JAPPL-01055-2011R2

1 Dyspnea-pain counter-irritation induced by inspiratory threshold loading: 2

a laser evoked potentials study 3 4 5

6 Guillaume BOUVIER*1, Louis LAVIOLETTE*1, Felix KINDLER1 7

Lionel NACCACHE 2, André MOURAUX 3, 8 Thomas SIMILOWSKI 1,4 **, Capucine MORELOT-PANZINI 1,4 ** 9

10 * The first two authors contributed equally to this work 11 ** The last two authors contributed equally to this work 12 13 14 15 1 Université Paris 6, ER10UPMC, Laboratoire de Physiopathologie Respiratoire, Paris, France 16 17 2 Assistance Publique-Hôpitaux de Paris, Groupe Hospitalier Pitié-Salpêtrière, Départements de 18 Neurologie et de Neurophysiologie, Paris, France 19 20 3 Université catholique de Louvain, Institut de Neuroscience, Louvain, Belgium 21 22 4Assistance Publique-Hôpitaux de Paris, Groupe Hospitalier Pitié-Salpêtrière, Service de 23 Pneumologie et Réanimation Médicale, Paris, France 24 25 26 27 Running title: Laser-evoked potentials and dyspnea 28 29 30 31 32 Corresponding author: 33 Capucine MORÉLOT-PANZINI 34 Laboratoire de Physiopathologie Respiratoire, 35 Service de Pneumologie et Réanimation, Groupe Hospitalier Pitié Salpêtrière, 36 47-83 boulevard de l’hôpital, 75651 37 Paris Cedex 13, France 38 Phone : +33 01 42 16 78 59 39 Fax : +33 0 1 42 16 77 87 40 Email : capucine.morelot@psl.aphp.fr 41 42

43

44

Articles in PresS. J Appl Physiol (January 19, 2012). doi:10.1152/japplphysiol.01055.2011

Copyright © 2012 by the American Physiological Society.

JAPPL-01055-2011R2

2/18

ABSTRACT 45

46

Background. 47

Experimentally induced dyspnea of the work/effort type inhibits, in a top-down manner, the 48

spinal transmission of nociceptive input (dyspnea-pain counter-irritation). Previous studies 49

have demonstrated that this inhibition can be assessed by measuring the nociceptive flexion 50

reflex (RIII). However, its clinical application is limited because of the strong discomfort 51

associated with the electrical stimuli required to elicit the RIII reflex. 52

Study objectives. 53

We examined whether the dyspnea-pain counter-irritation phenomenon can be evaluated by 54

measuring the effect of work/effort type dyspnea on the magnitude of laser-evoked brain 55

potentials (LEPs). 56

Methods. 57

Ten normal male volunteers were studied (age: 19-30 years). LEPs were elicited using a CO2 58

laser stimulator delivering 10-15 ms stimuli of 6±0.7 W over a 12.5 mm2 area. The EEG was 59

recorded using 9 scalp channels. Non-nociceptive somatosensory-evoked potentials (SEPs) 60

served as control. LEPs and SEPs were recorded before, during and after 10 minutes of 61

experimentally induced dyspnea (inspiratory threshold loading, ITL). 62

Results. 63

Pain caused by the nociceptive laser stimulus was mild. ITL consistently induced dyspnea, 64

mostly of the "excessive effort" type. Amplitude of the N2-P2 wave of LEPs decreased by 65

37.6±13.8% during ITL, and was significantly correlated with the intensity of dyspnea (r=0.66, 66

CI 95% [0.08-0.92, p=0.0319]). In contrast, ITL had no effect on the magnitude of non-67

nociceptive SEPs. 68

Discussion. 69

Experimentally induced-dyspnea of the work/effort type reduces the magnitude of LEPs. This 70

reduction correlates with the intensity of dyspnea. The recording of LEPs could constitute a 71

clinically applicable approach to assess the dyspnea-pain counter-irritation phenomenon in 72

patients. 73

74

75

76

Keywords 77

Breathlessness, diffuse noxious inhibitory controls, C fibers, Aδ fibers78

JAPPL-01055-2011R2

3/18

INTRODUCTION 79

Dyspnea is a distressing and debilitating symptom that is frequent in a wide variety of 80

clinical conditions (10, 17, 22). Dyspnea shares many characteristics with pain (2, 7), but its 81

intimate neurophysiological mechanisms are less precisely known. The pain-dyspnea analogy 82

is therefore useful to gain insights into the physiological determinants of dyspnea. For instance, 83

experimentally inducing a sensation of excessive respiratory work or effort — one of the major 84

forms of dyspnea — results in a marked inhibition of the nociceptive flexion reflex RIII in 85

healthy humans (26). This phenomenon is akin to counter-irritation, defined as the attenuation 86

of a preexisting pain by a novel heterotopic noxious stimulus. Counter-irritation is thought to 87

result, at least in part, from a descending modulation of spinal nociceptive transmission by 88

diffuse noxious inhibitory controls (DNICs), of which the source and pathways have been well 89

described (15, 30). The occurrence of a « dyspnea-pain counter-irritation » (DPCI) (26) in 90

response to a stimulus inducing a sensation of excessive respiratory work/effort reinforces the 91

analogy between certain forms of dyspnea and pain. Of clinical relevance, studying the 92

neurophysiological correlates of dyspnea-pain counter-irritation could provide a mean to 93

quantify dyspnea – e.g. through the potency of dyspnea to inhibit nociception – and to objectify 94

and quantify the effects of relieving interventions. However, using the nociceptive flexion 95

reflex (35) to study dyspnea-pain counter-irritation is not realistic in clinical practice, because 96

the technique is cumbersome and because the electrical stimulus required to elicit the 97

nociceptive response is usually perceived as very painful. 98

Infrared laser stimuli applied onto the skin can be used to briefly and selectively activate 99

heat-sensitive nociceptive free nerve endings and, thereby, elicit nociceptive event-related brain 100

potentials (29, 37). Although the stimulus activates nociceptive afferents, it is non-noxious (i.e. 101

it does not inflict a lesion) and is often perceived as only mildly painful. Laser evoked 102

potentials (LEPs) have been shown to be modulated by counter irritation (1, 21, 28, 32, 40). 103

Therefore, we hypothesized that the dyspnea-pain counter-irritation phenomenon can be 104

evaluated by measuring the effect of dyspnea on the magnitude of LEPs. To test this 105

hypothesis, LEPs were recorded in healthy subjects before, during and after 10 minutes of 106

experimentally induced dyspnea for 10 minutes. Furthermore, non-nociceptive somatosensory-107

evoked potentials (SEPs) were concurrently recorded to assess the specificity of the observed 108

effects. 109

110

METHODS 111

Subjects 112

JAPPL-01055-2011R2

4/18

After ethical approval (Comité de Protection des Personnes Ile-de-France VI, Groupe 113

hospitalier Pitié-Salpêtrière, Paris, France), 10 naive healthy caucasian male volunteers (age 114

19-30 years ; body mass index [BMI] 19.4-26.6 kg*m-2) were recruited to participate in the 115

study. They were free from any past medical history, chronic or recurrent pain symptoms, and 116

did not suffer from any acute condition at the time of the study. Women were deliberately 117

excluded to avoid any risk of interference with menstrual pain. All volunteers received detailed 118

information and gave written consent. 119

120

Experimental design 121

The subjects were instructed to avoid sleep deprivation and refrain from taking 122

analgesic medication, anti-inflammatory medication and alcohol or psychotropic substances at 123

least 48 h prior to the experiment. On the day of the study, they were instructed to have a light 124

meal and to empty their bladder immediately before the experimental session, to avoid any risk 125

of interferences from visceral sources of sensory input (3, 4). During the experiments, the 126

subjects sat comfortably in a semi-reclined examination chair, with their back and head fully 127

supported. 128

The experimental design is illustrated in Figure 1. Nociceptive laser-evoked potentials 129

(LEPs) were recorded during three distinct 10 minutes periods: before, during and after 130

inducing experimental dyspnea. After a 20 minutes rest period, non-nociceptive 131

somatosensory-evoked potentials (SEPs) were recorded using the same procedure. 132

133

Respiratory measurements 134

The subjects breathed through a facemask connected in series with a heated 135

pneumotachograph (3700 series, linearity range 0-160 L*min-1; Hans Rudolph, Kansas City, 136

MO) and a two-way valve (Hans Rudolph 2600 medium, Kansas City, MO). The experimental 137

apparatus had a resistance <1 cmH2O*L-1*s-1 and its dead space was ~ 100 mL. Special care 138

was taken to fit the mask as comfortably as possible. Ventilatory airflow (V’) was measured by 139

connecting the pneumotachograph to a linear differential pressure transducer (±5 cmH2O, 140

DP45-18, Validyne, Northridge, CA). Tidal volume (VT) was obtained by electrical integration 141

of flow. Inspiratory time (TI), expiratory time (TE), total time (TT), breathing frequency (f), 142

mean inspiratory flow (VT/TI) and duty cycle (TI/TT) were obtained using a custom MATLAB 143

routine (VariVent, MATLAB R14, MathWorks Inc, Boston, MA). Inspiratory airway opening 144

pressure was measured by a differential pressure transducer (±100 cmH2O, DP15-34, Validyne, 145

Northridge, CA) connected to a lateral port of the facial mask; peak negative values for each 146

JAPPL-01055-2011R2

5/18

respiratory cycle were used in the analysis. End-tidal carbon dioxide tension (PETCO2) was 147

measured at a lateral port of the facial mask with an infrared CO2 analyzer (Servomex 1505, 148

Plaine Seine Saint-Denis, France). All respiratory signals were recorded through a analog-149

digital converter (Maclab 16S, Powerlab System, AD Instruments, Castle Hill, Australia ; 150

sampling rate 2000 Hz) and Charttm software (Chart 5.0, AD Instruments, Castle Hill, 151

Australia). 152

153

Experimental dyspnea 154

Respiratory discomfort was induced by adding an inspiratory threshold loading (ITL) 155

device (range 7-41 cm H2O threshold inspiratory muscle trainer, Nr 730, Health Scan, NJ, 156

USA) to the inspiratory arm of the breathing circuit. The load was arbitrarily adjusted to 30 157

cmH2O for all subjects and maintained for 10 min. The subjects were asked to rate the sensory 158

intensity of their respiratory discomfort if any every minute throughout the experiment (namely 159

during the baseline, ITL and recovery period) using a visual analog scale ranging from from 0 160

(“no respiratory discomfort”) to 10 (“intolerable respiratory discomfort”). Following the 161

experiment, the subjects were asked to describe their respiratory sensations by choosing one or 162

several descriptors within the list proposed by Simon et al. (34), translated by Morelot-Panzini 163

et al. (26). 164

165

EEG recordings 166

The EEG was recorded with an average reference at Fz, Cz, Pz, C3, C4, T3, T4, A1 and 167

A2 (International 10-20 system), using active surface electrodes. The electrooculogram (Fp1, 168

Fp2) was concomitantly recorded using electrodes located above both eyes. The EEG signal 169

was amplified and digitized at 2 kHz using a V-Amp amplifier (Brain Products GmbH, 170

Gilching, Germany). Electrode impedances were kept below 5 kΩ. 171

172

Nociceptive laser evoked brain potentials (LEPs) 173

Nociceptive stimuli were applied perpendicular to the dorsum of the right hand using a 174

CO2 laser stimulator (Neurolas CO2 Laser System, Electronic Engineering, Firenze, Italy). 175

Beam diameter at target site was 4 mm. Prior to the recording, the energy of the laser pulse was 176

adjusted such as to elicit a clear pinprick sensation, which was detected by the subject with a 177

reaction-time <600 ms. Subjects were asked to rate the intensity of the elicited sensation using 178

a visual-analogue scale ranging from 0 (not painful) to 10 (intolerably painful) during this 179

setting phase only. The subjects were not asked to rate the sensation elicited by the laser 180

JAPPL-01055-2011R2

6/18

stimulation during inspiratory threshold loading, mostly because of time constraints (the 181

respiratory sensation was evaluated every minute and we wanted the subjects to fully 182

concentrate on the laser stimuli, see details below). On average, the energy of the 10-15 ms 183

laser pulses used to elicit LEPs was 5.44±0.96 mJ.mm-2. Of note, we used laser stimuli close to 184

the perceptual threshold to avoid any tolerance issue, given the large number of stimulations 185

involved by the experiment. A visible He-Ne laser, collinear with the CO2 laser, was used to 186

localize the stimulated area. To reduce the risk of nociceptor sensitization and/or habituation, 187

the target of stimulation was moved by a few millimeters between each stimulus. Each stimulus 188

was preceded by a verbal warning (2-3 s) and instructions to refrain from blinking. For each 189

recording, a total of 30 stimuli were applied, with a 10-s inter-stimulus interval. 190

Signal processing. EEG signals were analysed using the Brain Vision Analyser 2 191

software (Brain Products GmbH, Gilching, Germany), as follows. Scalp signals were 192

rereferenced to the earlobe electrodes (A1-A2) and band-pass filtered using a Butterworth zero-193

phase filter (from 0.5 to 30 Hz). EEG epochs lasting 2 s were then obtained by segmenting the 194

recordings from -500 ms to +1500 ms relative to stimulus onset. Baseline correction was 195

performed using the pre-stimulus time interval (-500 to 0 ms). Furthermore, epochs 196

contaminated by ocular artifacts were rejected by visual inspection. Finally, average waveforms 197

were then obtained for each subject and experimental condition (mean number of averaged 198

segments: 27.9±4.8). 199

The peak latencies and the baseline-to-peak amplitudes of the laser-evoked N1, N2 and 200

P2 waves were measured as follows. First, the P2 wave was identified at the vertex (electrode 201

Cz vs. A1A2) as the positive peak with maximal amplitude occurring between 200 and 500 ms 202

after stimulus onset. The N2 wave was also measured at the vertex, defined as the negative 203

peak preceding P2 and occurring between 150 and 300 ms after stimulus onset. The N1 wave 204

was measured at the contralateral temporal electrode T3 rereferenced to FZ, defined as the most 205

negative peak between 100 and 200 ms after stimulus onset. N1, N2 and P2 amplitudes were 206

measured from baseline to peak. Peak latencies were measured relative to the onset of laser 207

stimulation. 208

209

Non-nociceptive somatosensory evoked potentials (SEPs) 210

Transcutaneous electrical stimulation of the median nerve (left in 6 cases, right in 4) at 211

the wrist was used to elicit non-nociceptive SEPs. Constant-current 0.2 ms electrical pulses 212

were generated using an MEB 2200 Nihon Kohden, Tokyo, Japan) and applied using a 20 mm 213

bar electrode. The perception threshold was measured using 0.2 mA stepwise increases in the 214

JAPPL-01055-2011R2

7/18

intensity of the stimulating current. The intensity of the stimulation used to elicit SEPs was set 215

to twice the perception threshold. The average stimulation intensity was of 3.16±0.69 mA. A 216

total of 1157.9±67.5 pulses were applied, using a 500-ms inter-stimulus interval. 217

Signal processing. EEG signals were analysed using the Brain Vision Analyser 2 218

software (Brain Products GmbH, Gilching, Germany), as follows. Signals were rereferenced to 219

Fz and band-pass filtered using a Butterworth zero-phase filter (from 30 to 3000 Hz). EEG 220

epochs were then obtained by segmenting the recordings from -10 ms to +200 ms relative to 221

stimulus onset. Baseline correction was performed using the pre-stimulus time interval (-10 ms 222

to 0 ms). An automatic artefact detection was used to reject all signals with an amplitude 223

exceeding ±40 µV. Finally, average waveforms were obtained for each subject and 224

experimental condition (mean number of averaged segments: 1096.8±120.8). 225

The peak latencies and baseline-to-peak amplitudes of the N20, P25 and N140 waves of 226

SEPs were measured at the central electrode contralateral to the stimulated side (C3 or C4 vs. 227

Fz). The N20 was defined as the most negative peak occurring between 15 and 25 ms after 228

stimulus onset. The P25 was defined as the most positive peak occurring between 20 and 35 ms 229

after stimulus onset. The N140 was defined as the most negative peak occurring between 130 230

and 160 ms (Garcia-Larrea et al., 1995). N20, P25 and N140 amplitudes were measured from 231

baseline to peak. Peak latencies were measured relative to the onset of laser. 232

233 Statistical analysis 234

Results are reported as mean and standard deviation. Normality was assessed using the 235

Shapiro-Wilk test. The SEP P25 amplitude was not normally distributed, but became so after 236

logarithmic transformation. The effect of ITL on each of the different measures was assessed 237

using repeated-measures ANOVA with a subject factor and with the following three conditions: 238

‘before’, ‘during’ and ‘after’ ITL. Post-hoc comparisons were conducted using orthogonal 239

contrasts. Correlations coefficients between numerical measures were computed using 240

Pearson’s R. The corresponding effect-size was estimated as Cohen's d coefficient (6). 241

Comparisons were considered statistically significant when the probability p of a type I error 242

was below 5%. Statistical analyses were performed using Statistix 9 (Analytical software, 243

Tallahassee, FL, USA), Prism 4.0 (Graphpad software Inc, CA, USA) and JMP 7 (SAS inc, 244

NC, USA). 245

246

RESULTS 247

Ventilatory pattern 248

JAPPL-01055-2011R2

8/18

In response to ITL, V’E, VT, TI/TT and VT/TI significantly increased (all p < 0.0001 vs. 249

baseline), while f decreased significantly (p < 0.0001 vs. baseline). Inspiratory pressure became 250

more negative during ITL (∆ 29.1±3.2 cmH2O, p < 0.0001 vs. baseline). Mean PETCO2 251

decreased from baseline to ITL (44.7±4.1 and 38.8±5.1 mmHg for baseline and ITL 252

respectively, p<0.0001 vs. baseline). The ventilatory pattern returned to baseline values during 253

the recovery period, with minute-by-minute PETCO2 showing a steady increase over time, from 254

.9± . mmHg during the first minute of recovery to to . ± . mmHg during the last 255

of the tenth minute; p < 0.0001). These changes were similar during the "LEP run" (above 256

values) and the "SEP run". 257

258

Experimentally induced dyspnea 259

Figure 2 shows dyspnea ratings during the LEP run (panel A) and the SEP run (panel 260

B). ITL induced dyspnea in all the subjects (VAS = 5.11±1.66 cm, p<0.0001 vs. baseline). 261

Baseline and recovery dyspnea ratings were not significantly different (VAS = 1.27±1.1 cm 262

and 1.05±0.9 cm, respectively, p=0.15). There was no significant difference between the LEP 263

run and the SEP run (p=0.95). During ITL, the subjects characterized their respiratory sensation 264

mainly in terms of the “respiratory effort” locus of Simon's descriptors (Table 1). 265

Laser evoked potentials 266

The mean VAS pain rating corresponding to the laser perception threshold was 2.5±1.5 267

cm. Table 2 summarizes the LEP N1, N2 and P2 amplitudes and latencies for all the 268

experimental conditions. A reduction of the magnitude of the laser-evoked response was 269

consistently present in all the subjects during ITL (Figure 3, Figure 4). There was a statistically 270

significant reduction in N2-P2 amplitude (-37.6±13.8 %, p=0.0001, effect-size=0.86) (Figure 271

4). Of note, some LEPs were consistently preceded by a negative peak that was probably 272

elicited by the noise related to triggering the laser using the foot pedal (Figure 3, left). This 273

contamination disappeared entirely when the subjects were studied with soundproofing 274

headphones and did not affect the N2-P2 amplitude (Figure 3, right). 275

There was a statistically significant correlation across subjects between the intensity of 276

the ITL-induced dyspnea and the reduction, from baseline to ITL, in N2-P2 amplitude 277

(expressed as a %) (r = 0.66, 95% CI [0.08-0.92]; p = 0.0319) (Figure 5). 278

The amplitude of N1 was not significantly modulated by ITL (Δ -18%±33 ; p=0.15 vs. 279

baseline). There were no significant variations in the latencies of N1, N2 and P2 during ITL 280

compared to baseline (p=0.99 ; p=0.12 and p=0.33, respectively). During the recovery phase, 281

N2-P2 tended to baseline in 5 out of 10 subjects. 282

JAPPL-01055-2011R2

9/18

283

Somesthetic evoked potentials 284

The average intensity of electrical stimulation used to obtain SEPs for was 3.16±0.69 285

mA. Absolute values for amplitude and latency are shown in Table 3. The amplitude of N20-286

P25 components and the latency of the N20 component did not significantly vary with ITL 287

(p=0.65 and p=0.83 for N20-P25 amplitude and N20 latency respectively). The latency of the 288

P25 component significantly decreased during ITL (p=0.0053 vs. baseline). The amplitudes 289

and latencies of N140 did not change significantly during the 3 experimental conditions. 290

291

DISCUSSION 292

This study shows that experimental dyspnea of the work/effort type reduces the 293

magnitude of nociceptive laser-evoked cerebral potentials, with a significant relationship 294

between the reduction of LEP amplitude and the intensity of respiratory discomfort. 295

296 Methodological considerations 297

POPULATION SIZE 298

In spite of the small size of the study population that calls for caution regarding negative 299

comparisons, we did observe statistically significant results with a large effect-size allowing a 300

reasonable physiological discussion. 301

FIXED MAGNITUDE OF INSPIRATORY THRESHOLD LOADING 302

All subjects were exposed to the same inspiratory load (30 cmH2O) that thus 303

represented a variable proportion of their maximal inspiratory pressure. This is in contrast with 304

the choice made in a previous study (26), and probably explains the marked dispersion of the 305

dyspnea ratings between subjects (Figure 2). However, this methodological decision allowed us 306

to detect a quantitative relationship between the intensity of the induced dyspnea and the size of 307

the N2-P2 amplitude reduction (Figure 5), which is of importance in interpreting the findings 308

(see below). 309

310

Effect of dyspnea of the work/effort type on LEPs 311

We interpret the reduction of the N2-P2 amplitude induced by ITL as consistent with 312

the general hypothesis that certain forms of dyspnea inhibit nociception through a dyspnea-pain 313

counter-irritation phenomenon. Here, the observed effect was specific to nociception because 314

we did not observe significant SEPs changes that would have suggested ITL-related alterations 315

in the function of the posterior column-medial lemniscus pathway. In inhibiting LEPs (37.6 ± 316

JAPPL-01055-2011R2

10/18

13.8% reduction in the N2-P2 amplitude), ITL-induced dyspnea of the work/effort type 317

behaved like other heterotopic noxious conditioning stimuli that have been studied in a similar 318

manner. A 24-35% inhibition at the segmental level has been described in response to a cold 319

pressor stimulus inducing a pain rated circa 5.5 on a VAS (1). Capsaicin application eliciting 320

similar pain intensities has been shown to reduce the N2-P2 amplitude by 25 % (38). The same 321

proved true for thermal noxious stimuli (30% inhibition) (21). Muscular and cutaneous pains 322

can induce N2-P2 inhibition above 40% in magnitude (32). In these studies, the VAS ratings of 323

the conditioning pains were similar to the VAS ratings of the experimentally induced dyspnea 324

in the present study. 325

However, and pertinent to all the above studies, caution is needed before interpreting 326

the reduction of the LEPs N2-P2 amplitude in response to a conditioning stimulus as the result 327

of a descending modulation of nociceptive transmission. Firstly, LEPs may not always parallel 328

pain perception (20, 27): excitation of central pathways that is not phase-locked may also play a 329

role in pain perception independently of LEPs. Secondly, LEPs are sensitive to habituation, 330

namely they tend to decrease with repetitive stimulation (41). In addition, N2-P2 inhibition can 331

result from a modulation of attention (25), if the conditioning stimulus (in our case ITL) 332

distracts the subject away from the cutaneous noxious "primary" stimulus. This is why the 333

target of stimulation was moved by a few millimeters between each stimulus and why our 334

subjects were warned before each stimulation and asked to focus on the skin sensation (with the 335

aim of minimizing between-stimulation attentional variations). 336

Of note, the N140 component of the SEPs that is possibly sensitive to the focus of 337

spatial attention (11, 13) did not significantly vary across the experimental conditions (Table 338

2). Although the attentional conditions at the time of the stimulus were quite different for the 339

SEPs and the LEPs, this is a small indication that the putative impact of attention on the LEPs 340

amplitude in our subject was probably less than that of inspiratory loading. But perhaps the 341

most important argument for ITL-induced dyspnea being an important "direct" source of N2-P2 342

inhibition in our subjects is the significant and strong correlation that we observed across 343

subjects between the perceived intensity of the experimentally induced dyspnea and the 344

magnitude of the N2-P2 inhibition. While attention modulation and habituation might have 345

played a role, close to 40% of the variance of the N2-P2 amplitude was explained by dyspnea 346

intensity alone. The correlation between dyspnea and N2-P2 inhibition is in line with classical 347

counter-irritation experiments that point to a "dose-effect" relationship (23). A correlation was 348

also observed between the inhibition of the RIII spinal nociceptive reflex and the intensity of 349

dyspnea in a previous study (26). This correlation was apparent during the very course of the 350

JAPPL-01055-2011R2

11/18

ITL experiments: dyspnea increased over 5 minutes in spite of constant stimulation (wind-up 351

like dynamics), and RIII inhibition deepened in a parallel manner. Although the two studies 352

cannot be directly compared, the RIII inhibition induced by ITL was of greater magnitude 353

(circa 50%) than the LEP inhibition in our subjects (circa 35%). Yet the ITL-induced dyspnea 354

was more severe in the RIII experiments (average VAS ratings circa 8) than in the present 355

experiments (maximal VAS ratings circa 5). 356

Finally on this, other experiment-related noxious sensations could also have interfered 357

with the LEPs measured in our subjects, but very careful measures were taken so as to make the 358

subjects comfortable before and during the study (see methods). Vigilance changes could also 359

have had an effect (36), but ITL and the corresponding sensations can be expected to have 360

increased rather than decreased vilgilance. Yet an enhanced vigilance is expected to increase 361

N2-P2 amplitude. ITL-induced hypercapnia could have elevated the pain threshold (16) and 362

therefore interfered with the generation of the LEPs. This was not observed in our subjects who 363

expectedly tended to hyperventilate during ITL (42), with the exception of two of them. 364

Hypocapnia can be responsible for slower EEG responses (19, 24), but we did not observe 365

lengthened LEPs latencies. 366

All being considered, it seems reasonable to postulate that the N2-P2 inhibition in our 367

subjects really did follow inspiratory threshold loading and the associated increased work/effort 368

dyspneic sensation, at least in part. 369

Of note, the removal of the inspiratory load was not followed by a clear LEPs recovery 370

(Figure 4), although visible recovery was present in 5 of the 10 subjects (see results). This was 371

already observed with the nociceptive spinal reflex after the cessation of inspiratory threshold 372

loading (26). Similar observations exist in counter irritation studies, including LEPs ones that 373

show that recovery can be slow and partial, and is quite variable depending on experimental 374

paradigms (e.g. 20% recovery after 20 minutes to complete recovery within the first six 375

minutes following the removal of the conditioning stimulus(1, 21, 38). 376

377 Physiological significance 378

The nociceptive RIII flexion reflex, described in terms of the EMG response of a 379

muscle group to a painful electrical stimulation, is usually considered a defensive phenomenon 380

(33). It is a polysynaptic and multisegmental spinal reflex that causes a complex flexion 381

synergy of the stimulated limb (33). RIII inhibition by a conditioning noxious stimulus is not 382

observed in quadriplegic patients (31) and thus involves a supraspinal component (diffuse 383

noxious inhibitory controls, DNICs). Schematically, DNICs are thought to involve a spino-384

JAPPL-01055-2011R2

12/18

bulbo-spinal loop including the subnucleus reticularis dorsalis (SRD) in the caudal medulla (5). 385

Descending projections in the dorsolateral funiculus terminate in the dorsal horn at all levels of 386

the spinal cord and mediate the inhibition of nociception (15, 30). DNICs are triggered by 387

noxious stimuli only (18), suggesting a pivotal role for ascending C- and aδ-fiber input. Of 388

note, the occurrence of a « dyspnea-pain counter-irritation » (DPCI) (26) in response to a 389

stimulus inducing a sensation of excessive respiratory work/effort not only reinforces the 390

analogy between certain forms of dyspnea and pain, but it also points to a pivotal role of C- 391

and/or aδ-fibers in the pathogenesis of this particular sensation. Studies in patients with 392

thalamic lesions argue against a major role of attention among the determinants of RIII 393

inhibition by a heterotopic stimulus (9). Mental calculus was used by Morelot-Panzini et al (26) 394

as a control to test the effects of attention on the RIII reflex. No interference was observed. 395

This provides an additional argument to think that counter-irritation described in terms of the 396

RIII reflex does not involve prominent cortical processing, which is coherent with the 397

"brainstem nature" of the phenomenon. 398

In contrast, LEPs implicate several cortical and sub-cortical structures including the 399

thalamus, the anterior insula, the prefrontal cortex, the anterior cingular cortex and the 400

secondary somesthetic cortex(12, 39). They are influenced not only by the intensity of the laser 401

stimulation (12, 39) but also by the perceived intensity of the corresponding sensation (14). 402

They are modified by attentional tasks (28), and although their precise meaning and cortical 403

origins are still incompletely understood (39) they may be closely related to attentional 404

reorientation (27). Cortical processing is thus of major importance for LEPs. The present study 405

therefore suggests that experimental dyspnea of the work/effort type is liable to interfere with 406

nociception not only at the brainstem level, but also by perturbating the cortical processing of 407

laser-induced cutaneous pain. Of note, several of the cortical areas associated with LEPs are 408

also implicated in the sensation of dyspnea (8). 409

410

Conclusions and perspectives 411

Experimentally induced dyspnea of the work/effort type inhibits the RIII nociceptive 412

reflex (26) and laser evoked potentials (this study). The concept of dyspnea-pain counter-413

irritation is therefore extended, which provides yet another dyspnea-pain neurophysiological 414

analogy. 415

Practically speaking, the demonstration that experimentally induced dyspnea of the 416

work/effort type can be studied through the LEPs inhibition that it provokes makes LEPs a 417

JAPPL-01055-2011R2

13/18

promising tool as a quantifiable neurophysiological surrogate of this sensation that would not 418

be difficult to use in the clinical field. 419

420 GRANTS 421 422 The study was supported in part by a grant "Legs Poix" from "Chancellerie de l'Université de 423 Paris" and by "Association pour le Développement et l'Organisation de la Recherche en 424 Pneumologie et sur le Sommeil, ADOREPS". 425 426 GB was supported by a grant from the ALTADIR ("Association Ligérienne pour le Traitement 427 à Domicile, l’Innovation et la Recherche"). 428 429 LL was supported by a grant from the Fondation de l’Institut Universitaire de Cardiologie et 430 de Pneumologie de Québec (2010-2011) and by an European Respiratory Society Fellowship 431 (LTRF fellowship n°39-2011, 2011-2012). 432 433 434 435 DISCLOSURES 436 437 None of the authors have any conflicts of interest to report. 438 439 440 441

JAPPL-01055-2011R2

REFERENCES 442 443 1. Arendt-Nielsen L, and Gotliebsen K. Segmental inhibition of laser-evoked brain potentials 444

by ipsi- and contralaterally applied cold pressor pain. Eur J Appl Physiol Occup Physiol 64: 56-61, 445

1992. 446

2. Banzett R, and Moosavi S. Dyspnea and Pain: Similarities and Contrasts Between Two Very 447

Unpleasant Sensations. In: APS Bulletin2001. 448

3. Bouhassira D, Chollet R, Coffin B, Lemann M, Le Bars D, Willer JC, and Jian R. 449

Inhibition of a somatic nociceptive reflex by gastric distention in humans. Gastroenterology 107: 450

985-992, 1994. 451

4. Bouhassira D, Sabate JM, Coffin B, Le Bars D, Willer JC, and Jian R. Effects of rectal 452

distensions on nociceptive flexion reflexes in humans. Am J Physiol 275: G410-417, 1998. 453

5. Bouhassira D, Villanueva L, Bing Z, and le Bars D. Involvement of the subnucleus 454

reticularis dorsalis in diffuse noxious inhibitory controls in the rat. Brain Res 595: 353-357, 1992. 455

6. Cohen J. Statistical power analysis for the behavioral sciences. NJ: Lawrence Erlbaum 456

associates, 1988. 457

7. Comroe J. Dyspnea. Mod Concepts Cardiovasc Dis 25 347-349, 1956. 458

8. Davenport P, and Vovk A. Cortical and subcortical central neural pathways in respiratory 459

sensations. Respir Physiol Neurobiol 167: 72-86, 2009. 460

9. De Broucker T, Cesaro P, Willer JC, and Le Bars D. Diffuse noxious inhibitory controls in 461

man. Involvement of the spinoreticular tract. Brain 113 ( Pt 4): 1223-1234, 1990. 462

10. Desbiens N, Mueller-Rizner N, Connors A, and Wenger N. The relationship of nausea and 463

dyspnea to pain in seriously ill patients. Pain 71: 149-156, 1997. 464

11. Eimer M, and Forster B. Modulations of early somatosensory ERP components by transient 465

and sustained spatial attention. Exp Brain Res 151: 24-31, 2003. 466

JAPPL-01055-2011R2

15/18

12. Garcia-Larrea L, Frot M, and Valeriani M. Brain generators of laser-evoked potentials: 467

from dipoles to functional significance. Neurophysiol Clin 33: 279-292, 2003. 468

13. Garcia-Larrea L, Lukaszewicz A, and Mauguiere F. Somatosensory responses during 469

selective spatial attention: The N120-to-N140 transition. Psychophysiology 32: 526-537, 1995. 470

14. Garcia-Larrea L, Peyron R, Laurent B, and Mauguier F. Association and dissociation 471

between laser-evoked potentials and pain perception. Neuroreport 8: 3785-3789, 1997. 472

15. Gebhart G. Descending modulation of pain. Neurosci Biobehav Rev 27: 729-737, 2004. 473

16. Gronroos M, and Pertovaara A. A selective suppression of human pain sensitivity by 474

carbon dioxide: central mechanisms implicated. Eur J Appl Physiol Occup Physiol 68: 74-79, 1994. 475

17. Hammond E. Some preliminerary findings on the physical complaints from a prospective 476

study of 1,064,004 men and women. Am J Public Health Nations Health 54: 11-23, 1964. 477

18. Hu J. Response properties of nociceptive and non-nociceptive neurons in the rat's trigeminal 478

subnucleus caudalis (medullary dorsal horn) related to cutaneous and deep craniofacial afferent 479

stimulation and modulation by diffuse noxious inhibitory controls. Pain 41: 331-345, 1990. 480

19. Huttunen J, Tolvanen H, Heinonen E, Voipio J, Wikstrom H, Ilmoniemi RJ, Hari R, 481

and Kaila K. Effects of voluntary hyperventilation on cortical sensory responses. 482

Electroencephalographic and magnetoencephalographic studies. Exp Brain Res 125: 248-254, 1999. 483

20. Iannetti GD, Hughes NP, Lee MC, and Mouraux A. Determinants of laser-evoked EEG 484

responses: pain perception or stimulus saliency? J Neurophysiol 100: 815-828, 2008. 485

21. Kakigi R. Diffuse noxious inhibitory control. Reappraisal by pain-related somatosensory 486

evoked potentials following CO2 laser stimulation. J Neurol Sci 125: 198-205, 1994. 487

22. Kroenke K, Arrington M, and Mangelsdorff A. The prevalence of symptoms in medical 488

outpatients and the adequacy of therapy. Arch Intern Med 150: 1685-1689, 1990. 489

23. Le Bars D, Dickenson A, and Besson J. Diffuse noxious inhibitory controls (DNIC). I. 490

Effects on dorsal horn convergent neurones in the rat. Pain 6: 283-304, 1979. 491

JAPPL-01055-2011R2

16/18

24. Ledsome JR, Cole C, and Sharp-Kehl JM. Somatosensory evoked potentials during 492

hypoxia and hypocapnia in conscious humans. Can J Anaesth 43: 1025-1029, 1996. 493

25. Lorenz J, and Garcia-Larrea L. Contribution of attentional and cognitive factors to laser 494

evoked brain potentials. Neurophysiol Clin 33: 293-301, 2003. 495

26. Morelot-Panzini C, Demoule A, Straus C, Zelter M, Derenne J, Willer J, and Similowski 496

T. Dyspnea as a noxious sensation: inspiratory threshold loading may trigger diffuse noxious 497

inhibitory controls in humans. J Neurophysiol 97: 1396-1404, 2007. 498

27. Mouraux A, and Iannetti GD. Nociceptive laser-evoked brain potentials do not reflect 499

nociceptive-specific neural activity. J Neurophysiol 101: 3258-3269, 2009. 500

28. Plaghki L, Delisle D, and Godfraind J. Heterotopic nociceptive conditioning stimuli and 501

mental task modulate differently the perception and physiological correlates of short CO2 laser 502

stimuli. Pain 57: 181-192, 1994. 503

29. Plaghki L, and Mouraux A. EEG and laser stimulation as tools for pain research. Curr Opin 504

Investig Drugs 6: 58-64, 2005. 505

30. Pud D, Sprecher E, and Yarnitsky D. Homotopic and heterotopic effects of endogenous 506

analgesia in healthy volunteers. Neurosci Lett 380: 209-213, 2005. 507

31. Roby-Brami A, Bussel B, Willer JC, and Le Bars D. An electrophysiological investigation 508

into the pain-relieving effects of heterotopic nociceptive stimuli. Probable involvement of a 509

supraspinal loop. Brain 110 ( Pt 6): 1497-1508, 1987. 510

32. Romaniello A, Arendt-Nielsen L, Cruccu G, and Svensson P. Modulation of trigeminal 511

laser evoked potentials and laser silent periods by homotopical experimental pain. Pain 98: 217-228, 512

2002. 513

33. Sandrini G, Serrao M, Rossi P, Romaniello A, Cruccu G, and Willer J. The lower limb 514

flexion reflex in humans. Prog Neurobiol 77: 353-395, 2005. 515

JAPPL-01055-2011R2

17/18

34. Simon P, Schwartzstein R, Weiss J, Lahive K, Fencl V, Teghtsoonian M, and 516

Weinberger S. Distinguishable sensations of breathlessness induced in normal volunteers. Am Rev 517

Respir Dis 140: 1021-1027, 1989. 518

35. Skljarevski V, and Ramadan N. The nociceptive flexion reflex in humans -- review article. 519

Pain 96: 3-8, 2002. 520

36. Tiede W, Magerl W, Baumgartner U, Durrer B, Ehlert U, and Treede R. Sleep restriction 521

attenuates amplitudes and attentional modulation of pain-related evoked potentials, but augments 522

pain ratings in healthy volunteers. Pain 148: 36-42, 2010. 523

37. Treede R, Lorenz J, and Baumgartner U. Clinical usefulness of laser-evoked potentials. 524

Neurophysiol Clin 33: 303-314, 2003. 525

38. Valeriani M, Tinazzi M, Le Pera D, Restuccia D, De Armas L, Maiese T, Tonali P, and 526

Arendt-Nielsen L. Inhibitory effect of capsaicin evoked trigeminal pain on warmth sensation and 527

warmth evoked potentials. Exp Brain Res 160: 29-37, 2005. 528

39. Veldhuijzen D, Nemenov M, Keaser M, Zhuo J, Gullapalli R, and Greenspan J. 529

Differential brain activation associated with laser-evoked burning and pricking pain: An event-related 530

fMRI study. Pain 141: 104-113, 2009. 531

40. Watanabe S, Kakigi R, Hoshiyama M, Kitamura Y, Koyama S, and Shimojo M. Effects 532

of noxious cooling of the skin on pain perception in man. J Neurol Sci 135: 68-73, 1996. 533

41. Weiss T, Kumpf K, Ehrhardt J, Gutberlet I, and Miltner WH. A bioadaptive approach for 534

experimental pain research in humans using laser-evoked brain potentials. Neurosci Lett 227: 95-98, 535

1997. 536

42. Yanos J, Banner A, Stanko R, Gentry S, and Greenawalt K. Ventilatory responses to 537

inspiratory threshold loading in humans. J Appl Physiol 68: 2511-2520, 1990. 538

539

JAPPL-01055-2011R2

18/18

FIGURES 540 541 Figure 1. Schematic representation of the research protocol. 542

Each rectangle represents the duration over which the event-related potentials (laser-543

evoked potentials, LEPs, or somesthetic-evoked potentials, SEPs) were collected (10 min 544

for each condition, namely baseline, inspiratory threshold loading —ITL—, recovery). A 545

20 min pause separated the LEP session from the SEP session. 546

Figure 2. Time course of dyspnea intensity as measured with the VAS scale across the 547

three experimental conditions during the LEPs run (panel A) and the SEPs run (Panel B). 548

VAS, visual analog scale; ITL, inspiratory threshold loading. * = p<0.0001 vs. baseline. 549

Each bar represents the mean value, with indication of one SD. 550

Figure 3. LEPs tracings across the 3 experimental conditions in 2 typical subjects without 551

(A) and then with (B) sound isolation. The EEG traces are shown at the Cz-A derivation. 552

Polarity is negative up. The vertical line illustrates the time of laser stimulation. See 553

Methods for details. (AEP: auditory evoked potential, ITL: inspiratory threshold loading). 554

Figure 4. LEPs N2-P2 amplitude across the 3 experimental conditions for the 10 subjects 555

(mean and SD ; ITL, inspiratory threshold loading). The "*" denotes a statistically 556

significant difference, see results for details. 557

Figure 5. Correlation between the magnitude of the fall in N2-P2 amplitude during 558

inspiratory threshold loading (ITL) in % of baseline values and the intensity of dyspnea 559

reported by the subjects during ITL (average of the minute by minute dyspnea evaluations, 560

see methods). 561

562

563 564 565 566 567 568

Table 1. Descriptors chosen by the subjects to characterize their dyspnea inspiratory threshold loading

Descriptor of the respiratory sensation Number of subjects using this descriptor

Number of subjects considering this

descriptor as the main one

My breathing requires more work* 9 4 My breathing requires effort* 9 1 My breathing requires more concentration 5 1 I cannot take a deep breath 4 1 I cannot get enough air 3 1 I feel a hunger for more air 3 0 My breath does not go in all the way 2 1 I feel that I am suffocating 2 0 My chest feels tight 1 1 I feel that I am smothering 1 0 My chest is constricted 1 0 I feel that my breathing is rapid 1 0 My breathing is shallow 1 0

*These descriptors belong to the "work" cluster defined by Simon et al. (1989).

Table 2. Characteristics of the laser-evoked potentials in relation with experimental condition Laser evoked potentials (LEP) N1 N2 P2 N2-P2 Latency (ms) Amplitude (μv) Latency (ms) Amplitude (μv) Latency (ms) Amplitude (μv)

BASELINE n = 10 171.6 (23.6) 2.4 (5.3) 220.5 (34.6) 16.0 (5.7) 335.5 (35.9) 18.37 (2.8)

ITL n = 10 172.5 (14.9) 3.0 (2.8) 236.3 (29.5) 8.1 (3.6) 349.3 (44.5) 11.2 (4.8)

RECOVERY n = 10 182.2 (13.9) 1.6 (3.5) 231.3 (26.2) 9.5 (4.5) 347.1 (48.9) 11.2 (5.0)

ANOVA F = 1.40 P = 0.2715

F = 0.82 P = 0.4572

F = 2.52 p = 0.1082

F = 19.58 p <0.0001

F = 1.36 P = 0.2827

F = 13.41 P < 0.0003

Post-hoc contrasts Baseline vs. ITL ITL vs. recovery Baseline vs. recovery

P = 0.0001 P = 0.5859 P = 0.0006

P = 0.0011 P = 0.9998 P = 0.0012

ITL, inspiratory threshold loading. Values are mean (SD). P values for post-hoc contrasts are provided only when ANOVA significant.

Table 3. Characteristics of somesthetic evoked potentials in relation to experimental condition

Somesthetic evoked potentials (SEP) N20 P25 N140

Amplitude (μv)

Latency (ms)

Amplitude (μv)

Latency (ms)

Amplitude (μv)

Latency (ms)

BASELINE n = 10 1.7 (0.9) 19.6 (0.6) 0.81

(0.50-1.32) 25.6 (3.1) 0.4 (0.2) 144.7 (8.9)

ITL n = 10 1.5 (0.6) 19.4 (0.8) 1.0 (0.6) 25.0 (3.0)* 0.4 (0.1) 142.6 (9.2)

RECOVERY n = 10 1.5 (0.8) 19.7 (0.8) 1.2 (0.7) 25.3 (3.1) 0.4 (0.1) 146.6 (10.7)

ITL, inspiratory threshold loading Values are mean (SD), except for = values are mean (95% CI) *=p<0,01

Baseline ITL Recovery Baseline ITL Recovery

Laser-evoked potentials(LEPs run)

Somesthetic-evoked potentials(SEPs run)

10 min 10 min 10 min 10 min 10 min 10 min

SubjectInstallation 20 min

Dy

sp

ne

a i

nte

ns

ity

(V

AS

sc

ale

)

0

2

4

6

8

10

Experimental conditions

BASELINE ITL RECOVERY

B. SEPs run

*

1 bar = 1 minute

Experimental conditions

Dy

sp

ne

a i

nte

ns

ity

(V

AS

sc

ale

)

0

2

4

6

8

10

1 bar = 1 minute

BASELINE ITL RECOVERY

A. LEPs run

*

-400 400 600 800 1000 1200 1400-200 2000

Time (ms)

ITL

Baseline

Recovery

N2

10 uV

-400 400 600 800 1000 1200 1400-200 2000

Time (ms)

AEP

AEP

AEP

A. B.

P2

N2

P2

N2

P2

10 uV

N2

N2

N2

P2

P2

P2

N2-P

2 a

mp

litu

de (

uV

)

0

10

20

30

Baseline ITL Recovery

F = 13.41, p = 0.0003

**

ns

r = 0.66 95% CI[0.08 - 0.92], p = 0.0319

Mean dyspnea score during ITL (VAS score)

0 2 4 6 8 10

Δ N

2-P

2 a

mp

litu

de

baseli

ne v

s.

ITL

(%

)

0

10

20

30

40

50

60

70