A novel protocol to investigate the monoaminergic relationship between descending nociceptive and motor unit neuromodulation

Abstract

Chronic pain is a multifaceted condition involving complex interactions between sensory and motor systems, in part modulated by monoaminergic neurotransmitters such as serotonin (5-HT) and norepinephrine (NE). Diffuse noxious inhibitory control (DNIC) and motor unit (MU) facilitation are critical processes that reflect descending nociceptive inhibition [1]and motor neuromodulation, respectively. This study presents a novel protocol combining pain-related evoked potentials (PREP) recorded via electroencephalography (EEG) and high-density surface electromyography (HDsEMG) to assess the effects of an experimental painful stimulus—the cold pressor test (CPT)—on both nociceptive and MU neuromodulation. Healthy, young participants completed pre- and post-CPT assessments, with PREP and HDsEMG data collected in three experimental components. Although the sample size was small, promising trends in post-CPT responses for both nociceptive and MU parameters suggest the feasibility of this approach. The findings support the hypothesis that the CPT induces enhanced nociceptive inhibition and MU facilitation in healthy controls, likely mediated by increased widespread monoaminergic drive. Future studies with larger sample sizes and diverse participant populations, including those with chronic pain conditions, will be necessary to validate these results and explore the protocol’s ability to capture both nociceptive inhibition and facilitation.

I. Introduction

Chronic musculoskeletal and neuropathic pain has posed a tremendous societal burden for decades, with modest progress toward better interventions. [1] This is in part due to the complex nature of pain processing, as it involves and influences cognitive, affective, sensory, and motor circuits within the nervous system. [2] Nociception is defined as the neural process of encoding noxious stimuli, and is the sensory input that can be interpreted as pain once processed in the cortex. [3] Both ascending nociceptive and descending motor signals can be modulated via monoaminergic neurotransmitters originating primarily in the brainstem, including serotonin (5-HT) and norepinephrine (NE). [4] An example of this modulation is the “fight or flight response”, in which it is advantageous to temporarily inhibit nociceptive input from an injury, and facilitate motor function, while escaping from a predator. Conversely, these monoamines can, at times, have the opposite effect, where nociception is facilitated and motor function is inhibited. [4, 5] Due to the complexity of the involved neurophysiological circuitry, a one-size-fits-all approach cannot be applied to patients living with chronic pain. [6, 7] Consequently, our work aims to establish a novel protocol combining validated bioelectrical signal recording methods, in order to measure facilitatory and inhibitory processes within both the sensory and motor systems.

The physiological phenomenon of widespread nociceptive inhibition, or “pain inhibiting pain”, demonstrated in animal models, is known as diffuse noxious inhibitory control (DNIC). [8–10] Experimentally, components of the DNIC effect can be tested via heterotopic noxious conditioning stimulation, in which a noxious testing stimulus is quantified before and after a painful conditioning stimulus at a distant site. [11, 12] Immersion of a hand, or foot into cold water (cold pressor test, or CPT) has been shown to be a reliable conditioning stimulus in that it should induce DNIC effects in those without impaired nociceptive inhibition. [13, 14] Pain-related evoked potentials (PREP) are cortical responses to noxious peripheral inputs such as electrical stimulation, measured via surface encephalography (EEG). [15] Different PREP protocols have been used as testing stimuli to assess the DNIC effect. [16, 17] The change in PREP signal magnitude after the conditioning stimulus indicates its effect, where a decrease and increase in magnitude reflect nociceptive inhibition (DNIC) and facilitation, respectively.

While surface EEG signals have excellent temporal resolution, they only show net cortical activity without direct insight into signals at the spinal cord level. [18] The use of high-density surface electromyography (HDsEMG) on the other hand, has been shown to allow for more direct insight into descending monoaminergic drive at the spinal cord. [19] This is due to the one-to-one spike ratio between the motoneurons at the spinal cord and the muscle fibers they innervate. [20] Together, the spinal motoneurons and their respective muscle fibers are known as motor units (MU). Decades of extensive animal and human work, including work from our labs, have shown the neuromodulatory role 5-HT and NE play in MU discharge patterns. [21, 22] These monoamines drive discharge rate acceleration, amplification, prolongation, and attenuation, all of which can be estimated with HDsEMG. [23, 24]

Across chronic pain conditions, individuals frequently demonstrate impaired nociceptive inhibition and motor function, while nociceptive facilitation is increased, resulting in increased pain levels and physical limitations. [14, 25–28] While not yet fully understood, there is substantial scientific support for the roles the monoamines 5-HT and NE play in the neuromodulation of nociceptive processing and MU discharge patterns. [22, 29–31] However, to our knowledge, investigating these processes and their interplay simultaneously is a novel approach. Given the overlap in circuitry and clinical concurrence, we hypothesized that in healthy control participants, we would observe both improved nociceptive inhibition (DNIC) and MU facilitation. The ability to measure all of these facets within one protocol could aid in more comprehensive and specific evaluations, and consequently, more targeted interventions.

II. Methods

A. Participants

Participants were five healthy males (age 28.0 ± 1.87) recruited from the university and community, without history of chronic pain, or neurological pathology affecting sensorimotor function.

B. Experimental Protocols Overview

Participants were seated in a Biodex chair (Biodex Medical Systems, Shirley, NY), with the trunk and waist restrained. All experiments consisted of a test sequence before and after the conditioning stimulus (CPT) as outlined in TABLE I. For the CPT, the left foot was placed in a container with circulating cold water (0–2°C) for two minutes, with verbal indications of perceived pain in the foot on the numeric pain rating scale (NPRS; 0–100 where 0 indicates no pain and 100 the worst pain imaginable) at baseline and every 30 seconds. The CPT was performed immediately after completion of the pre-CPT testing (PREP, HDsEMG, or both), and post-CPT testing was initiated immediately after the CPT.

TABLE I.

Cold pressor test (CPT), high-density surface electromyography (HDsEMG), pain-related evoked potentials (PREP).

	Experimental Protocols
	Pre		Conditioning Stimulus	Post
A.		PREP	CPT		PREP
B.	HDsEMG		CPT	HDsEMG
C.	HDsEMG	PREP	CPT	HDsEMG	PREP

C. Experimental Procedures & Data Acquisition

Experiment A.

Participants were fitted with a 32-channel EEG cap using active electrodes mounted on a stretchable fabric cap arranged after the standard 10/20 system (BrainVision, Brain Products GmbH, Gilching, Germany). Electrodes were prepared with gel to ensure an impedance below 5 kΩ.

The palmar aspect of a participants’ right fifth finger was lightly abraded to decrease impedance. A flat disc bar electrode (10 mm discs, 30 mm apart center-to-center, DDA-30, Impulse Medical Technologies, Inc., Seattle, WA, USA) with the cathode proximal, was placed on the prepared site. Electrical stimulation was delivered via stimulator with a train delay generator (Digitimer DS7A and DG2A, Digitimer North America, LLC, Fort Lauderdale, FL, USA). Pulses were delivered at a 1 Hz frequency with a 500 μs pulse duration. Prior to data collection, the stimulus intensity for three types of sensory thresholds was determined in a x3 step-up and step-down protocol, with a consistent rate of 0.2 mA/s. Participants were instructed to indicate the onset and cessation of a light tapping sensation for electrical detection sensory thresholds. The first onset of pricking pain was used for pain thresholds. The target intensity for the PREP was defined as NPRS 50/100 similar to previous PREP studies, [15, 16] and “moderate pain that could be tolerated for two bouts of two minutes”.

EEG data was collected at 2.5 kHz in a quiet room with dimmed lights. To decrease artifacts due to eye movements and blinks, participants were instructed to maintain their gaze at a fixed target on the wall. Participants were also instructed to avoid grimacing as much as possible.

Each PREP condition lasted two minutes, resulting in 120 stimuli, which is consistent with, or above the number used in recent studies. [32] The onset of each stimulus was coded as an event marker to synchronize the stimulation with the raw EEG data. Participants were instructed to rate the stimulus intensity (NPRS) at baseline and every 30 seconds.

Experiment B.

The participant placed his right lower shank in the Biodex knee flexion/extension attachment, mounted to the Systems 2 Dynamometer (Biodex Medical Systems, Shirley, NY). The knee was flexed to 150°, or 30° of flexion in reference to the anatomically neutral position. To avoid unwanted movement, an additional strap was used to restrain the distal thigh. The resulting torques were sampled at 2048 Hz and smoothed offline with a 10 Hz low-pass filter.

The target muscle for this participant was the vastus lateralis (VL). After removing excess hair, the skin was lightly abraded to decrease impedance. One HDsEMG electrode array (64 electrodes, 13 × 5, 8 mm I.E.D., GR08MM1305, OT Bioelettronica, Inc., Turin, Italy) was placed over the distal VL and one across the distal posterior thigh across the hamstring muscles to monitor for antagonist activity. Ground electrodes were placed over the lateral malleolus and the patella, and a patient reference electrode was placed on the wrist. HDsEMG was collected with a Quattrocentro signal amplifier (OT Bioelettronica, Inc., Turin, Italy.) set to monopolar mode, a 2048 Hz sampling rate, x150 amplification and a 10–500 Hz band-pass filter. [33]

Prior to data collection, the HDsEMG signal quality was visually examined to ensure optimal signal-to-noise ratios. Maximum voluntary contractions (MVC) were performed three times for both knee extension (VL) and knee flexion (hamstrings), with one minute rest periods. Additional trials were performed if the three trials were not within 10%, or if the last torque was higher than the previous.

HDsEMG experimental trials consisted of isometric knee extension ramp contractions to 30% of the MVC. For each trial, participants slowly increased the contraction intensity at 3% MVC/s over 10 seconds, and then gradually decreased the contraction intensity at the same rate until they were relaxed. During each trial, participants had a screen in front of with a guideline for the target ramp and received real-time feedback of their torque output. Each trial included five seconds of baseline recordings before and after the triangle. Each HDsEMG condition (pre and post) consisted of three trials with 10 second rest periods between recordings.

Experiment C.

This experiment combined procedures from A. and B., starting with skin preparation of the fifth digit and electrode sites, and baseline sensory threshold testing. Afterwards, the participant was fitted with the EEG cap, and electrodes were gelled. For this participant, we chose the tibialis anterior as the target muscle for HDsEMG recordings, to test a more distal muscle and to further test our hypothesis of increased widespread monoaminergic drive in response to the CPT. We placed another HDsEMG electrode over the medial aspect of the soleus to monitor for antagonist activity. The participant placed his right foot in the Biodex footplate attachment, mounted to the Systems 2 Dynamometer (Biodex Medical Systems, Shirley, NY). The knee was fully extended, with the hips at 100° flexion and his right ankle at 100°, or 10° anatomical plantarflexion. In order to prevent loss of force transfer during testing, Velcro straps were placed around the foot. [33]

Once optimal signal quality of EEG and HDsEMG was ensured, MVC was performed for ankle dorsiflexion (tibialis anterior) and plantarflexion (soleus) as described above. The experimental trials consisted of dorsiflexion ramp contractions to 30% of its MVC with otherwise identical parameters. All other procedures were identical to experiments A. and B., except that both HDsEMG and PREP were performed before and after the CPT (see TABLE I).

D. Data Analysis

Pain-related evoked potentials (PREP)

Signal analysis of PREP was performed using the FieldTrip toolbox in Matlab (Matlab, The MathWorks, Inc., Natick, MA, USA). [34] The recordings were segmented into epochs from −200 ms to 800 ms, with 0 representing the stimulus onset. Epochs were then baseline corrected (-200 ms to 0 ms) and an 80-sample moving average filter was applied. The resulting signals were averaged across epochs at each selected electrode.

Our primary measure was the peak-to-peak amplitudes between the second (N2P2) negative and positive waves. Due to the small sample size, we used paired t-tests separately for each data from each individual channel, as well as for mean values of channels across participants to compare pre- and post-CPT conditions. Alpha values were set at 0.05 for statistical significance.

High-density surface electromyography (HDsEMG)

The HDsEMG signals were manually inspected and channels with substantial artifacts or noise were removed. Decomposition of MU spike trains was performed with a convolutive blind-source separation algorithm. [35] Once decomposed, manual editing was performed to correct decomposition errors using a custom-made graphical user interface in Matlab. For each MU spike train the instantaneous discharge rate was calculated as the inverse of the interspike interval and then smoothed using support vector regression. [23]

The descending monoaminergic effects on MU discharge patterns were estimated using ΔF (i.e., ΔFrequency), which quantifies the difference in discharge rate of a higher-threshold (test) MU relative to the discharge rate of a lower-threshold (reporter) MU. [36] This difference is also known as onset-offset hysteresis. [37] DeltaF values were averaged as ‘unit wise’ values, where each test unit receives a ΔF value based on a lower threshold reporter unit, and then the resulting ΔF values for that unit are averaged to give one ΔF per test unit. [23] Additionally, we used the brace height index (BH), a recently developed method by Beauchamp et al., which quantifies the deviation from linear discharge, as well as components of the rate of discharge changes. [38] As with PREPs, paired t-tests were used for ΔF and BH for pre- and post-CPT conditions within individual, and across participants.

Comparative analysis and directionality

To normalize PREP, DF, and BH values for more direct comparison, and to evaluate the directionality of changes pre- and post-CPT, we calculated indices reflecting biophysiological changes following the CPT, termed ΔCPT. We added PREP, ΔF, and BH subscripts accordingly. The indices vary between −1 and 1, where 0 equals no change, or neuromodulation. Since improved inhibition of nociceptive signal transmission can generally be considered a positive, values greater than 0 indicate a decrease in N2P2 peak-to-peak amplitude following the CPT (ΔCPT_PREP, see equation below). Negative values, indicate facilitated nociceptive signal transmission, as seen by a larger N2P2 peak-to-peak amplitude post-CPT.

$${\Delta \text{CPT}}_{\text{PREP}}=\frac{\left(\text{N}2\text{P}2\text{PreCPT}-\text{N}2\text{P}2\text{PostCPT}\right)}{\left(\text{N}2\text{P}2\text{PreCPT}+\text{N}2\text{P}2\text{PostCPT}\right)}$$

The indices for ΔF (ΔCPT_ΔF) and BH (ΔCPT_BH) follow the same principle, with the exception that positive values indicate facilitated motor function, whereas negative values indicate inhibited motor function.

$${\Delta \text{CPT}}_{\Delta \text{F},\text{BH}}=\frac{\left(\Delta \text{F},\text{BH}\text{PostCPT}-\Delta \text{F},\text{BH}\text{PreCPT}\right)}{\left(\Delta \text{F},\text{BH}\text{PreCPT}+\Delta \text{F},\text{BH}\text{PostCPT}\right)}$$

Consequently, positive values suggest an increase in descending monoaminergic drive related to DNIC and MU firing patterns, as well as a parallel relationship between these two systems.

III. Results

A. Bioelectrical signals

Table II. summarizes PREP data for Experiments A. and C., including the raw N2P2 peak-to-peak amplitude difference, as well as the DCPTPREP for channels C1, C2, and C3. All participants demonstrated decreased N2P2 amplitudes post-CPT, indicating a decreased nociceptive signal magnitude reaching the cortex. Despite the small sample size, channel C3 reached significance (p = 0.0032) with the paired t-test. Channels C1 and C5 did not reach significance with p = 0.1402 and p = 0.0776, respectively.

TABLE II.

PREP data from experiments A. and C.

		N2P2 Amplitude
Participant	Channel	preCPT	postCPT	Pre - Post	ΔCPTPREP
A.1	C1	1.5616	1.3489	0.2126	0.0731
	C3	4.8223	3.9415	0.8808	0.1005
	C5	8.6516	6.0979	2.5536	0.1731
A.2	C1	3.0005	2.2414	0.7591	0.1448
	C3	6.4214	5.1959	1.2255	0.1055
	C5	10.2365	8.8864	1.3501	0.0706
A.3	C1	2.3380	1.6312	0.7069	0.1781
	C3	8.6174	7.0893	1.5281	0.0973
	C5	14.1243	14.0435	0.0809	0.0029
C.	Cl	7.9744	5.2738	2.7006	0.2038
	C3	11.3513	9.8835	1.4678	0.0691
	C5	17.5704	16.2235	1.3469	0.0399

Data for HDsEMG data from experiments B. and C. are summarized in Table III. Both participants demonstrated positive values for ΔF and BH. However, neither reached significance with the paired t-test (ΔF p = 0.2043, BH p = 0.2178).

TABLE III.

DF and BH data for experiments B. and C.

Participant	Metric	preCPT	postCPT	Post - Pre	ΔCPT
B.	ΔF	2.6760	3.1850	0.5090	0.0868
	BH	30.2120	34.7000	4.4880	0.0691
C.	ΔF	2.4774	3.4932	1.0158	0.1701
	BH	27.7860	37.2400	9.4540	0.1454

IV. Discussion

This study serves as a preliminary investigation into a combined EEG and HDsEMG protocol to assess the effects of an experimental painful stimulus—the cold pressor test (CPT)—on nociceptive and motor unit (MU) neuromodulation. We approached this study in three experimental components (A, B, and C) to examine the effects of the CPT on pain-related evoked potentials (PREP), MU neuromodulation, and the combined response of both. While other studies employing HDsEMG typically collect more data to increase MU yield, the transient nature of the CPT-induced diffuse noxious inhibitory control (DNIC) effect required a shorter experimental window. To accommodate this time limitation, we conducted three ramp contractions with short rest periods in both pre- and post-CPT conditions, totaling approximately two minutes. This duration was selected to align with the CPT and PREP protocols for consistency.

Although our sample size was small, and changes in firing rate variability (ΔF) and motor unit behavior (BH) did not reach statistical significance, the positive trends in their respective ΔCPT values suggest the feasibility of this approach. These preliminary findings support our hypothesis that in healthy, young controls, the CPT would enhance both nociceptive inhibition (DNIC) and MU facilitation due to increased monoaminergic drive. Further experiments with a larger sample size are planned to strengthen these findings.

It is also important to note that 5-HT plays a complex role in descending modulation, potentially contributing to nociceptive facilitation in cases of aberrant processing, such as that observed in chronic pain conditions. Future studies should expand this protocol to include participants with various chronic pain conditions. This would help determine whether the protocol can reliably capture both nociceptive inhibition and facilitation and clarify how these processes relate to MU neuromodulation.

Beyond the limited sample size, another notable limitation of this study is the absence of female participants. Future investigations will need to address this imbalance, as sex-related differences in nociceptive and motor unit responses may provide valuable insights.

Clinical Relevance—

This provides the first step towards furthering our understanding of the interplay between nociceptive processing and motor impairments, two of the hallmark signs across chronic pain conditions.