300 and 400 msec, which is similar to the latency of the P3 or P300 component of the event-related potential (ERP) recorded in the oddball paradigm using auditory as well as electric stimuli (Barrett et al., 1987). Miyazaki et al. (1994) reported that the amplitude of P2 was significantly larger with a preceding waming signal than without. Furthermore, Beydoun et al. (1993) showed variability in the peak-to-peak amplitude of negative (N2) and positive (P2) components of pain SEPs by modulating various conditions. In particular, amplitude decreased significantly during both a distraction task and drowsiness and was absent during stage II sleep. In the oddball paradigm employing CO 2 laser stimuli, Towell and Boyd (1993) first demonstrated another positive potential following the P2, which appeared only in the
0168-5597/96/$15.00 © 1996 Elsevier Science Ireland Ltd. AIl rights reserved SSDI 00 1 3 - 4 6 9 4 ( 9 5 ) 0 0 1 90-5
EEP 94181
106
M.Kandaetal./Electroencephalography andclinicalNeurophysiology100(1996)105-114
oddball target task. Latency and amplitude of the preceding negative (N2) and positive (P2) peaks in response to target stimuli were not significantly different from those to the standard stimuli. They adopted mental counting for the oddball target task, and the potentials were recorded from only 1 electrode placed at Cz. Therefore, the field distribution of the potential, effects of different response tasks, and differences from other stimulus modalities remained unclear. The aim of the present study, therefore, was to elucidate the characteristics of the endogenous, cognitive component of pain SEPs following CO 2 laser stimulation by employing an oddball paradigm.
2. Methods 2.1. Subjects
Twelve normal subjects (9 men and 3 women; ail right handed) volunteered for this study. Each subject gave written informed consent before the experiment. Their ages ranged from 21 to 39 years (mean 28.2 years), and their height ranged from 158 to 180 cm (mean 167.2 cm). No particular medication was given for the present study. The subject was seated in a reclining armchair in a quiet and electrically shielded room, with the ambient temperature controlled at about 24°C, while their eyes were kept open. The electroencephalogram (EEG) was monitored to check the vigilance level of the subject, and, when it was reduced, as judged by the monitored EEG, the subject was instructed to keep the same level. 2.2. Stimulus 2.2.1. Pain stimulus
Pain stimuli were provided by a special CO 2 laser stimulator (Nippon lnfrared Industries Co. Ltd., Kawasaki, Japan). Its maximum power was 23 W, but an attenuator limited the output to approximately 12.5 W for safety reasons. The laser wave length was 10.6 /xm and the strength of power was 6 W with an irradiation duration of 60 msec, while the irradiation beam was adjusted to about 6 mm in diameter on the skin with the aid of a helium-neon laser (Kakigi et al., 1989). The caiculated stimulus intensity was approximately 12.8 m J / m m ~ which elicited a painful sensation like a "pinprick" in normal subjects. The use of the machine was approved by the Ethics Committee of Kyoto University School of Medicine. Stimuli were presented to either the radial or the ulnar side of the dorsum of the left hand, avoiding the central part (see Section 2.3) (Fig. lA), while the stimulated side was masked so that the subject could not see it. The stimulus spot on each side of the dorsum was slightly shifted for each stimulus in order to avoid sensitization, habituation and receptor fatigue. The interstimulus interval varied at random between 3 and 7 sec.
A. co2 Laser Stirn. Freq. stim. p=o.s
Rate stim. -_. 2
B. E]ectric Stirn. Freq. stim.
Rare stim.
p=0.8~.2
Fig. 1. Schematic diagram illustrating examples of the pain (A) and electdc stimulation (B) for the oddball paradigm. In A pain stimuli for targets (rare) were applied to the radial side of the left hand dorsum at a probability of 0.2, whereas those for standards (freq.) were applied to the ulnar side at a probability of 0.8. In B electric stimuli for targets were applied to the second digit of the left hand at a probability of 0.2, whereas those for standards were applied to the fourth digit at a probability of 0.8. For both modalities, 1 session consisted of 20 frequent (standard) and 5 rare (targe0 stimuli.
2.2.2. Electric stimulus
Percutaneous electric stimuli were presented to either the second or the fourth digit of the left hand with ring electrodes (Fig. lB). The cathode was placed at the proximal interphalangeal joint and the anode at the distal interphalangeal joint of the corresponding digit. The stimulus puise width was 0.2 msec, while its voltage was adjusted to 20% above the sensory threshold, which ranged from 80 to 140 V. The interstimulus intervai for electric stimulation also varied randomly between 3 and 7 sec. 2.2.3. Auditory stimulus
Tone bursts (100 msec duration, 10 msec rise/fall rime, 45 dB SPL above threshold) were presented binaurally through headphones. The frequency of the tone was either 1000 or 2000 Hz (see below). The interstimulus interval for auditory stimulation varied randomly between 1.5 and 2 sec. 2.3. Oddball paradigm
Pain stimuli for target (rare) trials were applied to either the ulnar or the radial side of the left hand dorsum, depending on the session, at a probability of 0.2, whereas those for standard (frequen0 triais were applied to the other side of the same hand at a probability of 0.8 (Fig. lA). Likewise, electric stimuli for target (rare) trials were applied to either the second or the fourth digit of the left hand, depending on the session, at a probability of 0.2, whereas those for standard (frequent) trials were applied to the other digit with a probability of 0.8 (Fig. lB). For auditory stimulation, the 2000 Hz tone was presented binaurally with a probability of 0.2 for target stimuli, whereas the 1000 Hz tone was given at a probability of 0.8 for standard stimuli.
M. Kanda et al./Electroencephalography
and clinical Neurophysiology 100 (1996) 105-114
The response tasks to the target stimuli were the same for each stimulus modality. Subjects were asked to either mentally count the number of infrequently presented target stimuli or to pressa button with the right hand in response to each target promptly and as correctly as possible. The response tasks were varied over sessions. One session consisted of 20 frequent (standard) and 5 rare (target) stimuli, irrespective of either the stimulus modality or the kind of response task. Recordings were performed using auditory, electric and pain stimulus modalities in this order. For each stimulus modality, 4 successive sessions using each of the mental count and button press tasks were given in this order. Consequently every subject was given 24 sessions in all. For the button press task for ail stimulus modalities, reaction time was measured from the onset of the target stimulus to the onset of the button press. The target side and the target digit for the first session for pain and electric stimulation, respectively, were determined randomly and counterbalanced across subjects, and were changed alternatively from session to session.
2.4. Recording In ail subjects, EEGs were recorded with 15 silver/silver chloride shallow cup electrodes (1 cm in diameter) placed on the scalp at Fpz, F3, Fz, F4, T3, C3, Cz, C4, T4, P3, Pz, P4, O1, Oz and 0 2 according to the International 10-20 System, using a Pathfinder with 32 recording channels (Nicolet Biomedical Inc., Wisconsin, USA). Each scalp electrode was referenced to linked earlobes. An electrode placed 1 cm just below the right eye was referenced to another electrode placed just above that eye for monitoring artifacts arising from blinks and eye movements. The ground electrode for pain and auditory stimulation was placed midway between Fpz and Fz, and that for electric stimulation was a stainless steel plate of 3.0 cm × 3.0 cm placed on the skin at the left antero-lateral part of the neck. Ail electrodes were filled with electroconductive jelly, and the impedance was maintained at less than 5 k J2. The bandpass filter was set to 0.05-30 Hz ( - 3 dB) for pain and electric stimulation and 0.05-60 Hz ( - 3 dB) for auditory stimulation.
107
2.6. Data analysis After confirming the reproducibility of wave forms among like sessions with respect to stimulus modality (pain, electric or auditory), oddball condition (target or standard) and response task (mental count or button press) for each subject, group average responses for the 12 modality/condition/response combinations were obtained by averaging data with respect to stimulus onset. The amplitude of each identifiable peak in the group averaged responses was measured from the baseline, which was determined by averaging the first 10% of the epoch for each electrode. For analyzing the latency and amplitude of N2 and P2 components of the pain SEPs, an ANOVA with repeated measurements was performed with factors of oddball condition and response task. For analysis of latency and amplitude of each of the 3 oddball components (P3) obtained using the different stimulus modalities, ANOVAs with repeated measurements were performed with factors of stimulus modality and response task. For evaluating the reaction time for the 3 kinds of stimulus modality, an ANOVA with repeated measurements was performed with stimulus modality as the single factor. Although a multiplicative dipole model is assumed for ERP components, the ANOVA model is based on an additive assumption and can consequently lead to misinterpretation of amplitude distribution (McCarthy and Wood, 1985; Naumann et al., 1992). In an attempt to reduce this problem for the topographic analyses of N2, P2 and P3, amplitudes were normalized by scaling the data relative to the maximum grand mean of the ERP amplitude, before performing ANOVA with the same designs as used for the analysis of the raw amplitude data. The significance level was set at P < 0.01. In addition, the P3 for each stimulus modality for each response task was displayed as topographic maps by using the Nicolet Pathfinder TMAP software, in which purely linear rectangular interpolation between electrode positions was performed on the normalized data.
3. Results
2.5. Averaging One epoch for pain and electric ERPs was 2 sec including a prestimulus period of 200 msec, and that for the auditory ERP was 1 sec including a prestimulus period of 100 msec. The sampling rate for the analog-to-digital conversion was 125 Hz for the former 2 ERPs and 250 Hz for the latter. For each session, responses to 20 frequent stimuli and those to 5 rare (target) stimuli were averaged separately time-locked to stimulus onset.
In the count task, it is impossible to detect target responses in which subjects failed to perform the task correctly. However, as the number of target stimuli reported by the subject after each session was correct, it was considered that there was no inaccurate performance. In the button press task, there were only a few target trials with incorrect responses, and these were excluded from the averaged response for each stimulus modality as were trials contaminated by artifact.
108
M. Kanda et al. / Electroencephalography and clinical Neurophysiolo gy 100 (1996) 105-114
Frequent
Rare
Button press
Count EOG"
EOG'
Count
PAIN
Fz
N2
Cz
P2
Cz
P3
PZ i
10pV +l
P2
i i i i , k l l l
, stim.
I 1800 mec
i
P3 i
i
i
i
i
[
I
1800 ms
stim.
I'''''',LI stim.
i
i
i
i
i
i
i
i
i
I
~' stirn,
1800 ms
1800 fl~ec
EOG -,
Fz
.
- ......
ELECTRIC E0G
Button F z press cz I
P3
P3 I
laoo stim,
PZ 10pV +l
I' ' stil.
''
'
I'
'''1 1800
mec
stim.
' ' ' ' ' '
'
t
3.1. W a v e f o r m s
I n the averaged wave forrns of pain SEPs o b t a i n e d in the oddball p a r a d i g m u s i n g C O 2 laser stimulation, early negative-positive potentials (N2-P2) were recorded regardless o f oddball conditions (frequent or rare) and response tasks (mental count or b u t t o n press) (Fig. 2). F o l l o w i n g these potentials, a positive deflection at a latency o f 572 msec for the count response and 508 msec for the b u t t o n press response was observed at Pz and Cz o n l y in the rare (target) conditions. In the grand average w a v e forms of pain SEPs across ail subjects obtained in the oddball p a r a d i g m (Fig. 3), a negative c o m p o n e n t was followed b y a positivity corres p o n d i n g to N 2 and P2, respectively, in both sets of w a v e forms. The latency of N 2 as well as that of P2 was similar b e t w e e n the oddball conditions and b e t w e e n the response tasks. Both N 2 and P2 were m a x i m a l at Cz in ail of the 4 sets o f c o m b i n a t i o n s (oddball c o n d i t i o n and response task) (Fig. 4). A n o t h e r positive c o m p o n e n t was seen after the P2, exclusively in the wave forms of target responses, which was considered to be an oddball c o m p o n e n t and was n a m e d P3 (Fig. 3). The peak latency of P3 was 572 msec
L
L
i
~
i
i
i
[
~soo
stim.
ms
EOG"
1800 msec
Fig. 2. Averaged wave forms of SEPs obtained in an oddball paradigm using CO z laser stimuli in a single subject. Each trace of superimposed wave forms is the result of averaging 2 like sessions in which either the ulnar or the radial side of the left hand dorsum was stimulated as oddball target. In this and the following figures, 3 selected EEG channels (Fz, Cz and Pz) and the electrooculogram (EOG) are shown. "Stim" denotes the rime of stimulus onset. Relative negativity in the exploring electrode is shown upward in this and the following figures. A positive potential (arrow) is seen only in the rare (target) conditions, at a latency of 572 msec for the count response and 508 msec for the button press response.
ms
i
ç
Fz
AUD. Cz Pz
,o.v~i
P3 I
i
i
h
i
P3 i
h
i
i
I
900 stim.
ms
.
.
.
.
.
.
i
L
i
I
900
stim,
ms
Fig. 3. Grand average wave forms obtained in the oddball paradigm across ail subjects. Solid lines show responses to the rare (target) stimuli, whereas interrupted lines show those to the frequent (standard) stimuli. In pain SEPs, a negative peak (N2) and the following positive peak (P2) are observed in the target as well as standard conditions for both response tasks (count and button press). Another positive potential (pain P3) can be seen following P2 only in the target condition. Electric P3 and auditory P3 also appear only in the target condition for both response tasks.
for the count response and 508 msec for the b u t t o n press response. The amplitude of P3 was m a x i m a l at Pz regardless of the response task (Fig. 5).
RARE
FREQUENT
~~
N~
-1
-1 Fz
Cz ELECTRODE
Pz
Fz
Cz
Pz
ELECTRODE
Fig. 4. Amplitudes of N2 and P2 of pain SEP at 3 electrodes (Fz, Cz and Pz). In each figure, amplitudes for the count response are plotted with solid lines, whereas those for the button press response are plotted with interrupted lines. Both N2 and P2 are maximal at the Cz electrode, regardless of the oddball condition or response task.
M. Kanda et al./Electroencephalography and clinical Neurophysiolo gy 100 (1996) 105-114
In the grand average wave forms of SEPs obtained in the oddball paradigm using electric stimulation (Fig. 3), a large positive component appeared maximally at Pz only in response to the target stimuli for both response tasks. This was called the electric oddball P3, The peak latency of this P3 was 428 msec for the count response and 404 msec for the button press response, and its amplitude was maximal at Pz regardless of the kind of response task (Fig. 5). Grand average wave forms of auditory evoked potentials (AEPs) across ail subjects obtained in the oddball paradigm showed a large positive comportent corresponding to the conventional P300 or P3 which appeared only in the target condition for both response tasks (Fig. 3). The peak latency of P3 was 352 msec for the count response and 364 msec for the button press response, and its amplitude was also maximal at Pz irrespective of the kind of response task (Fig. 5).
109
12> =-
/
COUNT BUTTON --O.-- ..-o-ELECTRIC --42-- ---oe-AUD. -..,b-- ...~..
.. ..-" .,~ ..." . , ~
PAIN
10-
III
oO~
~
"°
~,.....I
8-
..e
~
........•
13.
< 13.
6-
4
i Fz
i
[
Cz
Pz
ELECTRODE Fig. 5. P3 amplitude of 3 different kinds of ERPs at 3 electrodes (Fz, Cz and Pz). P3 is maximal at the Pz electrode regardless of the stimulus modality or response task.
Table 1 Mean latency and amplitude of N2 and P2 of pain SEPs and P3 components of 3 different kinds of evoked potentials obtained in the oddball paradigm in 12 subjects (mean + S.D.) Comp/RT
Electrode
Mod
Task
Cond
Mean dz S.D.
Cz Cz Cz Cz Cz Cz Cz Cz Pz Pz Pz Pz Pz Pz
Pz Pz Pz
p p p p p p p p p p e e a a p e a p e a
c c b b c c b b c b c b c b b b b b b b
freq rare freq rare freq rare freq rare rare rare rare rare rare rare rare rare rare rare rare rare
211 214 218 217 318 315 323 318 593 560 425 417 351 336 668 467 372 108 49 19
Cz Cz Cz Cz Cz Cz Cz Cz Pz Pz Pz Pz Pz Pz
p p p p p p p p p p e e a a
c c b b c c b b c b c b c b
freq rare freq rare freq rare freq rare rare rare rare rare rare rare
- 5.3 ___ -6.2+ - 6.6 i -8.0+ 8.1 + 10.8 + 8.7 _ 9.0_+ 10.6+ 10.5 + 11.2 dz 12.2 dz 11.7 + 13.4_+
Latency (msec) N2
P2
P3
RT
RT-P3
_+ 22 _+ 23 dz 19 dz 18 dz 22 dz 26 + 21 dz 29 + 3l + 54 dz 69 dz 58 dz 31 dz 38 dz 91 dz 93 dz 90 dz 94 _+105 _+ 82
Amplitude ( btV) N2
P2
P3
3.1 4.3 3.9 5.0 1.8 4.3 3.4 3.0 3.8 1.8 3.7 5.5 4.8 5.9
Comp, component; RT, reaction time; RT-P3, time difference from the P3 peak to RT; Mod, stimulus modality; p, e and a, pain, electric and auditory, respectively; c and b, count and button press response task, respectively; Cond, oddball conditions; freq and rare, frequent (standard) and rare (target) condition, respectively.
M. Kanda et al./ Electroencephalography and clinical Neurophysiology 100 (1996) 105-114
110
Table 2 A N O V A s for latency and amplitude of N2, P2 and P3, and for R T and RT-P3 Factor ( d f )
Latency N2
Response time P2
P3
RT
Amplitude RT-P3
N2
P2
P3
c(1, ll) t ( l , 11) c t ( l , 11) s (2, 22) t ( 1 , 11) st (2, 22) s (2, 22) N2 and P2: measured at Cz; P3: measured at Pz. Abbreviations of A N O V A terms: c = oddball condition; s = stimulus modality; t = response task. Significance values are calculated using Greenhouse-Geisser correction: - P > 0.01; * P < 0.01; * * P < 0.005; * * " P < 0.001.
Table 3 A N O V A s for scalp topography of N2, P2 and P3 with application of m a x i m u m normalization Condition
Modality
Factor (df)
N2
c(1, ll) t(l, Il) e ( 1 4 , 154) c t ( l , 11) et (14, 154) ce (14, 154) cet ( 14, 154)
. . . . . . -
P2
-
Factor (df)
where its amplitude was maximal among the 15 electrodes. An ANOVA on normalized P2 also showed no effect for the ANOVA terms described above, apart from electrode site (e) (Table 3).
P3
s ( 2 , 11) t ( 1 , 11) e ( 1 4 , 154) s t ( 2 , 22) et (14, 154) es (28, 308) est ( 2 8 , 3 0 8 )
Abbreviations of A N O V A terms: c = oddball condition; e = electrode; s = stimulus modality; t = response task. Significance values are obtained using Greenhouse-Geisser correction: P > 0.01; *** P < 0 . 0 0 1 .
3.2. N2 and P2 of pain SEPs 3.2.1. Latency and amplitude : Mean values of latency and amplitude for N2 and P2 of pain SEPs obtained in the oddball paradigm and measured at Cz for each combination of oddball condition and response task are shown in Table 1, and Table 2 shows the ANOVA results. For the 3 terms for ANOVA involving oddball condition (c), response task (t) and condition x response task interaction (ct), neither N2 nor P2 was significantly different in latency (Table 2). Similarly, the raw amplitude of the 2 components did not show statistically significant differences for the ANOVA terres. 3.2.2. Scalp topography An ANOVA on the N2 amplitudes normalized by scaling as a percentage of the Cz amplitude showed no effect of oddball condition (c), response task (t), condition X response task interaction (ct), electrode site X response task interaction (et), or condition X electrode site × response task interaction (cet), and only electrode site (e) showed significant differences (Table 3). Similarly, P2 was normalized as a percentage of the value at the Cz electrode
3.3. Oddball component (P3) 3.3.1. Latency and amplitude Mean values of latency and amplitude of the oddball component (P3) evoked in the 3 different tests are shown in Table 1. The iatency of P3 was significantly different between stimulus modalities (s), although it showed no difference between the response tasks (t) or for the stimulus modality x response task interaction (st) (Table 2). An ANOVA on the raw amplitude of P3 did not show differences between the stimulus modalities (s), response tasks (t) and stimulus modality X response task interaction (st) (Table 2). ANOVAs on reaction time (RT) obtained from the button press responses in the target conditions and the time difference between the corresponding P3 peak and the reaction time (RT-P3) showed significant differences between stimulus modalities (s) (Table 2). 3.3.2. Scalp topography An ANOVA on the normalized P3 amplitude (scaled with respect to amplitude at Pz) showed no effect of stimulus modality (s), response task (t), stimulus modality X response task interaction (st), electrode site X response task interaction (et), electrode site x stimulus modality interaction (es) and electrode site X stimulus modality X response task interaction (est), although electrode site (e) showed a significant difference (Table 3). Fig. 6 demonstrates topographic maps of the oddball component P3, comparing pain, electric and auditory modalities for each of the count and button press responses. In ail 6 maps, the P3 component was maximal at Pz, extended symmetrically to the bilateral parietal electrodes. The field distribution of each P3 on the left hemisphere was similar to the corresponding one on the right, and, therefore, further statistical
M. Kanda et a l . / Electroencephalography and clinical Neurophysiology 100 (1996) 105-114
111
P3 Count
Button press
Pain
Electric
Auditory
-1.0 ~
1.0
Fig. 6. Scaip topography of the oddball component P3 of pain, electric and auditory ERPs ail obtained in the oddball paradigm for count and button press responses. Note similar distribution pattem in all maps, which is maximal at Pz and extends symmetrically fo the bilateral parietal electrodes.
analyses comparing interhemispheric differences were not performed. 4. Discussion
The wave form of pain SEPs obtained in the present study demonstmted N2 and P2 components, both of which were present regardless of the oddball condition (rare, target or frequent, standard) and of the kind of response task (mental count or button press). Miyazaki et al. (1994) reponed that P2 increased in amplitude by applying a predictive waming signal, whereas N2 did not. Beydoun et al. (1993) reponed that the amplitude difference between
negative and positive peaks of CO 2 laser SEPs, which correspond to N2 and P2, respectively, diminished by distraction and during lowered arousal state. Moreover, even early cortical components of the electric SEPs have been suggested to be modulated by attention (Desmedt et al., 1983; Garda-Larrea et al., 1991). In the present study, however, neither the peak latency nor the amplitude of N2 and P2 was significantly different according to the oddball condition or response task and there was no condition × response interaction. This observation also confirmed and extended the finding by Towell and Boyd (1993) which used the count task and showed non-significant changes in the peak-to-peak amplitude (N2-P2) between the oddbali
112
M. KarMa et aL / Electroencephalography and clinieal Neurophysiology I00 (1996) 105-114
conditions. Furthermore, we found that the scãlp distribution of N2 and P2 was not different between the response tasks or between the oddball conditions. These findings suggest that the oddball type of manipulation does not significantly alter either N2 or P2. In addition to N2 and P2, another positive component (P3) following the P2 was recognized exclusively in the oddball target conditions for both count and button press responses, with a mean peak latency of 593 msec for the count response and 560 msec for the button press response, and a mean amplitude of 10.6/zV for the former and 10.5 /zV for the latter. The late positive component reported by Towell and Boyd (1993) had a mean peak latency of 621 msec and a peak-to-peak amplitude (N2-P3) of 22.5 /zV. Although these values were slightly different from those of the P3 observed in the present study, this component is considered to be a cognitive component because it appeared exclusively in the oddball target condition irrespective of the response tasks. While they used only 1 electrode placed at Cz (Towell and Boyd, 1993), we used 15 electrodes to delineate scalp distribution of the potential. As a result, we showed that the P3 potential was maximal at Pz, in contrast with the P2 which was maximal at Cz, suggesting different generator sources for these 2 positive components. Becker et al. (1993), by studying SEPs produced by intracutaneous electrical stimulation of a fingertip, reported that the pain and intensity components were larger at Cz than at Pz, and vice versa for the cognitive component. According to Donchin et al. (1978), endogenous components of ERPs for any stimulus modality are characterized by: (1) the component must be non-obligatory responses to stimuli; (2) the amplitude, latency, and scalp distribution of the potential are often invariant to changes in the physical parameters of the eliciting stimulus; (3) the variance of the endogenous component is normally accounted for by variation in the tasks assigned to the subject. Therefore, it is conceivable that, although P2 might be mainly related to pain perception depending on physical parameters of the pain stimulus a n d / o r the subject's attention directed to it, P3 is an endogenous component related to the "cognition," which means "categorization processes" through which a subject makes a decision to perform the task in an oddball paradigm. There was a significant difference in the P3 latency between the 3 stimulus modalities (pain, electric and auditory) (Table 2). The P3 latency of the electric ERP was longer than that of the auditory ERP, confirming previous reports (Pieton et al., 1984; Barrett et al., 1987). P3 latency of the pain ERP was much longer than that for electric or auditory ERPs. It has been suggested that the initial cortical response evoked by CO 2 laser stimulation of the hand occurs approximately 150 msec after the stimulus onset due to the slow impulse conduction through A 3 fibers. This is estimated as 19.2 +_ 7.2 m / s e c (Adriaensen et al., 1983), and 9 - 1 4 m / s e c based on pain SEPs (Bromm and Treede, 1987; Kakigi et al., 1991). Assuming tbat the time
for the impulse arrival at the cerebral cortex for electric and auditory stimuli is about 20 msec after stimulation, the latency of pain P3 is estimated to be about 130 msec longer than that of electric or auditory P3. According to the present results (Table 1), the latency difference of P3 between pain and eiectric ERPs and between pain and auditory ERPs was 168 msec and 242 msec, respectively, for the count response, and 143 msec and 224 msec, respectively, for the button press response. These findings suggest that, although the difference in conduction time may be an important factor for causing the latency difference, other factors might also play a part. Our experimental design for recording electric and auditory P3 was similar to that ernployed by Barrett et al. (1987), who explained that the latency difference between electric and auditory P3 was due to the possible effects of differences in task difficulty or in the number of channels to which the subject had to pay attention; single channel for auditory stimulation and 2 channels for electric stimulation (2 fingers; one for the target and the other for standard). Furthermore, since in the present study the irradiation site for CO 2 laser stimulation was changed for each stimulus to avoid possible receptor sensitization and habituation, the subject had to pay attention hOt only to the side of stimulation on tbe left hand dorsum but also to the stimulus spot within each side of the hand dorsum. It is therefore most likely that all these effects of task difficulty in the oddball paradigm employing a pain stimulus compared with electric and auditory stimuli explain why the latency of pain P3 is longer than estimated theoretically. There was also a significant difference in the reaction time (RT) between the 3 stimulus modalities (Table 1), which can be largely explained, as in case of P3 latency, by differences in timing of the impulse arrival at the cerebral cortex as well as the effects of task difficulty. The present results also showed that the time difference from P3 to RT (RT-P3) was shortest for the auditory ERP and longest for the pain ERP, with that for tbe electric ERP in between (Table 2). It bas been reported that the association between P3 latency and RT is affected by various factors including evaluation of the stimulus, instructions about speed vs. accuracy and factors of task difficulty (Hillyard and Kutas, 1983; Pfefferbaum et al., 1983; Magliero et al., 1984). Therefore, it is likely that these factors, especially evaluation of the stimulus, affected response processes which produced differences in RT-P3 between the 3 stimulus modalities in the present study. Donchin et al. (1986) pointed out in their review that, while many different processes determine RT, only a subset of these processes detelxnines the P3 latency. I t i s possible that the initial recognition of pain sensation might have actually occurred earlier than the P3 and served as a reference for response processes. Thus, i t i s conceivable that, while N2 and/or P2 may be the reference onset, P3 of the pain ERP is related to categorization processes corresponding to those of auditory P3 and possibly of electric P3 as well.
M. Kanda et al./Electroencephalography and clinical Neurophysiology 100 (1996) 105-114
With regard to the effects of response task, the latency of pain P3 in the count response was not significantly different from that in the button press response. This result is in conformity with previous reports that the latency of P3 is determined by the duration of categorization processes and is relatively independent of the time required for response selection and execution processes (Kutas et al., 1977; McCarthy and Donchin, 1981; Pfefferbaum et al., 1986). According to Johnson (1988), one of the experimental variables that influences P3 amplitude is stimulus meaning in terms of task complexity, stimulus complexity and stimulus value. Although our experimental design differed in stimulus meaning between the 3 stimulus modalities in a manner described above, the amplitude of P3 was not significanfly different between the 3 ERPs. As the amplitude of P3 in the averaged wave form depends entirely on trial-by-trial latency stability (Johnson, 1988), it is likely that poorly synchronized averaging with respect to stimulus onset, which might be associated with longer response latency, might reduce the amplitude of pain P3, and possibly of electric P3 as well. Another possible explanation might be that the P3 amplitude is inversely related to the influence of task difficulty on central processing operations (Polich, 1987). The scalp distribution of P3 was maximum at Pz regardless of the type of response task for ail 3 ERPs, and extended anteriorly to Cz and symmetrically to P3 and P4. Therefore it is suggested that the brain structures participating in cognitive processes are almost the same for the 2 different response tasks and the 3 different stimulus modalities studied in the present experiment. Some investigators have found that the topography of P3 depends on stimulus modality (Barrett et al., 1987; Johnson, 1989a, b), but modality independence has aiso been reported (Snyder et al., 1980; Polich et al., 1991). The latter finding is in conformity with the present result which showed neither significant interaction between modality and electrode site nor a significant main effect of modality on P3 topography (Table 3). In a positron emission tomography (PET) study, Talbot et al. (1991) demonstrated that painful heat caused significant activation in the contralateral primary as well as secondary somatosensory cortices, and anterior cingulate. It is likely that functions of all these cerebral areas serve as candidates for sources of P3-evoking volleys such as those reported for the somatosensory P3 (Bruyant et al., 1993). According to Naumann et al. (1992), a modality-dependent difference in the scalp topography of P3 depends on overlapping by the preceding vertex potential. However, in the present study the latency difference between P2 and P3 of pain ERPs was so long irrespective of the response task (Table 1), that the effects of P2 possibly terminated before development of the P3, and it is unlikely that the P3 is overlapped by P2. Therefore, an experimental design to study pain ERPs such as the present pain P3 may provide a
I 13
future key to elucidate the cortical areas that take part in P3-evoking processes irrespective of the kind of stimulus modality.
Acknowledgements This study was partly supported by Grants-in-Aid for Scientific Research 06404031, for Priority Scientific Research 06260225, and for New Program 06NP0101 from the Japan Ministry of Education, Science and Culture.
References Adriaensen, H., Gybels, J., Handwerker, H.O, and Van Hees, J. Response properties of thin myelinated (A-8) fibres in human skin nerves. J. Neurophysiol., 1983, 49:111-122. Barrett, G., Neshige, R. and Shibasaki, H. Human auditory and somatosensory event-related potentials: effects of oesponse condition and age. Electroenceph. clin. Neurophysiol., 1987, 66: 409-419. Becker, D.E., Yingling, C.D. and Fein, G. Identification of pain, intensity and P300 components in the pain evoked potential. Electroenceph. clin. Neurophysiol., 1993, 88: 290-301. Beydoun, A., Morrow, T.J., Shen, J.F. and Casey, K.L. Variability of laser-evoked potentials: attention, arousal and lateralized differences. Electroenceph. clin. Neurophysiol., 1993, 88:173-181. Bromm, B. and Treede, R.-D. Pain related cerebral potentials: late and ultralate components. Int. J. Neurosci., 1987, 33: 15-23. Bruyant, P., Garcla-Larrea, L. and Mauguière, F. Target side and scalp topography of the somatosensory P300. Electroenceph. clin. Neurophysiol., 1993, 88: 468-477. Desmedt, J.E., Nguyen, T.H. and Bourget, M. The cognitive P40, N60 and PI00 components of somatosensory evoked potentials and the earliest electrical signs of sensory processing in man. Electroenceph. clin. Neurophysiol., 1983, 56: 272-282. Donchin, E., Ritter, W. and McCallum, W.C. Cognitive psychophysiology: the endogenous component of the ERP. In: E. Callaway, P. Tueting and S.H. Koslow (Eds.), Event-Related Brain Potentials in Man. Academic Press, New York, 1978:349-411. Donchin, E., Karis, D., Bashore, T.R., Coles, M.G.H. and Gratton, G. Cognitive psychophysiology and human information processing. In: M.G.H. Coles, E. Donchin and S.W. Porges (Eds.), Psychophysiology: Systems, Processes and Applications. Guilford Press, New York, 1986: 244-267. Garcfa-Larrea, L., Bastuji, H. and Mauguière, F~ Mapping study of the somatosensory evoked potentials during selective spatial attention. Electroenceph. clin. Neurophysiol., 1991, 80: 201-214. Hillyard, S.A. and Kutas, M. Electrophysiology of cognitive processing. Annu. Rev. Psychol., 1983, 34: 33-61. Johnson, Jr., R. The amplitude of the P300 component of the event-related potential: review and synthesis. In: P.K. Ackles, J.R. Jennings and M.G.H. Coles (Eds.), Advances in Psychophysiology, Vol. I1. JAI Press, Greenwich, CT, 1988: 69-138. Johnson, Jr., R. Auditory and visual P300s in temporal lobectomy patient: evidence for modality-dependent generators. Psychophysiology, 1989a, 26: 633-650. Johnson, Jr., R. Developmental evidence for modality-dependent P300 generators: a normative study. Psychophysiology, 1989b, 26:651-667. Kakigi, R., Shibasaki, H. and Ikeda, A. Paiu-related somatosensory evoked potentials following CO 2 laser stimulation in man. Electroenceph. clin. Neurophysiol., 1989, 74: 139-146. Kakigi, R., Endo, C., Neshige, R., Kuroda, Y. and Shibasaki, H. Estimation of conduction velocity of A 6 fibers in humans. Muscle Nerve, 1991, 14: 1193-1196.
114
M. Kanda et al./Electroencephalography
and clinical Neurophysiology 100 (1996) 105-114
Kutas, M., McCarthy, G. and Donchin, E. Augmenting mental chronometry: the P300 as a measure of stimulus evaluation time. Science, 1977, 197: 792-795. Magliero, A., Bashore, T.R., Coles, M.G.H. and Donchin, E. On the dependence of P300 on stimulus evaluation process. Psychophysiology, 1984, 21: 171-186. McCarthy, G. and Donchin, E. A metric for thought: a comparison of P300 latency and reaction rime. Science, 1981, 211: 77-80. McCarthy, G. and Wood, C.C. Scalp distributions of event-related potentials: an ambiguity associated with analysis of variance models. Electroenceph. clin. Neurophysiol., 1985, 62: 203-208. Miyazaki, M., Shibasaki, H., Kanda, M., Xu, X., Shindo, K., Honda, M., Ikeda, A., Nagamine, T., Kaji, R. and Kimura, J. Generator mechanisms of pain-related evoked potentials following CO 2 laser stimulation of the hand: scalp topography and effect of predictive warning signal. J. Clin. Neurophysiol., 1994, 11: 242-254. Naumann, E., Huber, C., Maier, S., Plihal, W., Wustmans, A., Diedrich, O. and Bartussek, D. The scalp topography of P300 in the visual and auditory modalities: a comparison of three normalization methods and the control of statistical type II error. Electroenceph. clin. Neurophysiol., 1992, 83: 254-264. Pfefferbaum, A., Ford, J., Johnson, R., Wenegrat, B. and Kopell, B.S. Manipulation of P3 latency; speed vs accuracy instructions. Electroenceph. clin. Neurophysiol., 1983, 55: 188-197. Pfefferbaum, A., Christensen, C., Ford, J.M. and Kopell, B.S. Apparent response incompatibility effects on P3 latency depend on the task. Electroenceph. clin. Neurophysiol., 1986, 64: 424-437.
Picton, T.W., Stuss, D.T., Champagne, S.C. and Nelson, R.F. The effects of age on human event-related potentials. Psychophysiology, 1984, 21: 312-325. Polich, J. Task difficulty, probability and inter-stimulus interval as determinants of P300 from anditory stimuli. Electroenceph. clin. Neurophysiol., 1987, 68: 311-320. Polich, J., Brock, T. and Geisler, M.W. P300 from auditory and somatosensory stimuli: probability and inter-stimulus interval. Int. J. Psychophysiol., 1991, I1: 219-223. Snyder, E., HiUyard, S.A. and Galambos, R. Similarities and differences among the P3 waves to detected signais in three modalities. Psychophysiology, 1980, 17: 112-122. Talbot, J.D., Marrett, S., Evans, A.C., Meyer, E., Bushnell, M.C. and Duncan, G.H. Multiple representations of pain in human cerebral cortex. Science, 1991, 251: 1355-1358. Towell, A.D. and Boyd, S.G. Sensory and cognitive components of the CO 2 laser evoked cerebral potential. Electroenceph. clin. Neurophysiol., 1993, 88: 237-239. Treede, R.-D., Kief, S., Hölzer, T. and Bromm, B. Late somatosensory evoked cerebral potentials in response to cutaneous heat stimuli. Electroenceph. clin. NeurophysioL, 1988, 70: 429-441. Xu, X., Kanda, M., Shindo, K., Fujiwara, N., Nagamine, T., Ikeda, A., Honda, M., Tachibana, N., Barrett, G., Kaji, R., Kimura, J. and Shibasaki, H. Pain-related somatosensory evoked potentials following CO e laser stimulation of foot in man. Electroenceph. clin, Neurophysiol., 1995, 96: 12-23.