Optimizing the Measurement of Contact Heat Evoked Potentials by Andrew Bagshaw | Papers by Andrew

ORIGINAL RESEARCH ARTICLE Optimizing the Measurement of Contact Heat Evoked Potentials Tracy Warbrick,*† Stuart W. G. Derbyshire,* and Andrew P. Bagshaw* Abstract: Our aim was to determine whether single trial averaging could improve quantification of contact heat evoked potentials measured from fixed position contact heat stimulation. Event-related brain potentials were measured in response to contact heat stimuli applied to the arm and the leg of 10 subjects via a circular thermode, using fixed and varied thermode positions at 41°C and 51°C. Contact heat evoked potentials were successfully recorded from varied position stimulation of the leg at 51°C in 80% of subjects, but from only 60% of subjects using a fixed position. Contact heat evoked potentials were only identified in a small number of subjects when stimulating at 41°C. The amplitude of the N2-P2 complex and pain intensity ratings were larger for the varied compared with the fixed thermode position and were also larger when stimulating the arm. Automated single trial analysis of the data resulted in larger amplitude of the N2-P2 complex than standard averaging. In conclusion, we have demonstrated that the reduced contact heat evoked potentials amplitude seen for fixed location stimulation can be improved with single-trial averaging. Key Words: EEG, fMRI, Pain, Human, Brain. (J Clin Neurophysiol 2009;26: 117–122) T he generation of pain event-related brain potentials (ERPs) requires accurate onset of nociceptor activation and thus a strong, brief, phasic stimulus (Bromm, 1984). Currently, laser evoked potentials (LEPs) are the most reliable and commonly used method for investigating pain ERPs (Treede, 2003). However, new contact heat stimulation technology can deliver appropriate stimuli to generate event related cortical responses (Chen et al., 2001; Valeriani et al., 2002). Consequently, measuring Contact Heat Evoked Potentials (CHEPs) in pain research is becoming increasingly popular (Chen et al., 2002; Granovsky et al., 2005; Greffrath et al., 2007; Iannetti et al., 2006). However, CHEPs tend to have longer latencies and smaller amplitudes than LEPs (Iannetti et al., 2006) and are not consistently evoked in all subjects at lower intensities (Chen et al., 2006). Stimulus duration has been shown to influence the latency of pain ERP components (Iannetti et al., 2004) and the longer latency of CHEPs is most likely due to the fact that contact heat stimuli have a longer rise time than laser stimuli. In addition, higher latency jitter in single trials is observed in CHEPs compared with LEPs, which contributes to the reduced amplitude seen in CHEPs (Iannetti et al., 2006). This reduced reliability of CHEPs recording needs further investigation to maximize the use of the contact heat stimulator in pain research. One possible way to optimize the use of CHEPs is to investigate appropriate analysis strategies. The potential problems with CHEPS, such as latency jitter and reduced amplitude, are increased by standard averaging procedures that can decrease amplitude, cause waveform distortion and lead to inaccurate peak latency estimation (Iannetti et al., 2006). A novel strategy for EP analysis involving automated measurement of single trials aims to disclose the maximum biologic information from a single trial response, thus avoiding the negative consequences of standard averaging (Mayhew et al., 2006). The method is potentially useful when CHEPs are recorded under circumstances where the quality of measured responses is compromised. For example, when experimental constraints require fixed location stimulation or when stimulation of the lower limbs is necessary. Both fixed location stimulation and stimulation of the leg result in CHEPs with reduced amplitudes when compared with varied location stimulation (Greffrath et al., 2007) or stimulation of other sites (Chen et al., 2006; Granovsky et al., 2006). However, suboptimal stimulation methods are sometimes necessary, for example, when stimulation of multiple sites is required. Furthermore, if contact heat stimulation is to be used as an alternative to laser stimulation its potential for use during simultaneous recording of ERP and MRI data should be considered. Simultaneous EEG-fMRI recordings should be possible owing to the MR compatibility of the Contact Heat Evoked Potential Stimulator (CHEPS, Medoc Ltd, Ramat Yishai, Israel). Practically, however, access to the arm can be limited in some MRI scanners, potentially making varied stimulation locations difficult. Thus, fixed location stimulation of the arm or stimulation of the leg would be necessary. Given that fixed position stimulation of the arm reduces CHEP amplitude and stimulation of the leg is less reliable for eliciting CHEPs (Chen et al., 2006), identifying methods for maximizing the measured responses is essential. This has implications for using the contact heat stimulator in combined ERP and fMRI pain research, and for research/clinical practice requiring multiple site stimulation. It is possible that by adopting an appropriate analysis strategy, such as the single trial averaging described above, responses at stimulation sites that generally yield poor data can be optimized. Thus, the purpose of this study was to determine whether single trial analysis of CHEPs could improve the data obtained from fixed location stimulation and stimulation of the leg in comparison with standard averaging. METHOD Ten volunteers (2 male), with a mean age of 23.1 years ( 4.7 years) participated in the study. Written informed consent was obtained from all subjects and the protocol was approved by the local Research Ethics Board. A Contact Heat Evoked Potential Stimulator (CHEPS, Medoc Ltd., Ramat Yishai, Israel) was used to generate heat stimuli. The Contact heat stimulator consists of a circular thermode with a diameter of 27 mm. The thermode comprises a heating thermoil (Minco Products, Inc., Minneapolis, MN) covered with a 25 m layer of thermoconductive plastic (Kapton, thermal conductivity at 23°C of 0.1– 0.35 W/m/K). Two thermocouples are embedded 10 m within this conductive coating, which contacts the skin directly, thus giving an esti117 From the *School of Psychology and Birmingham University Imaging Centre (BUIC), University of Birmingham, Birmingham, United Kingdom; and †Department of Psychiatry, Neuropsychiatric Research Laboratory, HeinrichHeine University of Dusseldorf, Dusseldorf, Germany. Address correspondence and reprint requests to Tracy Warbrick, Ph.D., Rheinische Kliniken Dusseldorf, Kliniken der Heinrich-Heine Universitat Dussel¨ ¨ ¨ dorf, Neuropsychiatrisches Forschungslabor, Bergische Landstrasse 2, 40629 Dusseldorf, Germany; e-mail: t.warbrick@fz-juelich.de. ¨ Copyright © 2009 by the American Clinical Neurophysiology Society ISSN: 0736-0258/09/2602-0117 Journal of Clinical Neurophysiology • Volume 26, Number 2, April 2009 T. Warbrick et al. Journal of Clinical Neurophysiology • Volume 26, Number 2, April 2009 mate of skin temperature at the thermode surface. The thermofoil permits a heating rate of 70°C/s and the Peltier device permits a cooling rate of 70°C/s. In the current study, cooling began immediately after the attainment of the target heat pulse temperature. Contact heat stimuli were delivered at 41°C and 51°C from a baseline of 32°C, and were delivered to the right volar forearm and the right lower leg. Two stimulation methods were used: fixed position of the thermode and variable position of the thermode. In fixed position conditions the thermode was attached to the subject and remained in the same place for the duration of the block. During the variable position conditions the thermode position was rotated between one of three positions 5 cm apart after every trial within that block. This 2 2 2 (temperature stimulation method limb) design resulted in eight conditions, the order of which was counterbalanced across subjects. One block of 30 trials was delivered for each condition, giving a total of 240 trials. A variable interstimulus interval of 8 to 12 seconds was used. Subjects were asked to provide a pain intensity rating after each stimulus using the Gracely intensity scale (Gracely, 1992). This is a 20 point scale anchored with “no sensation” at zero and containing verbal descriptors ranging from “very weak” (rating of 3) to “extremely intense” (rating of 18). A mean pain rating was then calculated for each condition for each subject. EEG was recorded continuously from 128 Ag/AgCl active electrodes contained in a nylon electrode cap (BioSemi, Amsterdam, Netherlands), relative to a linked mastoid reference. The electrodes were placed according to the 10-5 electrode system (Oostenveld and Praamstra, 2001). Eye movements were monitored using bipolar electrode arrangements on the outer canthi for horizontal movements and below each eye (on the lower orbital portion of the orbicularis oculi muscle) for vertical movements. All signals were sampled at 512 Hz and were amplified with a bandpass of 0.10 to 100 Hz by BioSemi Active-Two amplifiers. Offline processing of the EEG signals was performed using BrainVision Analyzer software (BrainProducts, Munich, Germany). Continuous EEG was segmented from 100 milliseconds before stimulus onset to 1,000 milliseconds after stimulus onset. Trials containing eye movement artifacts and those containing amplitude differences greater than 150 V were excluded from analysis. A baseline correction was performed for each trial using the 100 milliseconds prestimulus period. To obtain latency and amplitude measurements two analysis strategies were adopted. The first was standard averaging, where for each subject trials were averaged across each condition. Peaks were then detected manually in each averaged CHEPs waveform. The second strategy involved automated single trial analysis using a multiple linear regression approach (Mayhew et al., 2006). This linear regression algorithm has been developed into a graphical user interface that runs in the MATLAB (Mathworks Inc.) environment, for further details of this new EP analysis tool readers are referred to Mayhew et al., (2006). For each subject a basis set is derived from the time averaged data and is regressed against the EP response in each trial. This provides a peak latency and a baseline to peak amplitude for the N2 and P2 waves. RESULTS A2 2 2 (temperature limb stimulation method) analysis of variance revealed that the 51°C stimulus produced higher pain ratings than the 41°C stimulus (F(1,9) 40.4, P 0.001). Stimulation of the arm resulted in higher pain ratings than the leg (F(1,9) 90.7, P 0.001), and ratings were higher for the variable stimulus location than for the fixed location (F(1,9) 16.4, P 0.003). In addition there was an interaction effect between intensity and method (F(1,9) 14.8, P 0.004); pain ratings were higher for variable than for fixed location stimuli at 51°C, but there was no difference at 41°C. Mean pain ratings for each condition are presented in Fig. 1 and Table 1. In the grand average CHEPs waveforms at Cz a negative peak and a positive peak were identified (Fig. 2) with mean latencies FIGURE 1. Subjective pain intensity ratings for fixed and variable stimulus locations at 41°C and 51°C for the arm and the leg. Higher pain ratings were observed for the 51°C stimulus the 41°C stimulus (F(1,9) 40.4, P 0.001), for stimulation of the arm than the leg and for the variable stimulus location than for the fixed location. TABLE 1. Mean and Standard Deviation (SD) for Amplitude and Latency of Contact Heat Evoked Potentials (CHEPS) for Fixed and Variable Location Stimulation of the Arm and the Leg at 51°C When Analyzed Using Standard Averaging (Std Avg) and Automated Single Trial Averaging (ST Avg) Arm Fixed Std Avg Mean N-P amplitude, V Mean N latency, ms Mean P latency, ms ST Avg Mean N-P amplitude, V Mean N latency, ms Mean P latency, ms Pain ratings, 41°C Pain ratings, 51°C 10.7 (4.7) 483 (85) 578 (78) 15.8 (6.2) 456 (45) 578 (64) 2.6 (1.5) 12.7 (4.6) Variable 19.9 (13.5) 441 (43) 546 (35) 25.5 (8.8) 425 (54) 546 (61) 3.7 (2.6) 14.1 (3.7) Fixed 10.7 (11.9) 503 (57) 599 (35) 16.9 (17.1) 489 (55) 652 (74) 1.3 (1.1) 7.7 (5.1) Leg Variable 20.5 (16.9) 477 (23) 604 (49) 27.9 (20.5) 475 (91) 656 (58) 1.4 (0.09) 10.2 (5.6) The table also includes mean pain ratings for all conditions. 118 Copyright © 2009 by the American Clinical Neurophysiology Society Journal of Clinical Neurophysiology • Volume 26, Number 2, April 2009 Optimizing Contact Heat Evoked Potentials FIGURE 2. Grand average Contact Heat Evoked Potentials (CHEPs) at Cz for the 51°C and the 41°C stimuli applied to the arm and the leg for each stimulation method. Each grand average was created from data for the subjects displaying a measurable cortical response to heat stimulation in each condition, the number of subjects included for each condition can be seen in the figure legend. The negative (N) and positive (P) peaks identified for quantification of CHEPs amplitude and latency are marked on the bottom right figure. CHEPs were clearer and larger in amplitude for variable location stimulation than for fixed location stimulation and for the 51°C stimuli than the 41°C stimuli. CHEPs were similar in amplitude for the arm and the leg. across conditions of 476 milliseconds and 582 milliseconds, respectively, the peaks are equivalent to the N2 and P2 cited in other CHEPs papers (Greffrath et al., 2007; Iannetti et al., 2006). Figure 2 illustrates that the CHEPs in the 41°C conditions are noisy and unclear, for this reason it was not possible to manually identify CHEPs at the single subject level, particularly for stimulation of the leg. The number of subjects where CHEPs could be identified in each condition is presented in Figure 2. The mean latency and peak amplitude of the N2 and P2 for each condition for the 51°C stimulus using both standard averaging and single trial analysis can be seen in Fig. 3 and Table 1. It was anticipated that automated single trial analysis would allow more objective measurement of CHEPs in the 41°C conditions, however, amplitude values similar to background noise were obtained. It is also possible that stimulation at 41°C elicits a C-fiber response which would be observed later than the CHEPs latencies reported here and would require a longer analysis window. To this end, additional analysis was conducted on a time window extending 2,000 milliseconds post stimulus. The signal to noise ratio for the 41°C conditions was low ( 10) making identification of responses unlikely. Thus, the 41°C data were not included in further analysis. It is possible that a later, C fiber response is also demonstrated in the 51°C data, however due to the lack of an observable peak no analysis was performed outside of the initial 1,000 milliseconds post stimulus window described above. A 2 2 2 (EP measure limb stimulated stimulation method) repeated measures analysis of variance was conducted for N2-P2 amplitude, N2 latency and P2 latency. For N2-P2 amplitude single trial analysis resulted in higher amplitude than standard Copyright © 2009 by the American Clinical Neurophysiology Society averaging (F(1,5) 7, P 0.04, 2 0.58). N2-P2 amplitude was also larger for the variable stimulus location compared with fixed location stimuli (F(1,5) 10.1, P 0.03, 2 0.67). There was no main effect for limb and no interaction effects. The latency of N2 was longer for stimulation of the leg than the arm (F(1,5) 9.1, P 0.03, 2 0.65). There was a measure by limb interaction effect (F(1,5) 14.17, P 0.01, 2 0.75). Post hoc comparisons showed that latency was longer for the leg than the arm when analyzed using the standard average (F(1,5) 15, P 0.01, 2 0.75), however, for single trial analysis the effect was marginal (F(1,5) 4.6, P 0.08, 2 0.48). There were no other significant effects for N2. P2 latency was longer for single trial analysis than standard averaging (F(1,5) 9.4, P 0.03, 2 0.65) and for the leg than the arm (F(1,5) 15.4, P 0.01, 2 0.76). There were no other significant effects for P2. The stacked plots in Fig. 4 represent all trials for each condition for one subject. The individual trials are presented with the first trial at the bottom of the plot and the last trial at the top of the plot. Each trial is shown with the amplitude color coded to illustrate consistent patterns across trials. A consistent negativity between 400 and 500 milliseconds and a consistent positivity between 500 and 600 milliseconds can be seen for stimulation with a varied location on the arm and the leg. This negativity and positivity represent the N2 and the P2 of the CHEPs at the single trial level. It is worth noting that these patterns are not seen for the fixed stimulus position conditions. The plot demonstrates habituation of the response to a fixed stimulus for both the arm and the leg: the positive and negative patterns are no longer apparent at the top 119 T. Warbrick et al. Journal of Clinical Neurophysiology • Volume 26, Number 2, April 2009 FIGURE 3. Mean ( SD) peak to peak amplitude, N2 and P2 latencies for each condition at 51°C, for both single trial analysis and standard averaging. Each mean was created from subjects displaying a measurable response in each condition, the number of subjects included in each condition can be seen in Fig. 2. Single trial averaging results in larger amplitude measures than standard averaging for all stimulation methods. of the plot (i.e., the later trials) for the fixed location, but are still visible for the varied location. Pearson’s product moment correlation coefficient r showed that mean subjective pain intensity ratings were only related to N2-P2 amplitude in standard average CHEPs for fixed position stimulation of the arm (r(9) 0.7, P 0.03). For the variable position stimulation of the arm there was a trend for a positive relationship (r(8) 0.6, P 0.09). However, for stimulation of the leg this trend was not as strong for fixed or variable position stimuli (r(5) 0.7, P 0.08 and r(7) 0.2, P 0.81, respectively). A similar pattern was also seen for ST amplitude within subjects, although none of the correlations reached statistical significance. For the arm both fixed stimuli and variable stimuli displayed a trend for a positive correlations (r(8) 0.6, P 0.1) and (r(8) 0.5, P 0.2), respectively. Whereas for the leg only fixed stimuli demonstrated the same trend (r(5) 0.58, P 0.2), with varied stimuli showing no evidence of a positive relationship (r(5) 0.1, P 0.9). We found a positive relationship between pain intensity ratings and N2-P2 amplitude for fixed position stimulation of the arm only. However, the trend for a positive relationship between pain ratings and N2-P2 for variable stimulation of the arm and fixed stimulation of the leg show strong correlations (r 0.5, Cohen, 1988) which suggests the failure to reach statistical significance is due to a lack of statistical power given that CHEPs were not identified in all subjects. The reason for the absence of a trend for a positive correlation for variable stimulation of the leg is unclear. The findings from previous research are equivocal; a positive relationship between pain intensity and CHEPs amplitude has been shown for the arm (Chen et al., 2002; Granovsky et al., 2005) and the leg (Granovsky et al., 2006), however, one study (Chen et al., 2006), found no relationship between pain ratings and CHEPs amplitude for stimulation at any site. Further research is required to address the apparently small relationship between intensity ratings and CHEPs amplitudes recorded from the leg. Comparison of automated single trial analysis and standard averaging for 51°C stimulation CHEPs revealed a higher N2-P2 amplitude for automated single trial analysis. Furthermore, latency differences were observed for both the N2 and P2. The most likely explanation for these amplitude and latency differences is the waveform distortion effects apparent in standard averaging (Iannetti et al., 2006; Mayhew et al., 2006). The potential for reduction of waveform distortion, in terms of peak latency and peak amplitude, when using automated single trial analysis is particularly pertinent for responses to contact heat given the latency jitter in CHEPs from trial to trial. This latency jitter is likely to be related to stimulus parameters, such the rise time of the stimulus temperature (Iannetti et al., 2006), a level contact with the thermode (Baumgartner et al., 2005; ¨ Casey, 2006), and the potential for a discrepancy between stimulation temperature and temperature at the nociceptive nerve terminals (Baumgartner et al., 2005). Given that these factors are an integral ¨ part of CHEPs stimulation, and solutions to these potential problems Copyright © 2009 by the American Clinical Neurophysiology Society DISCUSSION Our results demonstrate that single trial averaging is more effective for quantifying CHEPs than standard averaging when suboptimal stimulation parameters are used. Contact heat evoked potentials were recorded successfully from stimulation of the arm and the leg at 51°C using a variable stimulus position. The CHEPs were slightly longer in latency than pain EPs from other stimulation methods, consistent with other studies (Granovsky et al., 2005, 2006; Iannetti et al., 2006). This longer latency for CHEPs has been related to differences in stimulus variables such as location, temperature, duration, and surface area of the stimulus (Granovsky et al., 2005), and the slower rise time of skin temperature for contact heat stimulation compared with laser stimulation (Chen et al., 2002). This latency effect is consistent and a characteristic of CHEPs. In line with stimulus intensity dependant increases in ERP amplitude seen for other somatosensory stimuli (Bromm, 1984), the CHEPs presented here were larger in amplitude for the 51°C stimulus than the 41°C stimulus. 120 Journal of Clinical Neurophysiology • Volume 26, Number 2, April 2009 Optimizing Contact Heat Evoked Potentials FIGURE 4. Stacked plots of Contact Heat Evoked Potential (CHEPs) from stimulation of the arm and the leg at 51°C for fixed and variable thermode positions for one representative subject. In the stacked plot each trial is shown with the amplitude color coded. Although the data itself has not been smoothed across trials, the resulting image has been smoothed within BrainVision Analyzer for visualization purposes. A consistent negativity between 400 –500 milliseconds and a consistent positivity between 500 and 600 milliseconds can be seen for stimulation with a varied location on the arm and the leg. This negativity and positivity represent the N peak and the P peak of the CHEPs at the single trial level, and are indicated with arrows, for the fixed position stimuli the arrows on the diagram point to where the consistent negativity and positivity would be expected to illustrate their absence. These patterns are not seen for the fixed stimulus position conditions, demonstrating habituation of the response to a fixed stimulus for both the arm and the leg. are not readily available, it is essential to use appropriate analysis tools to optimize CHEPs measurement. Contact heat evoked potentials could not be clearly identified in all subjects when stimulating at 41°C, particularly when stimulating the leg. This is consistent with other studies reporting a failure to evoke CHEPs in all subjects at lower stimulation intensities (Granovsky et al., 2005; La Pera et al., 2002; Truini et al., 2007). The generation of CHEPs at exclusively painful temperatures suggests that CHEPs are related to the coding and intensity of the painful stimulus (La Pera et al., 2002), this is supported by the physiology of contact heat stimulation. The vertex (Cz) potential evoked at the latency of the CHEPs discussed here is thought to be Copyright © 2009 by the American Clinical Neurophysiology Society a result of A mechanothermal nociceptive afferent stimulation (Harkins et al., 2000; Itskovitch et al., 2000), the same fibers stimulated to produce pain LEPs. Furthermore, CHEPs have demonstrated a similar dipole model explaining their scalp distribution to that of LEPs (Valeriani et al., 2002). As LEPs are considered to be driven by nociceptive fibers only, CHEPs can also be considered of nociceptive origin. A further explanation for the absence of CHEPs at 41°C is that the baseline temperature of 32°C activates a number of warm/ low threshold heat nociceptive fibers which would result in a reduction of afferent volley synchrony to subsequent heat pulses (Harkins et al., 2000). This might be particularly apparent when the 121 T. Warbrick et al. Journal of Clinical Neurophysiology • Volume 26, Number 2, April 2009 probe is stationary and at a constant temperature several seconds before the heat pulse (Harkins et al., 2000). It is also possible that stimulation at 41°C elicits a C-fiber response (Granovsky et al., 2005) owing to the lower activation threshold of C-fibers. The C-fiber response is later than the A response due to a slower conduction velocity, as demonstrated using LEPs (Plaghki and Mouraux, 2003). Therefore, it would be reasonable to expect a later CHEPs response for the 41°C stimulus. However, we were not able to identify this in the data presented here, even with the more objective automated single trial analysis. A factor contributing to this could be the low single to noise ratio for the 41°C conditions, which could be resolved by having more trials per condition at lower temperatures. The ability to record CHEPs at lower temperatures would provide greater flexibility in experimental design and clinical applications, for example, when comparing responses across stimulus intensities. Further investigation of CHEPs elicited at lower intensities is required. Given the expected smaller amplitude of low temperature CHEPs it is recommended that single trial averaging is applied to avoid loss of information through standard averaging. A reduction in CHEPs amplitude was seen for stimulation at a fixed location compared with a varied location. Peripheral habituation of CHEPs has been demonstrated previously (Greffrath et al., 2006), where an increased habituation effect in fixed compared with varied location stimuli was found. It is likely that some central habituation occurs as a small degree of habituation was seen for varied location stimuli, but the larger degree of habituation for fixed location stimuli can be attributed to peripheral habituation, or fatigue of peripheral nociceptors (Greffrath et al., 2007). This effect seems to be particularly prominent in contact heat research in comparison with other stimulation methods, particular the use of lasers. It is not clear why this effect is seen more frequently for CHEPs than LEPs. One possibility is that in LEP research the laser beam is generally moved for each stimulus preventing nociceptor fatigue and cutaneous lesions from laser stimulation. For example, when comparing CHEPs and LEPs Iannetti et al., (2006) moved the laser beam after each stimulus but kept the contact heat thermode in a constant position throughout each stimulation block. It is possible that the reduction in amplitude of CHEPs for fixed position stimulation is a result of nociceptor physiology rather than an effect specific to heat stimulation. Our findings indicate that single trial averaging results in larger amplitude measurements for fixed location stimulation, indeed single trial averaging of the fixed location stimulation yielded similar amplitudes to standard averaging of the variable location stimuli (Fig. 3). This indicates that appropriate analysis methods can be beneficial when it is necessary to make compromises in experimental design. In contrast to previous research (Chen et al., 2006; Granovsky et al., 2006), we found similar amplitudes for CHEPs evoked from stimulation of the leg and the arm. Despite similar amplitude for CHEPs evoked from the arm and the leg, CHEPs were only clearly identified in 80% of subjects for variable stimulation of the leg (less for fixed stimulation) in comparison with 100% of subjects for both fixed and variable stimulation of the arm. The reasons for this are unclear, however, it is apparent that stimulation of the leg does evoke CHEPs in most subjects. We anticipated that using single trial averaging would improve the amplitude values gained from stimulation of the leg, which is indeed the case when CHEPs are present. However, the problem is a lack of signal (low SNR) in the two subjects that did not display clear CHEPS rather than amplitude information simply being lost during standard averaging. Single trial averaging can therefore provide no additional benefit for the data from these two subjects. It is worth bearing in mind that fixed position stimulation of the arm resulted in measurable CHEPs in all subjects, but with reduced amplitude compared with varied location stimuli. However, as described above, this can be improved by using single trial averaging. It would seem that optimizing CHEPs recording requires a trade off between maximizing the amplitude of responses and the reliability of evoking CHEPs. In conclusion, we demonstrated CHEPs in all subjects when stimulating the arm and in most subjects when stimulating the leg. The reduced amplitude for fixed location stimulation on the arm was improved with ST averaging. Our findings demonstrate that recording CHEPs in situations requiring fixed location stimulation or multiple site stimulation can be facilitated by appropriate consideration of stimulation and analysis methods. REFERENCES Baumgartner U, Cruccu G, Iannetti GD, et al. Laser guns and hot plates. Pain. ¨ 2005;116:1–3. Bromm B. Pain-related components in the cerebral potential: experimental and multivariate statistical approaches. In: Bromm B ed. Pain measurement in man. Neurophysiological correlates of pain. Amsterdam: Elsevier Science, 1984. Casey KL. Laser guns and hotplates revisited. Pain. 2006;120:325–327. Chen AC, Niddam DM, Arendt-Nielsen L. Contact heat evoked potentials as a valid means to study nociceptive pathways in human subjects. Neurosci Lett. 2001; 316:79–82. Chen ACN, Niddam DM, Crawford HJ, et al. Spatial summation of pain processing in the human brain as assessed by cerebral event related potentials. Neurosci Lett. 2002;328:190 –194. Chen IA, Hung SW, Chen YH, et al. Contact heat evoked potentials in normal subjects. Acta Neurol Taiwan. 2006;15:184 –191. Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. Hillsdale, New Jersey: Lawrence Erlbaum Associates, Inc., 1988. Gracely RH. Evaluation of multi-dimensional pain scales. Pain. 1992;48:297– 300. Granovsky Y, Granot M, Nir RR, et al. Objective measurement of subjective pain perception by contact heat evoked potentials, 5th Congress of the European Federation of IASP Chapters (EFIC), Istanbul, 2006. Granovsky Y, Matre D, Sokolik A, et al. Thermoreceptive innervation of human glabrous and hairy skin: a contact heat evoked potentials analysis. Pain. 2005;115:238 –247. Greffrath W, Baumgartner U, Treede RD. Habitiuation of human heat pain and ¨ evoked potentials reflects fatigue of peripheral nociceptors. Acta Physiol. 2006;186(suppl 1):121. Greffrath W, Baumgartner U, Treede RD. Peripheral and central components of ¨ habituation of heat pain perception and evoked potentials in humans. Pain. 2007;132:301–311. Harkins SW, Price DD, Roy A, et al. Somatosensory evoked potentials associated with thermal activiation of type II A d mechanoheat nociceptive afferents. Int J Neurosci. 2000;104:93–111. Iannetti GD, Leandri M, Truini A, Zambreanu L, Cruccu G, Tracey I. A delta nociceptor response to laser stimuli: selective effect of stimulus duration on skin temperature, brain potentials and pain perception. Clin Neurophysiol. 2004;115:2629 –2637. Iannetti GD, Zambreanu L, Tracey I. Similar nociceptive afferents mediate psychophysical and electrophysiological responses to heat stimulation of glabrous and hairy skin in humans. J Physiol. 2006;577:235–248. Itskovitch VV, Fei DY, Harkins SW. Psychophysiological and psychophysical responses to experimental pain induced by two types of cutaneous thermal stimuli. Int J Neurosci. 2000;105:63–75. La Pera D, Valeriani M, Niddam DM, et al. Contact heat evoked potentials topainful and non-painful stimuli: effects of attention towards stimulus properties. Brain Topogr. 2002;15:115–123. Mayhew SD, Iannetti GD, Woolrich MW, et al. Autmated single-trial measurement of amplitude and latency of laser evoked potentials (LEPs) using multiple linear regression. Clin Neurophysiol. 2006;117:1331–1344. Oostenveld R, Praamstra P. The five percent electrode system for high-resolution EEG and ERP measurement. Clin Neurophysiol. 2001;112:713–719. Plaghki L, Mouraux A. How do we selectively activate skin nociceptors with a high power infrared laser? Physiology and biophysics of laser stimulation. Neurophysiol Clin. 2003;33:269 –277. Treede RD. Neurophysiological studies of pain pathways in peripheral and central nervous system disorders. J Neurol. 2003;250:1152–1161. Truini A, Galeotti F, Pennisi E, et al. Trigeminal small-fibre function assessed with contact heat evoked potentials in humans. Pain. 2007;132:102–107. Valeriani M, Le Pera D, Niddam DM, et al. Dipolar modelling of the scalp evoked potentials to painful contact heat stimulation of the human skin. Neurosci Lett. 2002;318:44 – 48. 122 Copyright © 2009 by the American Clinical Neurophysiology Society