Pain-Related Suppression of Beta Oscillations Facilitates Voluntary Movement

Abstract

Increased beta oscillations over sensorimotor cortex are antikinetic. Motor- and pain-related processes separately suppress beta oscillations over sensorimotor cortex leading to the prediction that ongoing pain should facilitate movement. In the current study, we used a paradigm in which voluntary movements were executed during an ongoing pain-eliciting stimulus to test the hypothesis that a pain-related suppression of beta oscillations would facilitate the initiation of a subsequent voluntary movement. Using kinematic measures, electromyography, and high-density electroencephalography, we demonstrate that ongoing pain leads to shorter reaction times without affecting the kinematics or accuracy of movement. Reaction time was positively correlated with beta power prior to movement in contralateral premotor areas. Our findings corroborate the view that beta-band oscillations are antikinetic and provide new evidence that pain primes the motor system for action. Our observations provide the first evidence that a pain-related suppression of beta oscillations over contralateral premotor areas leads to shorter reaction times for voluntary movement.

Introduction

Acute physical pain is a warning signal that primes the motor system for fast and adaptive motor responses. Withdrawal reflexes and voluntary movements in response to acute painful stimuli have been well characterized (Kerr et al. 1999; Ploner, Gross, Timmermann, Schnitzler 2006; France et al. 2009; Schulz et al. 2012). However, very little is known about the cortical processes that underlie voluntary movements when an individual is already experiencing pain. One's ability to influence how pain alters human motor performance will be enhanced by understanding the cortical processes that underlie this relationship. In the current study, we use high-density electroencephalography (EEG) during limb movements (Ofori et al. 2015) to determine the neural basis for the initiation and control of voluntary movements during ongoing pain.

Endogenous and exogenous experimental pain models have been used to examine the relationship between pain and the kinetics and kinematics of motor function (for reviews see Hodges and Tucker 2011; Bank et al. 2013). Elbow flexion movements following the injection of hypertonic saline into the biceps muscle reduce movement acceleration, amplitude, and peak velocity, and lengthen reaction time (Ervilha et al. 2004). Similar effects have been found in patients with musculoskeletal pain (Shum et al. 2005; Roijezon et al. 2010), and on lower-limb and whole-body movements using chemically induced joint pain and muscle pain (Madeleine et al. 1999; Henriksen et al. 2010). Coupling pain induced by electrical stimulation with the onset of movement also leads to an interference effect as reflected by longer reaction times (Babiloni et al. 2010, 2014), and other evidence shows that reaction times are longer when individuals respond to painful laser stimuli when compared with tactile stimuli (Ploner, Gross, Timmermann, Schnitzler 2006). Collectively, these findings are at odds with the notion that pain can serve as a warning signal that primes the motor system for fast and adaptive motor responses (Shackman et al. 2011; Morrison et al. 2013). Anatomical evidence for the integration of pain and motor systems comes from a virus tracing study which shows that the spinothalamic tract system reaches multiple targets in the midcingulate cortex (Dum et al. 2009). Evidence for a pain-related priming of the motor system comes from human functional neuroimaging studies of pain processing that report activation in regions including the primary motor cortex, supplementary motor cortex, and midcingulate cortex (Gelnar et al. 1999; Wager et al. 2013; Misra and Coombes 2015). However, behavioral responses that occur while an individual is also experiencing pain (i.e., ongoing pain) are rarely measured or manipulated in neuroimaging studies (Bank et al. 2013). Hence, there is very little direct evidence that pain-related activation of the motor system actually facilitates voluntary behavioral responses.

High-density EEG studies have associated a pain-related lengthening of reaction time with increased power in the alpha band over sensorimotor cortex (Babiloni et al. 2010, 2014). In addition to modifying power in the alpha band, pain processes also suppress beta-band oscillations over sensorimotor cortex (Crone et al. 1998; Ploner, Gross, Timmermann, Pollok et al. 2006; Mancini et al. 2013), which scale with stimulus intensity and may reflect top-down processing related to the current sensorimotor state (Engel and Fries 2010; Hauck et al. 2015). Motor processes also suppress beta-band oscillations over sensorimotor cortex (Brinkman et al. 2014; Ofori et al. 2015). Consistent with the idea that increased beta power is antikinetic, faster reactions times are associated with lower beta power, and slowed or inhibited reactions are associated with high beta power (Gilbertson et al. 2005; Pastotter et al. 2008; Zhang et al. 2008; Pogosyan et al. 2009; Perfetti et al. 2011; Swann et al. 2011). Voluntary movements and pain-eliciting stimuli are therefore both associated with the suppression of beta oscillations over sensorimotor cortex, and beta oscillations are known to be antikinetic. However, previous studies show that, when the onset of an externally applied painful stimuli and movement are simultaneous, there are interference effects and slower reaction times. Hence, faster pain-related reaction times may require the onset of the painful stimulus to precede the onset of movement. Here, we address this issue by collecting high-density EEG data, while participants made visually guided arm movements during an ongoing pain-eliciting stimulus on the same arm. We tested the hypothesis that ongoing pain will activate the contralateral sensorimotor areas and shorten reaction time through the suppression of beta oscillations.

Materials and Methods

Participants

Twenty-four right-handed adults (21 ± 2 years; 11 female) with normal or corrected-to-normal vision were recruited for this study. Participants were asked not to consume caffeine or use any hair products prior to the experimental session on the day of testing. Prior to the experimental session, participants provided informed consent. This experimental study was approved by the local Institutional Review Board. Since depression and anxiety have been shown to influence how experimental pain is experienced (Rhudy and Meagher 2000; Weiss et al. 2011), all participants completed the State segment of the STAI Anxiety Inventory (Spielberger 1983), the Beck Depression Inventory (BDI: Beck and Steer 1987), and the Tampa Scale of Kinesiophobia (TSK: Kori et al. 1990). Data from one female participant was removed from all analyses based on outlier reaction time data (>2.5 standard deviation [SD] from mean). As a result, all analyses were based on data from 23 individuals (20.5 ± 1.8 years; 10 female). All questionnaire scores were within the normal range of responses given by adult participants (STAI: M = 29.7, SD = 6.9; BDI: 3.8, SD = 3.9; TSK: M = 31.1, SD = 5.0) (Spielberger 1983; Beck and Steer 1987; Thomas et al. 2008).

Thermal Stimulation

Thermal stimuli were delivered to the biceps of the right arm using a 573-mm² CHEPS thermode (PATHWAY System, Medoc Advanced Medical Systems, Chapel Hill, NC, USA). The CHEPS thermode was set to heat up at a rate of 70°C/s and cool down at a rate of 40°C/s. The high rates enabled reaching the requisite warm- and pain-eliciting temperatures in under 0.2 s. Hence, the onset and offset of heat was precisely controlled. The temperature needed to elicit a warm sensation and moderate pain was determined for each participant prior to the experiment using a calibration paradigm. During the calibration, the thermode was placed on the biceps of their right arm, and heated to 8 different temperatures (range: 41–48°C), each lasting for 4 s and repeated twice in a pseudorandom order. Trials were separated by 30 s of rest. After each stimulation, participants made a verbal rating on a scale from 0 to 10, 0 being no change in sensation, 5 being moderate pain, and 10 being intolerable pain (Paris et al. 2013; Misra et al. 2014; Misra and Coombes 2015). Intolerable pain was defined as the pain at which the thermode had to be removed from the biceps. Participants were informed that they could give noninteger ratings. For each participant, the temperature corresponding to a rating of 1–2 was used to elicit a warm sensation and a temperature corresponding to a rating of 5–6 was used to elicit moderate pain during the experiment.

Experimental Design and Task

Movements were completed in a movement-only (MO) condition in which the thermode remained at a baseline temperature (30°C), a movement-warm (MW) condition in which the thermode was heated to elicit a warm sensation (M = 42.33°C, SD = 1.14), and during a movement-pain (MP) condition in which the thermode was heated to elicit moderate pain (M = 45.8°C, SD = 1.4). The experimental setup for the movement task was similar to prior work (Ofori et al. 2015). As shown in Figure 1A, participants sat upright in a chair with their right arm supported on a cantilever beam attached to a custom-made manipulandum. The thermode was strapped to the biceps brachii of the right arm. As shown in Figure 1B, a computer monitor displayed the start position (red box), the target (white box), and a green cross-hair whose position was controlled by the participant's arm displacement. The target presented on the screen was located 72° relative to the start position. Participants were required to move the cross-hair from the start position to the target by flexing the elbow and were instructed to do so as quickly and accurately as possible. Each participant completed 6 practice trials and then performed 51 trials in each condition in an interleaved sequence. Hence, the participants always knew the type of trial they were about to perform.

Figure 1.

Setup, participant view, and trial timeline. (A) Participants performed a visually guided voluntary movement with the right arm using an elbow manipulandum, while EMG signal was collected from the right biceps brachii and brachioradialis muscles, and EEG signal was collected from the scalp. The thermode was applied to the biceps brachii of the right arm, 4 cm above the EMG electrode. (B) The participant controlled a cross-hair displayed on the screen with the elbow manipulandum. At the beginning of the trial, the cross-hair was positioned at the starting location (red box). At the cue for movement, the participant moved the cross-hair as quickly and accurately as possible to the white target box. (C) There were 2 conditions in the experiment, movement-only and movement-pain. In the movement-pain condition, heat came on 2.5 s after the start of trial and lasted for 4 s. An auditory cue for movement occurred 2.5 s after heat onset. Another auditory cue 6.5 s after heat onset signaled return to the initial position. The trial ended 9.5 s after heat onset, at which point the participant gave a verbal rating of the heat sensation on a scale of 0–10; 0 being no sensation, 2 being heat but no pain, 5 begin moderate pain, and 10 being intolerable pain. The movement-only condition was identical to the movement-pain condition except that the thermode remained at 30°C. (D) Typical timelines of the EMG signal, velocity, and acceleration during one movement-pain trial.

Figure 1C shows the timeline of events during one trial in the MP condition. The beginning of each trial was signaled by an auditory cue (Trial start), which also triggered the EEG, electromyography (EMG), and kinematic data collection. Heat onset started 2.5 s after the first auditory cue and lasted for 4 s. A second auditory cue was presented 2.5 s after heat onset and participants responded by completing the arm flexion task as quickly and accurately as possible. EMG, velocity, and acceleration data from a portion of one trial are shown in Figure 1D. A final auditory cue was presented 6.5 s after heat onset and cued the participants to move their arm back to the start position for the next trial. Participants moved their arm back to the start position at a comfortable pace. The trial ended 9.5 s after heat onset, at which point a black screen was presented on the visual display. Participants then verbally reported the sensation felt during the trial on a scale of 0–10, 0 being no change in sensation, 2 being heat but no pain, 5 being moderate pain, and 10 being intolerable pain. Ratings from individual trials were not used to recalibrate thermode temperature for future trials.

Data Acquisition

The MotionMonitor system (Innovative Sports Training, Inc., Chicago, IL, USA) was configured to synchronize and record data in real time from EMG, EEG, and kinematic systems.

Kinematic Data Acquisition

The kinematic data were collected with an angular displacement transducer (Trans-Teck, Ellington, CT, USA), which was mounted on the axis of rotation of the manipulandum. The angular displacement was transmitted and digitized at 1000 Hz using a USB-1616HS-BNC A/D board (Measurement Computing, Norton, MA, USA).

EMG Data Acquisition

Surface EMG data were collected with the Delsys Trigno Wireless System (Delsys Inc., Boston, MA, USA). Participants were prepared by rubbing ethyl alcohol on the biceps brachii and brachioradialis muscles of the right arm. Two channels were used to record the electromyographic signals from the 2 muscle groups. EMG signal gain was 909 ± 45 V/V and the sampling rate was 1000 Hz.

EEG Data Acquisition

EEG data were collected with the ActiveTwo system that comprised 128 Ag-AgCl active electrodes and a 256-channel AD-box (BioSemi, Amsterdam, the Netherlands). The ActiveTwo system replaces the conventional ground electrode with 2 separate electrodes: common mode sense (CMS) and driven right leg. These 2 electrodes form a feedback loop, which drives the average potential of the participant as close as possible to the reference voltage in the AD-box. The electrode offsets, the running averages of the voltages measured between the CMS and each active electrode, were measured before the start of each session and were kept below 40 µV.

The active electrodes were connected to a head cap that was in a preconfigured 128-electrode montage covering the entire scalp. The head cap size was selected based on the participant's head circumference (50–54, 54–58, or 58–62 cm). Standard protocols for placement of the cap were followed such that the apex electrode, which corresponded with the Cz electrode of the international 10–20 system (Jasper 1958), lied midway between the tragi and midway between the nasion and the inion. The ActiveTwo system has an analog input and a digital output. The scaling between the digital output range (2²³ to −2²³) and the analog input range (262 144 to −262 144 mV) was 32 bits/mV. EEG signals were sampled at 2048 Hz for recording the raw data. When the EEG, EMG, and kinematic signals were exported for analysis, the EEG data were downsampled to 1000 Hz for temporal synchronization with the EMG and kinematic data.

Data Analysis

All behavioral and electrophysiological data were analyzed with programs written in MATLAB (The Mathworks, Natick, MA, USA, 2014).

EMG Analysis

The EMG signals from the biceps brachii and brachioradialis were processed to identify the onset of the EMG burst and to quantify muscle activity before the movement. Raw EMG signal was first low-pass filtered at 50 Hz using a Butterworth fourth-order dual-pass filter to remove the line-noise picked up from the thermode. Thereafter, the signal was detrended and rectified. For identification of the onset of EMG burst, the rectified EMG signal was again low-pass filtered at 15 Hz using a Butterworth fourth-order dual-pass filter. The signal during the 1-s time window immediately prior to the movement cue was considered as the baseline EMG signal, and its mean and SD was calculated. The onset of EMG burst was defined as the time after the movement cue when the EMG voltage exceeded the mean baseline EMG voltage by 3 SDs.

To quantify muscle activity before movement, we integrated the rectified EMG signal prior to movement. Integrated EMG (iEMG) signal was computed using trapezoidal numerical integration on a uniformly spaced grid in 3 separate time windows: 1) before heat onset (−1.5 to 0 s), 2) between heat onset and the movement cue (0 to 2.5 s), and 3) between the movement cue and the first EMG burst (2.5 to ∼2.8 s). As the magnitude of iEMG signal is proportional to the span of the integration window, iEMG signal in each window was normalized by dividing it by the duration of the respective window.

Kinematic Analysis

Angular position data were low-pass filtered at 5 Hz using a Butterworth fifth-order dual-pass filter. Angular velocity was calculated by numerical differentiation of the angular position data. Angular acceleration was calculated by numerical differentiation of the angular velocity data. Movement onset was defined as the time at which the acceleration exceeded 20% of the peak acceleration and was identified by searching backward from the time of the peak acceleration. Reaction time was defined as the duration between the movement cue and movement onset. The end of movement was defined as the time when the velocity first returned to zero after movement onset. The timeline beyond this point included fine motor adjustments to position the cross-hair in the target box. Average velocity was calculated by dividing the total angular displacement between movement onset and movement end by the total time elapsed between the two. The average acceleration between movement onset and movement end was close to zero because this includes acceleration and deceleration phases. Average acceleration was therefore calculated by dividing the change in velocity between movement onset and time of peak velocity (i.e., time at which acceleration first returned to zero) by the total time elapsed between the two. Movement error was defined as the difference in degrees between the target box position and the position of the cross-hair when velocity first returned to zero. The square of the movement error was averaged across trials of a condition and then square-rooted to obtain the root-mean-square error (RMSE) for that condition.

EEG Preprocessing

EEG data were analyzed using custom MATLAB scripts and the EEGLAB toolbox (Delorme and Makeig 2004). EEG data were bandpass-filtered between 0.5 and 150 Hz. EEG channels were screened for large fluctuations as follows. The mean SD was calculated by averaging the SDs of all channels. Any channel with a SD greater than 4 times the mean SD was rejected and then interpolated from the neighboring channels. The interpolation procedure weighted the remaining channels with the inverse of their respective distance from the channel to be reconstructed. On an average, 2 channels per trial per participant were rejected and interpolated. The EEG signals were then re-referenced to the global average of all EEG channels. Heat onset was defined as the event for aligning the EEG signal across trials. Thereafter, 6 s long epochs, from 1.5 s before the event to 4.5 s after the event, were extracted from each trial and aligned at the event.

EEG Source Localization

EEG data were prepared for further analysis by concatenating all epochs within each condition within each participant. The Infomax independent component analysis (ICA) algorithm (Bell and Sejnowski 1995) with the natural gradient feature (Amari et al. 1996) was employed to decompose the concatenated EEG data of each participant into maximally independent components. The ICA enabled manual detection and removal of stereotyped eye, muscle, and line-noise artifacts, after which a second round of ICA was run to maximize brain-source-related independent components. Source localization of each independent component was performed using the DIPFIT toolbox, which computes an equivalent current dipole model that best explains the scalp topography of each independent component. Equivalent current dipole model was fitted using a nonlinear optimization technique (Scherg 1990) and the analytical spherical head model from Brain Electrical Source Analysis (BESA GmbH, Gräfelfing, Germany), which employs 4 spherical surfaces (skin, skull, cerebrospinal fluid, and cortex) to model the electrical properties of the human head (Kavanagh et al. 1978). The BESA head model has a template channel location map based on the standard 10-5 system. The actual channel location map was mapped onto the BESA template channel location map, which in turn was co-registered to the Montreal Neurological Institute (MNI-152) brain template. Independent components for which 15% or more of the channel variance in the scalp map could not be accounted for by a single equivalent dipole were removed from further analysis (Bigdely-Shamlo et al. 2013). Independent components for which the equivalent dipoles were located outside the MNI brain (minimum distance to the MNI brain surface being larger than 1 mm) were also removed from further analysis.

EEG Measures

Three measures were calculated for each independent component: event-related spectral perturbation (ERSP), event-related potential (ERP), and intertrial coherence (ITC). Sinusoidal wavelet transform was employed to convert the independent component EEG data from the time domain to the time–frequency domain to obtain the spectral estimate at each frequency and time point. Spectral estimates were baseline normalized by subtracting the mean baseline log power spectrum from each spectral estimate. The complex norms of the normalized spectral estimates were squared and averaged across components to obtain the ERSP value at each frequency and time (Delorme and Makeig 2004). The baseline-normalized spectral estimate at each frequency and time was length-normalized by its complex norm, which is also its vector length, and averaged across trials to obtain the ITC measure, also known as the phase-locking factor (Tallon-Baudry et al. 1996). ERPs were obtained by averaging the timeline of independent components that were source-localized in the spatial vicinity of each other (explained below).

EEG Measure Projection Analysis

Measure projection analysis (MPA) is a statistical method for characterizing the localization and consistency of EEG measures across sessions of EEG recordings (Bigdely-Shamlo et al. 2013). It allows the use of EEG as a 3D cortical imaging modality with near-cm scale spatial resolution.

MPA was employed through the Measure Projection Toolbox of EEGLAB and included the following 5 steps. 1) The locations of equivalent dipoles were overlaid on a 3D grid of voxels placed with 8-mm spacing inside the MNI brain template. 2) EEG measures for the dipole-localized independent components were spatially smoothed using a truncated 3D spatial Gaussian kernel. The SD of the Gaussian was set to 12 mm and its extent was truncated to 3 SDs (36 mm) to prevent spurious effects from distant dipoles. 3) A pair-wise independent component similarity matrix was constructed by calculating the signed mutual information between each independent component pair measure. The similarity matrix was used to calculate convergence, which was the voxel-wise expected value of measure similarity. Convergence was a scalar and was larger for areas in which the measures associated with local independent components were similar. A P-value for the convergence value at each voxel was obtained by bootstrap statistics. False discovery rate (FDR) testing was employed to correct for multiple comparisons at a group-wise error rate of 0.05 (Benjamini and Hochberg 1995). A critical P-value of 0.002 was obtained from FDR testing and hence voxels with convergence P < 0.002 were selected to define the “measure convergence subspace” of voxels at which the local similarity of independent component measures was statistically significant. 4) Within the measure convergence subspace, voxel groups that sufficiently differed from each other in their measure were identified by using affinity clustering, which automatically groups voxels into clusters, referred to as domains, based on the maximum allowed correlation between cluster exemplars. Maximal exemplar-pair similarity was set to a correlation value of 0.9, which is consistent with previous evidence (Ofori et al. 2015). To define anatomical regions, the toolbox incorporates the probabilistic atlas of human cortical structures provided by the LONI project (Shattuck et al. 2008). 5) Condition difference testing: for each participant, we calculated a projected measure in each voxel for each condition within each of the domains identified in step 4. The projected measure at a voxel was then weighted by the dipole density at that voxel, summed across all domain voxels, and normalized by the total domain voxel density to obtain a weighted-mean measure for that voxel. Bootstrapped t-tests with 1000 sample-sets were applied to the collection of participant-mean projected measures to determine significant differences between the MP and MO conditions and between the MP and MW conditions. Significant values (P < 0.05) for each contrast were retained and nonsignificant values were masked with zeros (Bigdely-Shamlo et al. 2013).

Trials were cumulatively eliminated at 3 different stages of the analysis. First, trials in which reaction time was <150 or >800 ms were excluded from all analyses. Second, trials in which EMG bursts could not be detected were excluded from all analysis. Third, trials with excessive noise in the EEG signal were eliminated. The whole trial was eliminated if it failed to meet the criterion at any one of these stages. A mean total of 89% (range: 86–92%) of trials were included in the final behavioral, EMG, and EEG analyses.

Outcome Measures and Statistical Analysis

Movement initiation was assessed using measures of reaction time. Movement kinematic measures included mean and peak acceleration, and mean and peak velocity. RMSE was used to quantify movement accuracy. For each measure, we compared the MP with the MO condition and the MP with the MW condition using two-tailed paired Student's t-tests. Significance was determined with a P < 0.05. iEMG was used to assess muscle activity prior to movement onset. For each muscle group at each time window, separate Student's t-tests were used to compare iEMG between the MP and MO conditions and between the MP and MW conditions. Significance was determined with a P < 0.05.

ERSP analyses were used to assess changes in spectral power within and across conditions. ERP and ITC analyses were used to assess whether the increases in power within and across conditions were due to phase locking. Separate between-condition analyses were conducted for ERSP, ERP, and ITC measures. Bootstrapped t-tests with 1000 sample-sets were applied to the collection of participant-mean projected measures to determine significant differences between the MP and MO conditions and between the MP and MW conditions. For all measures, significance was determined with a P < 0.05.

Results

Pain Ratings

The average rating in the MP condition corresponded with moderate pain (5.63 ± 1.09) and was significantly greater than the average rating in the MW condition (1.40 ± 0.74) and MO condition (0.11 ± 0.19). Details of the statistical comparison are shown in Table 1.

Table 1.

Average behavioral, kinematic, and accuracy measures for the 3 conditions

	Movement-pain		Movement-warm		Movement-only		MP-MW	MP-MO
	Mean	SD	Mean	SD	Mean	SD	P-value	P-value
Rating (0–10)	5.6	1.1	1.4	0.7	0.1	0.2	10−11	2 × 10−16
RT (ms)	327.2	75.5	349.3	80.1	355.5	79.7	0.030	0.002
Avg Vel (°/s)	113.4	52.7	109.9	49.3	114.7	49.7	0.817	0.722
Peak Vel (°/s)	231.2	98.7	225.7	93.6	230.9	95.2	0.894	0.894
Avg Acc (°/s2)	809.1	550.6	799.2	560.0	802.0	540.6	0.810	0.810
Peak Acc (°/s2)	1572.5	1121.1	1515.6	1060.7	1558.5	1121.1	0.861	0.861
RMSE (°)	3.4	2.2	3.6	2.2	4.0	2.4	0.833	0.164
CE (°)	1.3	2.3	1.4	2.0	1.9	2.6	0.540	0.084
VE (°)	2.8	1.3	3.2	1.6	2.9	1.3	0.196	0.389

Sensation ratings were higher and reaction times (RT) were shorter in the movement-pain condition when compared with the movement-only and movement-warm conditions. No other measure was significantly different across conditions. FDR correction was applied within each independent measure to control for multiple comparisons and FDR-adjusted P-values are reported. Degrees of freedom were 22.

Vel, velocity; Acc, acceleration; RMSE, root-mean-square error; CE, constant error; VE, variable error; SD, standard deviation; MP, movement-pain; MW, movement-warm; MO, movement-only.

Movement Initiation

Table 1 shows behavioral, kinematic, and accuracy measures and corresponding statistical tests that compare the MP with the MO condition, and the MP with the MW condition. The mean reaction time was 327 ms in the MP condition, 349 ms in the MW condition, and 355 ms in the MO condition. The black bar in Figure 2A shows that reaction time in the MP condition was 28 ms shorter than in the MO condition (P < 0.05). The white bar in Figure 2A shows that reaction time in the MP condition was 22 ms shorter than in the MW condition (P < 0.05). Could the shortening of reaction time be due to a pain-related increase in activity of the biceps brachii or the brachioradialis prior to movement onset? To quantify muscle activity before movement, we integrated the rectified EMG signal for both muscles in 3 time windows. Figure 2B shows the iEMG signal for the brachioradialis in each time window. As the magnitude of iEMG signal is proportional to the span of the integration window, iEMG signal in each window was normalized by dividing it by the duration of the respective window. There was no statistically significant difference in the normalized iEMG signal between the conditions in any of the time windows for either muscle group. These data suggest that pain-eliciting and warm thermal stimulations did not change muscle activity in the arm prior to movement onset and that peripheral factors are unlikely to play a major role in pain-related reductions in reaction time. Our findings are consistent with the previous evidence that found that saline-induced muscle pain has negligible effects on surface EMG measurements (Graven-Nielsen et al. 1997; Madeleine and Arendt-Nielsen 2005).

Figure 2.

Reaction time data. (A) Reaction time (RT) in the movement-pain (MP) condition was significantly shorter than in the movement-only (MO) condition (black bar) and significantly shorter than in the movement-warm (MW) condition (white bar). (B) Duration-normalized integrated EMG (iEMG) signal of brachioradialis in the 3 conditions for 3 different epochs. There were no statistically significant differences between conditions in any of the epochs. This suggests that muscle activity was not affected by pain and hence, cannot explain the pain-related decrease in RT. All error bars show standard error.

Movement Kinematics and Accuracy

Table 1 shows that no significant difference in mean or peak velocity, or mean or peak acceleration was found when comparing the MP condition with the MO and MW conditions. The accuracy of task performance was not affected by stimulation as there was no significant difference in the RMSE between conditions. As shown in Figure 3, plotting RMSE against average velocity reveals an interindividual speed-accuracy tradeoff in the MO condition (Fig. 3A: R² = 0.59, P < 0.05), MW condition (Fig. 3B: R² = 0.52, P < 0.05), and the MP condition (Fig. 3C: R² = 0.50, P < 0.05). Hence, pain was associated with shorter reaction times while movement kinematics and movement accuracy remained unaffected. We next examine the EEG signal to look for neural correlates of the pain-related decrease in reaction time.

Figure 3.

Average velocity versus root-mean-square error (RMSE). An inter-individual speed-accuracy trade-off was evident in all 3 conditions with R² = 0.59 in the movement-only condition, R² = 0.52 in the movement-warm condition, and R² = 0.50 in the movement-pain condition. All correlations were significant with a P < 0.001.

Source-Space Cortical Dynamics: Event-Related Spectral Perturbations

We performed MPA on the 128-channel EEG data from our participants. As a result of this analysis, we found 3 domains in cortical regions which exhibited similar ERSP measures across participants. The anatomical localization of each domain and the number of participants and independent components contributing to each domain are listed in Table 2. The first domain was localized in the hemisphere contralateral to the moving limb in premotor, motor, and somatosensory regions. The second domain was localized in the ipsilateral hemisphere in premotor, motor, and somatosensory regions. The third domain was localized in the medial parietal cortex, but it did not exhibit any significant differences between conditions and hence was not included in further analyses.

Table 2.

Statistically significant ERSP domains identified by measure projection analysis (MPA)

Domain	P	#Par	#ICs	BA	Probability	Description
Contralateral	<0.002	19	49	BA 6	0.70	Premotor and supplementary motor
				BA 4	0.20	Primary motor
				BA 3	0.10	Primary somatosensory
Ipsilateral	<0.002	22	104	BA 6	0.49	Premotor and supplementary motor
				BA 4	0.27	Primary motor
				BA 3	0.24	Primary somatosensory
Medial	<0.002	21	61	BA 7	0.42	Somatosensory association
				BA 5	0.25	Somatosensory association
				BA 4	0.18	Primary motor
				BA 3	0.08	Primary somatosensory
				BA 6	0.07	Premotor and supplementary motor

The voxel-wise P-value threshold was set at 0.002 to achieve a group-wise false discovery rate of 0.05. Of the 23 participants, 19 contributed 49 independent components to the contralateral domain, 22 contributed 104 independent components to the ipsilateral domain, and 21 contributed 61 independent components to the medial domain. The medial domain did not exhibit any significant between-condition differences and, hence, it was excluded from further analysis. The Brodmann areas each domain covers and the probability of each Brodmann area being covered was based on the brain atlas provided by the LONI project (Shattuck et al. 2008).

#Par, number of contributing participants; #ICs, number of contributed independent components; BA, Brodmann area.

ERSP in the Contralateral Domain

Figure 4A shows the contralateral domain that was localized to the premotor areas that included premotor cortex and supplementary motor area, as well as the primary motor and primary somatosensory cortices. As shown in Figure 4A and Table 2, the majority of the domain was localized to premotor areas (70%). Figure 4B shows the ERSP measure from the contralateral domain in each condition. The event for ERSP calculations was the heat onset at time zero. A duration of 1.5 s before the event was the baseline relative to which spectral perturbations were calculated up to 4.5 s after the event. Wavelet analysis of the ERSP data led to truncation of 0.75 s from each end of the ERSP plot. In all conditions, increased power is evident in the theta band right at movement onset and a decrease in power in alpha and beta bands can be seen after movement onset (Pastotter et al. 2008; Cruikshank et al. 2012; Ofori et al. 2015). Figure 4C shows the contrast plot thresholded at a P-value of 0.05 for differences in ERSP between the MP and MO conditions (MP-MO). Figure 4D shows the contrast plot for differences in ERSP between the MP and MW conditions (MP-MW). Figure 4C shows that immediately after heat onset, the MP condition shows an increase in theta power and a decrease in upper beta power (22–30 Hz) when compared with the MO condition. The MP–MW contrast shown in Figure 4D did not show this same pattern in theta and upper beta, which suggests that the initial increase in theta power and decrease in upper beta power maybe related to stimulation onset irrespective of the intensity of the stimulus. Figure 4C,D shows that, after heat onset but before the movement cue, the MP condition shows a significant reduction in alpha and lower beta power (13–18 Hz) when compared with the MO and the MW condition.

Figure 4.

Measure projection analysis (MPA) of the event-related spectral perturbation (ERSP) measure yielded statistically significant domains of brain activity. (A) The first significant domain (P < 0.002) was found in the premotor area of the contralateral hemisphere. Anatomical localization of the domain, shown on the MNI brain template, was based on the brain atlas provided by the LONI project (Shattuck et al. 2008) and is listed in Table 2. (B) ERSP measures from the premotor area for the movement-only (MO) condition, the movement-warm (MW) condition, and the movement-pain (MP) condition are shown. The timing of the heat onset was treated as the event in all conditions. The 1.5-s time window before the event served as the baseline relative to which spectral perturbations were calculated over the 4.5-s window after the event. Sinusoidal wavelet transform, which was employed to convert the EEG data to time–frequency space, led to truncation of 0.75 s from each end of the ERSP plot. (C) Contrast plots showing statistically significant differences (P < 0.05) in the ERSP between MP and MO conditions. (D) Contrast plots showing statistically significant differences (P < 0.05) in the ERSP between MP and MW conditions. Differences in the baseline were unsubstantial in both contrasts. On the basis of our a priori hypothesis of the effect of beta power on reaction time, and the significant contrast between the ERSP plots in the 2 conditions, we identified an ERSP region in a 1-s long window from 9 to 18 Hz just before the movement cue; the black boxes. (E) The area enclosed in the black boxes were extracted from the ERSP plot of each participant and Pearson's correlation coefficient and its t-statistics were calculated between pixels at each location and reaction times across the participants in each condition. Statistically significant correlation coefficients are shown enclosed in black curves (P < 0.05, uncorrected). All pixels in a region of interest above the dashed line (13–18 Hz) were averaged within participants and within conditions to obtain an averaged ERSP value for each participant and each condition. Linear regression was performed between the averaged ERSP values and reaction times in respective conditions. (F) Significant correlations between ERSP values and reaction time were found in the movement-only condition (black squares) and the movement-pain (red triangles) condition. (G) Significant correlations between ERSP values and reaction time were found in the movement-warm condition (black circles). Data for the movement-pain (red triangles) condition shown in (F) are shown again in Figure 5G to show a contrast between MP and MW.

On the basis of our a priori hypothesis of the effect of beta power on reaction time, and the significant contrast between the ERSP plots between conditions, we identified an ERSP region in a 1 s long window from 9 to 18 Hz just before the movement cue; the black boxes in Figure 4C,D. The area enclosed by the black boxes was extracted from the ERSP plot of each participant and Pearson's correlation coefficient and its t-statistic were calculated between pixels at each location and reaction times across the participants in each condition. ERSP values from 13 to 18 Hz correlated positively with reaction time in all 3 conditions as shown by the large clusters of red above the dashed lines in Figure 4E. The higher the ERSP value, the slower the participant was to initiate movement. Statistically significant correlation coefficients are shown enclosed in black curves (P < 0.05, uncorrected). ERSP values at frequencies below 13 Hz did not correlate with reaction time, suggesting that the ERSP–reaction time relationship was specific to the lower beta band.

All pixels above the dashed line (13–18 Hz) in the region of interest were averaged within participants and within conditions to obtain an averaged ERSP value for each participant and each condition. Linear regression was performed between the averaged ERSP values and reaction times in the respective conditions. Figure 4F shows that significant correlations were found in the MO condition (black squares; R² = 0.47, P < 0.01) and the MP condition (red triangles: R² = 0.45, P < 0.01). Figure 4G shows that significant correlations were also found in the MW condition (black circles; R² = 0.35, P < 0.01). For contrast purposes, Figure 4G shows the same MP data (red triangles) that are shown in Figure 4F.

On the basis of the between-condition contrasts, we also extracted ERSPs from regions of interest in theta (4–8 Hz), alpha (8–12 Hz), and upper beta (18–24 Hz) prior to movement and calculated Pearson's correlations between these values and corresponding reaction times. We did not find any significant correlations (all P > 0.05).

ERSP in the Ipsilateral Domain

In the right hemisphere, which was ipsilateral to the arm performing the motor task and experiencing the pain-eliciting stimulus, one statistically significant domain was found with a P < 0.002. Figure 5A shows the ipsilateral domain that was localized to the premotor cortex and supplementary motor cortex, as well as the primary motor cortex and primary somatosensory cortex (Table 2). Premotor areas accounted for 49% of the domain. The primary motor cortex accounted for 27% of the domain, and the primary somatosensory cortex accounted for 24% of the domain. The ERSP plot from this domain in each of the conditions is shown in Figure 5B. In all conditions, increases in theta power emerge at movement onset and decreases in alpha and beta power are evident at movement onset and persist during movement. Figure 5C,D shows the ERSP contrast plots that details statistical differences between conditions (5C: MP-MO; 5D: MP-MW), thresholded at a P-value of 0.05. Consistent with activity in the contralateral domain, the MP condition shows a significant increase in power in the theta band following heat onset compared with the MO condition but not compared with the MW condition. The MP condition also showed a significant reduction in alpha and beta power prior to movement onset when compared with each of the other conditions.

Figure 5.

ERSP data in the ipsilateral domain. (A) The second significant ERSP domain (P < 0.002) was found in the premotor region of the ipsilateral hemisphere. Anatomical localization of the domain, shown on the MNI brain template, was based on the brain atlas provided by the LONI project (Shattuck et al. 2008) and is listed in Table 2. (B) ERSP measures from the ipsilateral premotor area are shown for each condition. The timing of heat onset was treated as the event in all conditions. The 1.5-s time window before the event served as the baseline relative to which spectral perturbations were calculated over the 4.5-s window after the event. Sinusoidal wavelet transform, which was employed to convert the EEG data to time–frequency space, led to truncation of 0.75 s from each end of the ERSP plot. (C) Contrast plots showing statistically significant differences (P < 0.05) in the ERSP between MP and MO conditions. (D) Contrast plots showing statistically significant differences (P < 0.05) in the ERSP between MP and MW conditions. The black box identified in Figure 4C,D (9–18 Hz) was used to examine correlations between RT and ERSP in the ipsilateral domain. (E) Pearson's correlation coefficient and its t-statistics were calculated between pixels at each location and reaction times across the participants in each condition. Statistically significant correlation coefficients are shown enclosed in black curves (P < 0.05, uncorrected). (F,G) ERSP values averaged across the region of interest between 13 and 18 Hz did not exhibit any correlation with reaction time in any condition. This highlights the spatial specificity of the correlations observed in the contralateral premotor domain (Fig. 4E).

On the basis of our a priori hypothesis of the effect of beta power on reaction time, between-condition contrasts, and our finding in the contralateral domain, we explored correlations between ERSP and reaction times in the ipsilateral domain. We used the same area as in the contralateral domain (9–18 Hz, 1 s prior to movement); the black boxes in Figure 5C,D. The area enclosed by the boxes was extracted from the ERSP plot of each participant and Pearson's correlation coefficient and its t-statistic were calculated between pixels at each location and reaction times across participants in the 3 conditions. Figure 5E shows that the correlations were much weaker for the MO condition and negligible for the MW and MP conditions when compared with correlations in the contralateral domain. Very few pixels crossed the statistical significance limit of P < 0.05. Pixels above the dashed line in the region of interest (13–18 Hz) were then averaged within participants and within conditions to obtain an averaged ERSP value for each participant and each condition. Linear regression was performed between the averaged ERSP values and reaction times in the respective conditions. Consistent with the pixel-wise correlations shown in Figure 5E, No significant correlations were found for the MO condition (Fig. 5F: R² = 0.05, P = 0.34) or the MP condition (Fig. 5F: R² = 0.04, P = 0.4), or the MW condition (Fig. 5G: R² = 0.05, P = 0.37). Hence, even though the decrease in power in alpha and beta bands was evident in bilateral premotor areas, the ERSP–reaction time correlation was specific to the contralateral premotor area and the lower beta band.

On the basis of the between-condition contrasts, we also extracted ERSPs from regions of interest in theta (4–8 Hz), alpha (8–12 Hz), and upper beta (18–24 Hz) frequencies prior to movement and calculated Pearson's correlations between these values and corresponding reaction times. We did not find any significant correlations (all P > 0.05). ERSP findings from the contralateral and ipsilateral domains provide evidence that the mechanism for a pain-related reduction in reaction time is associated with the suppression of beta-band oscillations between 13 and 18 Hz that are predominantly localized to the contralateral premotor area.

Source-Space Cortical Dynamics: Event-Related Potentials and Intertrial Coherence

Figure 4C shows an increase in theta power in the MP condition when compared with the MO condition immediately following heat onset. This finding is consistent with the previous studies that have associated experimental pain with increased theta band power, which can be attributed to a combination of phase-locked and nonphase-locked oscillations to the onset of pain (Bromm and Treede 1991; Ohara et al. 2004; Iannetti et al. 2005). We used ERP and ITC analyses to determine whether the increase in theta power in the contralateral ERSP premotor domain could be attributed to phase-locked and/or nonphase-locked oscillations to the onset of heat stimulation. A statistically significant (P < 0.001) ERP domain was found in an area that was localized to the premotor cortex and the supplementary motor cortex (40%), and the dorsal (31%) and ventral (23%) anterior cingulate cortex. This ERP domain partially overlapped with the contralateral ERSP premotor region. We found evoked potentials triggered by the heat stimulus within 1 s of the heat onset. We also looked for phase coherence across trials and found a statistically significant (P < 0.05) ITC domain that was localized to the premotor cortex and the supplementary motor cortex (19%), the dorsal (31%), and ventral (36%) anterior cingulate cortex, and the temporal cortex (9%). The ITC domain also showed partial overlap with the ERSP and the ERP domains in the premotor areas. In the ITC plot, we found evidence for phase locking in frequencies up to 6 Hz within 1 s of the heat onset. In line with the previous evidence, the ERP and ITC analysis together show that phase-locking partially contributed to the increase in power in the theta band in the contralateral ERSP premotor area after heat onset. Theta power immediately following heat onset was not significantly different between the MP condition and the MW condition, which suggests that the increase in theta power in the MP when compared with the MO condition may not be specific to pain and instead may reflect processes associated with the onset of heat stimulation.

Discussion

Our experiment produced 3 novel results. First, we demonstrate that voluntary movements are initiated more quickly during ongoing pain while movement kinematics and accuracy remain unaffected. Second, prior to movement onset, a pain-related decrease in alpha and beta power between 9 and 18 Hz was found in a brain region that was predominantly localized to the premotor cortex and supplementary motor area. Third, power in the beta band between 13 and 18 Hz before movement in the contralateral premotor area correlated positively with reaction time in the MP condition, the MW condition, and the MO condition. Power in the alpha band between 9 and 13 Hz in the contralateral premotor area did not correlate with reaction time in any condition. Also, power in the alpha and beta in the ipsilateral premotor area did not correlate with reaction time in any condition. Our findings corroborate the view that beta-band oscillations are antikinetic and provide the first evidence that a pain-related suppression of beta oscillations over contralateral premotor areas leads to faster reaction times for voluntary movement.

Pain-Related Facilitation of Voluntary Movement

In the current study, movement initiation was cued by an auditory stimulus 2.5 s after the onset of an exogenous pain-eliciting stimulus and was therefore not influenced by interference effects in central processing that may occur when pain and movement onset are coupled (Babiloni et al. 2010). Longer reaction times have also been reported when movements are initiated following a pain-eliciting stimulus when compared with a tactile stimulus (Ploner, Gross, Timmermann, Schnitzler 2006). Slow conduction velocities in peripheral and spinal pain pathways may explain why reaction times to pain-eliciting stimuli are longer when compared with tactile stimuli (Ferrington et al. 1987; Kakigi et al. 2005; Ploner, Gross, Timmermann, Schnitzler 2006). Reaction times averaged 327 ms in the MP condition which is longer than simple reaction time tasks where speed rather than accuracy is emphasized, but shorter than the previously reported reaction times when pain-eliciting stimuli trigger or coincide with movement onset (417–900 ms) (Ploner, Gross, Timmermann, Schnitzler 2006; Babiloni et al. 2010). Although reaction time may be longer when triggered by a pain-eliciting stimulus, the time between the first evoked brain-response and motor output is shorter when the incoming signal is painful when compared with nonpainful (Ploner, Gross, Timmermann, Schnitzler 2006). Faster central processing of pain is in agreement with the current data and is consistent with the notion that pain primes the motor system for action (Shackman et al. 2011; Morrison et al. 2013). Our observations also complement previous studies that show shorter reaction times when pain-eliciting stimuli are delivered prior to movement onset (Raij et al. 2004; Perini et al. 2013). We add to this literature by demonstrating that shorter reaction times are likely to be driven by central rather than peripheral mechanisms and are related to the suppression of beta-band oscillations in contralateral premotor areas.

Antikinetic Role of Beta Oscillations

The suppression of beta-band oscillations is directly related to the disinhibition of neuronal populations involved in voluntary movement (Perfetti et al. 2011; Brinkman et al. 2014). We corroborate these observations by showing that reaction times correlated positively with beta-band power between 13 and 18 Hz in all experimental conditions. Importantly, relative to the MO and MW conditions, the MP condition led to faster reaction times and a larger reduction in beta power between 13 and 18 Hz, which is consistent with literature on the antikinetic role of beta oscillations. For instance, voluntary movements are slowed when they are triggered during spontaneous bursts of beta activity (Gilbertson et al. 2005), and when beta oscillations are entrained using transcranial alternating-current stimulation (Pogosyan et al. 2009). Increased cortical beta activity is associated with successful stop trials in Go/NoGo tasks in both human and nonhuman primates (Zhang et al. 2008; Swann et al. 2011), and increased beta activity has been demonstrated during an inhibition-of-return paradigm in which response to a target is slowed if the target is presented at the same location as the preceding target (Pastotter et al. 2008). The opposite pattern also holds true, with reductions in beta power associated with faster movements. Prestimulus beta power from 12 to 18 Hz over left sensory-motor electrodes also predicts reaction time of a voluntary reaching movement, with greater reductions in beta power associated with shorter reaction times (Perfetti et al. 2011). Our findings support the hypothesis that beta-band oscillations are antikinetic. We also show for the first time that pain suppresses beta oscillations in areas beyond the primary sensorimotor cortex in contralateral premotor regions that are associated with the planning of voluntary movement.

Motor and Pain Processing in Contralateral Premotor Areas

Motor- and pain-related changes in brain oscillations have typically focused on electrodes overlying sensorimotor cortex, with data often presented from single electrodes or clusters of electrodes over sensorimotor cortex (Babiloni et al. 2010; Perfetti et al. 2011; Schulz et al. 2012). Here, we analyzed our EEG data with MPA, which is considered a 3D functional imaging tool (Bigdely-Shamlo et al. 2013). MPA uses a blind source separation technique (Bell and Sejnowski 1995) to identify independent components in the EEG signal, and a source localization technique (Scherg 1990) for spatial localization of the identified independent components. Although independent component analyses and source localization of EEG signals have been employed for several years (Makeig et al. 1996, 1997) combining their results across participants is nontrivial. Various algorithms can be used to cluster independent components at the group level, but this methodology has too many tunable parameters to maintain objectivity and reproducibility of results (e.g., weights to ERP and ERSPs). MPA offers a reasonable solution to the problem of comparing EEG sources across individuals in 3D brain space. Source-localized ICs of each participant are projected onto a 3D grid placed over a standard MNI brain template. MPA performs statistical comparisons of a single EEG measure across participants in this 3D grid and identifies voxels that show a high degree of consistency across participants with near-cm scale spatial resolution (Bigdely-Shamlo et al. 2013). Using this novel approach, we identified ERSP domains in the contralateral and ipsilateral sensorimotor cortex that were predominantly localized to premotor areas.

Premotor areas have the anatomical substrate to influence motor processing and motor output via connections with the anterior midcingulate cortex (Dum et al. 2009; Perini et al. 2013; Hoffstaedter et al. 2014; Misra and Coombes 2015) and the primary motor cortex (Picard and Strick 1996, 2001, 2003). Engaging or stimulating the primary motor cortex is a well-established approach for pain relief (Hirayama et al. 2006; Passard et al. 2007; Lima and Fregni 2008; Paris et al. 2013; Misra et al. 2014), but much less is known about how pain influences voluntary movement. The current data suggest an important role for contralateral premotor areas in the motor–pain relationship. The lower beta power in the contralateral premotor area before movement correlated positively with reaction time, but no correlation was found in the ipsilateral premotor area. These findings suggest that although the suppression of beta power due to pain was bilateral, correlations between brain oscillations and behavior resolved to a specific effector according to the demands of the experimental task and the corresponding motor plan (i.e., moving only the right upper limb). Support for the involvement of premotor areas in movement planning is well established (Rushworth et al. 2003; Ojakangas et al. 2006; Kantak et al. 2012). Single-neuron recordings in monkey link premotor cortex activity to target location and arm use prior to movement (Hoshi and Tanji 2000). In humans, reaction time tasks lead to robust activation in premotor cortex (Sugiura et al. 2001), and transcranial magnetic stimulation studies show that perturbing premotor cortex increases reaction time (Schluter et al. 1998; Mochizuki et al. 2005). In addition to the supplementary motor area, the location of the contralateral premotor domain in the current study is more consistent with dorsal rather than ventral premotor cortex (Mayka et al. 2006). This is in agreement with evidence that the dorsal premotor areas integrate sensory information into motor commands and encode the parameters of movement when compared with the ventral premotor cortex which is more engaged during prehension and the manipulation of objects (Hoshi and Tanji 2000, 2007; Davare et al. 2006). Given that the task parameters such as target location, arm use, and hand position were identical between conditions in the current paradigm, we attribute the change in beta power in contralateral premotor areas during movement preparation to the experience of pain.

Pain influenced reaction time but endpoint accuracy and movement kinematics were not different between experimental conditions. This suggests that movement initiation and movement control are distinct processes that can be influenced independently (Carlsen et al. 2011), and that central processing speed rather than central processing quality was affected by pain. Movement adaptation to pain can protect individuals from further pain or injury, and in endogenous pain models where movements elicit pain, this can lead to a change in movement quality (Hodges and Tucker 2011). For instance, during walking, pain leads to a decrease in the normal counter rotation of the thorax and abdomen (Lamoth et al. 2002, 2004), and back pain is associated with increased movement stiffness of the trunk, and decreased trunk movement in anticipation of arm movements (Mok et al. 2007). In each of these cases, the gross features of the task are maintained, but movement quality is adapted to reduce pain and protect the body from injury. Hence, pain-related changes in movement quality may emerge when the movement itself is the pain-eliciting stimulus. Exogenous versus endogenous pain in the context of movement is an important distinction because whereas endogenous muscle pain that is elicited or exacerbated by movement may be associated with protective control strategies that lead to slower and stiffer movements, exogenous pain may serve as a warning stimulus that primes the motor system to initiate faster movements without compromising movement kinematics or endpoint accuracy. The extent to which pain alters movement may also be driven by the constraints of the motor task, the temporal relationship between onset of the pain-eliciting stimulus and the onset of movement, the predictability of pain, and the instructions given to the participants. For instance, during a force-matching task that requires individuals to be as accurate as possible, endogenous muscle pain leads to very few differences in performance, although adaptations in motor unit recruitment have been demonstrated (Tucker et al. 2009). In the current study, participants were instructed to move as quickly and as accurately as possible, the amplitude and duration of the pain-eliciting stimulus was both predictable and independent of the movement, onset of the pain-eliciting stimulus preceded the movement, and EMG activity before and during the movement was not different between conditions.

In the current study, the pain-eliciting stimulus delivered to the arm may have led to an increase in attention to that arm prior to movement, resulting in faster responses. A second explanation is that the pain-eliciting stimulus may have enhanced the perception of the cue to initiate movement. These interpretations are consistent with the previous evidence demonstrating pain-related changes in attention (Bantick et al. 2002; Ploner et al. 2011; Van Ryckeghem et al. 2013), and the previous evidence of lower beta power reflecting top-down processes that enhance perception and promote fast responses to task relevant stimuli (Perfetti et al. 2011). However, data from the MW condition suggest that one should be cautious in attributing the faster reaction times in the MP condition exclusively to attentional or perceptual processes. Given that pain is an unpleasant sensory and emotional experience, a third interpretation of our findings is that pain may have elicited emotional arousal, which has also previously been shown to lead to shorter reaction times (Coombes et al. 2007, 2009). Other evidence shows that controlling movements in emotional contexts also engages the premotor cortex (Coombes et al. 2012). Indeed, emotional state and pain may both prime the motor system for action (Frijda 1986, 2009; Peyron et al. 2000; Pessiglione et al. 2007; Schmidt et al. 2009). The current findings extend the previous evidence of a role for contralateral premotor areas in movements during emotional arousal to movements during ongoing pain.

Conclusion

Previous studies show that when voluntary movements are concurrent with a pain-eliciting exogenous stimulus, reaction times are longer and relate to changes in alpha oscillations over sensorimotor cortex. Here, we show for the first time that ongoing pain leads to shorter reaction times, which in turn are positively correlated with lower beta power prior to movement in contralateral premotor areas. Our findings support the hypothesis that beta-band oscillations are antikinetic and are consistent with the notion that exogenous physical pain primes the motor system for action. Our observations provide the first evidence of a pain-related suppression of beta oscillations over contralateral premotor areas and the consequent shortening of reaction time of a voluntary movement.

Funding

This work was supported by the National Institutes of Health (NS058487).