Optimum electrical sinusoidal frequency stimulation to activate C-nociceptors

Electrical sinusoidal stimulations between 3 and 8 Hz are better suited to activate human and porcine C-nociceptors than stimuli of higher frequency (>10 Hz).

Abstract

Introduction:

Electrical sinusoidal stimulation has been shown to reliably activate C-fibers. Here, we explored the optimum sinusoidal frequency for C-nociceptor responses.

Objectives:

Identify the optimum sinusoidal frequency to activate single C-nociceptors in pigs and to provoke pain and axon reflex flare in humans.

Methods:

Single C-fibers (n = 60) were recorded from pig saphenous nerves in vivo in response to transcutaneous sinusoidal electrical stimulation delivered at frequencies of 3 to 100 Hz with increasing amplitudes. In 22 human subjects, 30 sinusoidal pulses were delivered with 3 to 20 Hz at 0.0125 to 0.4 mA amplitudes and perception thresholds, evoked pain (numeric rating scale, 0–10) and axon reflex flare (full-field laser perfusion imaging) were monitored in 4 seconds intervals.

Results:

Single C-nociceptors (36 “polymodal” [C-mechano-heat-sensitive], 15 “very high threshold, and 9 “silent” [C-mechano-insensitive]) were activated more easily at frequencies below 8 Hz as compared with frequencies >10 Hz. In humans, pain and flare responses were also found to be frequency dependent. Pain threshold assessments were particularly sensitive at a frequency of 4 Hz. Highest pain ratings were recorded from 3 to 6 Hz at stimulation amplitudes of 0.2 mA. Significant axon reflex flare responses developed at low amplitudes of 0.05 and 0.1 mA at frequencies of 4 and 8 Hz only.

Conclusion:

Sinusoidal frequencies between 3 and 8 Hz are better suited to activate C-nociceptors than stimuli above 10 Hz and are expected to facilitate C-nociceptor recruitment for their functional assessment in volunteers and chronic pain patients.

1. Introduction

Rectangular stimulation is used traditionally for controlled transdermal electrical skin nerve fiber activation in humans. The recruitment of primary afferent nerve fiber subtypes thereby depends on the grade of myelination of the axon. Large-caliber neurons (A-fibers) have very low electrical activation thresholds whereas unmyelinated neurons (C-fibers) are characterized by higher thresholds. High electrical activation thresholds may be a limiting factor for C-nociceptors stimulation, in particular for its assessment in pain patients. Moreover, A-fiber activation by electrical rectangular pulses will also reduce spinal processing of C-fiber input (“gate-control”). Changing the electrical stimulation profile to a sinusoidal waveform improves the transdermal neuronal activation relative to the conventional rectangular pulses.^5,9 Changing the electrode type from standard patch-electrodes to small diameter pin electrodes has also been shown superior to activate nociceptive fibers.^8,17 In addition, parameters of pulse duration and pulse frequency of transcutaneous electrical stimulation have been shown to influence the strength of neuronal activation.^{7,10,13,15,23,28,32}

Altering the waveform of the electrical stimulus combined with a pin-pointed electrode configuration for transdermal application provides selective stimulation of unmyelinated C-nociceptors^11,12,26 and allows the assessment of C-fiber function in eg, patients with neuropathic pain^12,16 or chronic itch.²⁵ Explorative animal in vitro and in vivo studies confirmed the efficacy of 4 Hz sinusoidal waveform stimulation to electrically activate C-nociceptors with a minor contribution of A-delta nociceptors.^12,21 Thereby, both “polymodal” (heat- and mechanically sensitive, CMH) and “silent” (mechanically insensitive, CMi) C-nociceptors are stimulated, of which the latter also contribute to the development of a localized skin reddening (“axon reflex flare”) surrounding the stimulation site.²⁷

In this study, we systematically varied the frequency and intensity of sinusoidal electrical stimulation for C-nociceptor activation in identified single nerve fibers of the pig. We used the same approach in humans to identify the optimum frequency for the induction of pain and axon reflex flare using psychophysics and full-field laser perfusion imaging (FLPI) for the objective assessment of “silent” C-nociceptor activation.

2. Methods

2.1. Animals

Male Domestic German Landrace pigs (Sus Scrofa domesticus) 12 ± 4 week old with an average weight between 20 and 25 kg were generally anesthetized, monitored for cardiorespiratory function and temperature, and maintained under volume-controlled mechanical ventilation. After cleaning and disinfection, an incision was made at the ventral mid-tight and 6 cm of the saphenous nerve exposed for recordings. The studies obtained ethical approval for animal experimental procedures by the Ethics committee of the regional government (Karlsruhe, Baden-Wuerttemberg, Germany), Approval Number G-78/18, and were performed within the Medical Faculty Mannheim of Heidelberg University.

2.1.1. Pig single nerve fiber recordings

Approximately 1 cm of a fascicle was carefully separated from the saphenous nerve under visual control using a microscope (LEICA M320 Full HD; Wetzlar, Germany), desheathed, and fine strands were splatted from the nerve fascicle according the teased fibre technique and placed on a gold wire electrode for recording.^18,20 The receptive field of the afferent nerve fibres was localized by scratching and squeezing slightly the hind limb skin, evoking a characteristic discharge. Within this receptive field, 2 noninsulated microneurography electrodes (FHC Inc, Bowdoin, ME) were attached to the skin and delivered by means of a constant current stimulator (DS7A; DIGITIMER Ltd, Welwyn Garden City, United Kingdom) electrical 20 mA pulses (rectangular shape, 0.5 ms width, 0.25 Hz). The electrodes were inserted intradermally at a site where a single action potential was elicited and appeared repeatedly in the recording window (DAPSYS 8.0 software package) with a fixed retention time (time-locked action potential). Signals were preamplified (Bio-Signalverstärker; AvereSolutions, Erlangen, Germany), filtered, and amplified (bandwidth 100–3000 Hz, Model 3364; Krohn-Hite Corp, Brockton, MA) for discrimination and latency measurements (DAPSYS 8.0 software package; Brian Turnquist, Bethel University, MN). The different C-fibre classes were classified according to their responsiveness to mechanical, heat, and electrical stimulation^20,32 into mechano-thermal–sensitive C-nociceptors (“polymodal”, CMH), very high-threshold (>150 mN) mechano-thermal–sensitive nociceptors (VHT), and mechano-insensitive (>600 mN) C-nociceptors (“silent”, CMi). For transcutaneous electrical sinusoidal stimulation, a pair of L-shaped blunted bipolar platinum–iridium electrodes (diameter 0.4 mm separated by 2 mm, Cephalon, Netherlands) was attached to the skin and inside the receptive field of the unit. Sinusoidal electrical stimuli were generated by a constant current stimulator (A395; WPI, Worcester, MA) and delivered at frequencies from 3 to 100 Hz (3–4–6–8–10–20–50–100 Hz). Duration of sinusoidal stimuli were 4 seconds (3 Hz), 3 seconds (4 Hz), 2 seconds (6 Hz, 8 Hz), and 1 second (10–100 Hz). Sinusoidal amplitudes were applied in increments of 0.01 mA and in 5 seconds intervals (controlled by DAPSYS 8.0 software package) until a response of at least 2 action potentials of the unit were recorded, which was defined as threshold amplitude of activation.

2.2. Human psycho-physics and axon reflex flare assessment

The Ethic Committee II of the Medical Faculty Mannheim at the University of Heidelberg approved the study protocol and experimental procedures. Twenty-two sex-matched subjects (age 35 ± 14 years) participated in the study. Two experimental sessions were performed for: (1) psycho-physical frequency-amplitude nociceptor stimulation assessment; (2) FLPI frequency-amplitude axon reflex flare development measures. In each session, 30 sinusoidal pulses were delivered within a time period of 10 seconds. For transdermal stimulation, we placed a pair of blunt pin platinum wire electrodes of 0.4 mm diameter, positioned 2 mm apart from one another and mounted in an applicator printed on a 3D-printer, onto the subjects' volar forearm. Sinusoidal pulses were delivered by a constant current stimulator (DS5; Digitimer Ltd, Welwyn Garden City, United Kingdom) connected to a Digital-Analogue Converter (DAQ) (NI USB-6221; National Instruments, Austin, TX). The frequency and amplitude of the applied pulses were controlled by custom-written software (DAPSYS 8; Brian Turnquist, Bethel University, MN) defining the time required for one completed sinusoidal cycle plus the time interval between each cycle to match the demanded stimulation frequency needed to deliver 30 sinusoidal pulses within 10 seconds (Fig. 1).

Figure 1.

Sinusoidal stimulation protocol. Diagram of the 30 sinusoidal pulses delivered at frequencies of 3–20 Hz for a 10-second period. Pulses were applied with amplitudes of 0.0125–0.4 mA. Note that the sinusoidal frequency was defined by the intersinusoidal interval required to deliver the 30 pulses within the 10-second period, as depicted in the right panel. One sinusoidal cycle was applied within 0.33 seconds with the peak-to-peak time indicating the delivered sinusoidal frequency. A second sinusoidal cycle is shown inside the dashed box. Sinusoidal pulses were delivered with ascending intensities at random sinusoidal frequency.

2.2.1. Session 1

2.2.1.1. Psycho-physical assessment of frequency-amplitude sinusoidal nociceptor stimulation

In this trial, we randomly chose a sinusoidal frequency of 3–4–6–8–10–20 Hz and delivered electrical stimuli for a duration of 10 seconds and with increasing amplitudes, from lowest (0.0125 mA) to the highest (0.4 mA). All amplitudes (0.0125–0.025–0.05–0.1–0.2–0.4 mA) were investigated for each frequency. Subjects were instructed to indicate when they first perceived a sensation (perception threshold) and when the stimulus was felt painful (pain threshold) on a numeric rating scale (NRS, end points 0 [no pain] and 10 [maximum pain imaginable]). On further increase of the stimulation amplitude, subjects should finally indicate when the applied stimulus was felt as painful NRS 3 (suprathreshold pain). A maximum amplitude of 0.4 mA was applied and not further increased, even if subjects' pain was reported NRS <3. The test series (all frequencies tested for all amplitudes) was repeated and averaged values calculated for the statistics.

2.2.2. Session 2

2.2.2.1. Full-field laser perfusion imaging of sinusoidal frequency-amplitude vasodilation

The development of an axon reflex flare surrounding the stimulation site was recorded real time with full-field laser perfusion imaging (FLPI-2; Moor Instruments Ltd, Axminster, United Kingdom; Fig. 2). The scanner head was mounted perpendicular to the volar forearm at a distance of 25 cm to the skin surface. The subject's forearm was rested on a table and in a vacuum cushion (HEK medical GmbH, Ascheberg, Germany) to avoid movement. The zoom lenses focused images covered a 10 × 10-cm skin surface area. Images were recorded with a resolution of 750 × 580 pixels at an acquisition rate of 100 Hz (exposure time 20 ms) and captured in 4 seconds intervals (0.25 Hz).

Figure 2.

Specimen of full-field laser perfusion images for axon reflex recording. Full-field laser perfusion images recorded from a left female forearm skin during 0.1 mA electrical 4 Hz sinusoidal stimulation (flux image top left and simultaneously captured photograph image top right). A handheld bipolar electrode was used to deliver the transcutaneous electrical stimuli. The corresponding axon reflex flare is depicted in the bottom panel at 8 seconds (left) and 20 seconds (right) after stimulation.

Sinusoidal stimuli were delivered with increasing amplitudes for each frequency of 3–4–6–8–10–20 Hz. Each test was performed at a different skin site, ie, frequencies were assessed at randomly chosen distal–medial–proximal skin area of the left and right volar forearm. The FLPI-sequence started with the recording of 3 baseline scans plus an additional scan in which a rectangular cardboard (4 cm²) was placed for standardized area calibration. The baseline recording was halted, the 10-second sinusoidal stimuli were delivered with lowest amplitude (0.0125 mA), and the FLPI recordings continued to capture 10 scans requiring 40 seconds. Thereafter, FLPI-recording was halted again, the next highest current amplitude delivered (0.025 mA), and FLPI-recordings continued again for 40 seconds capturing another 10 scans. This process was repeated with incrementing sinusoidal current amplitudes up to a maximum of 0.2 mA. If at this stimulation amplitude, no erythema development was clearly visible online, a sinusoidal amplitude of finally 0.4 mA was applied (n = 9 subjects). After either 0.2 or 0.4 mA stimuli, FLPI-recordings were continued for 3 minutes to confirm the presence of a persistent vasodilation.

Full-field laser perfusion imaging-images of each sequence were analyzed offline using dedicated software (Moor FLPI-2 V1.1 software package). A cutoff blood flux value was determined in the baseline scans of each sequence to discriminate a transient blood flow increase evoked by the electrical sinusoidal stimulation. Mean blood flux values plus 2-fold standard deviation were defined as cutoff threshold and calculated at baseline condition around the skin site of intended sinusoidal stimulation. The number of contiguous pixels exceeding this baseline threshold value of skin blood flow perfusion was calculated for each scan after stimulation and dedicated to the area of axon reflex flare development according to the number of pixels identified within the rectangular cardboard calibration area. The average size of the axon reflex flare was calculated across the 10 images recorded during the 40-second scan. To avoid any skin blood flow artefact immediately after electrical stimulation, which potentially could have been evoked by the mechanical pressure due to the placement or withdraw of the electrode from the skin, the axon reflex flare analysis of each FLPI sequence was performed 8 seconds after stimulation offset and electrode removal.

During the protocol of FLPI-monitoring, we also recorded the pain intensity perceived by the subject in response to the sinusoidal stimuli. Similar to the assessment in session 1, volunteers were instructed to estimate the magnitude of pain on the NRS (end points 0 and 10) during the 10-second period of sinusoidal stimulation.

2.3. Statistics

Data were analyzed using STATISTICA 7.1 software package (StatSoft Inc, Tulsa, OK) with the Shapiro-Wilk test to evaluate normality and analysis of variance, including the factorial groups “frequency,” “amplitude,” and “nerve fiber class.” P-values <0.05 were considered significant. Graphical representations were performed using GraphPad Prism version 10.4.3 (GraphPad, Boston, MA). Single nerve fibre action potentials of CMH-, VHT-, and CMi-nociceptors were analyzed with DAPSYS 8.0 software (DAPSYS; Brian Turnquist, Bethel University, St Paul, MN) and sinusoidal current amplitudes (mA) needed to evoke at least 2 action potentials determined for each frequency. Identified amplitude values additionally were normalized to the grand mean stimulation amplitude delivered across all frequencies. Data are depicted as mean ± SEM.

Data of experiments performed in humans (psycho-physics and FLPI-assessment) were analyzed by analysis of variance and Bonferroni post hoc test within the factors “frequency” and “amplitude. Spearman rank correlation analysis was used to identify a stimulus amplitude-dependent correlation between pain NRS and FLPI axon reflex flare. Data are depicted as mean ± SD.

3. Results

3.1. Single nerve fiber recordings in pig

A total of 36 “polymodal” CMH-, 15 VHT-, and 9 “silent” CMi-nociceptors were recorded and analyzed for the thresholds of sinusoidal current amplitudes required to evoke at least 2 action potentials during each of the tested frequency of 3–4–6–8–10–20–50–100 Hz. We observed a significant difference of sinusoidal activation thresholds between the 3 fiber classes (ANOVA F [2, 34] = 5.89, P < 0.0001) in the order CMH < VHT < CMi nociceptors (Fig. 3A). Activation thresholds increased with increasing frequency, in particular for sinusoidal frequencies above 10 Hz (Fig. 2A). In general, lowest sinusoidal amplitudes were required for C-nociceptor activation at 3 Hz and highest at 100 Hz (ANOVA F [7, 252] = 12.4, P < 0.0001). No interaction, however, was identified between the factorial groups “fiber class” and “sinusoidal frequency” (2-way ANOVA F [14, 238] = 0.5, P > 0.94).

Figure 3.

Single nerve fiber recordings in pig in vivo. Threshold amplitudes of sinusoidal stimuli (mA) were delivered at frequencies of 3 to 100 Hz to induce at least 2 action potentials in C-mechano-heat-sensitive “polymodal” (CMH) (n = 36), very high threshold (VHT) (n = 15), and C-mechano-insensitive “silent” (CMi) (n = 9) nociceptors (A). Nociceptors seem to be activated at lower current intensities for frequencies below 6 Hz. “Polymodal” CMH nociceptors have lower sinusoidal activation thresholds than VHT and “silent” (CMi) C-nociceptors. Delta activation threshold amplitudes to the applied grand mean amplitude (B). Values below the dotted line indicate lower amplitudes than the grand mean and vice versa. Significantly higher amplitudes were needed to stimulate C-nociceptors with sinusoidal frequencies of 20 to 100 Hz. Data are presented as mean ± SEM, asterisk indicate P < 0.05 (analysis of variance).

To establish a better comparison of the frequency-dependency for sinusoidal activation in each nociceptor class, data of sinusoidal activation thresholds were normalized by subtracting the grand mean value calculated across all frequencies from the stimulation amplitude recorded at a given frequency. Thereby, negative values represent lower activation thresholds as compared with the overall average amplitudes (Fig. 3B). We observed a significant frequency-dependency of the sinusoidal amplitude required for single nerve fiber activation (ANOVA F [7, 245] = 11.742, P < 0.0001). This was independent the recorded nociceptor class (ANOVA F [2, 33] = 0.0031, P > 0.9) and without interaction between the factors “fiber class” and “sinusoidal frequency” (2-way ANOVA F [14, 231] = 0.477, P > 0.9). At sinusoidal frequencies lower than 8 Hz, the current amplitudes required to excite pig nociceptors were below the average current needed for excitation and significantly smaller as compared with the stimulus activation thresholds recorded at 20 to 100 Hz (Fig. 3B).

3.2. Pain and axon reflex erythema in volunteers

Two separate sessions were performed to assess the response to a sinusoidal frequency-amplitude–dependent nociceptor activation in humans.

3.2.1. Session 1

3.2.1.1. Psycho-physics

In our psycho-physical assessment of sinusoidal frequency-amplitude stimulation, we applied 10-second sinusoidal stimuli at a defined amplitude (from lowest 0.0125 mA to highest 0.4 mA) and delivered at randomly selected frequencies of 3–4–6–8–10–20 Hz. Irrespective of the frequency, approximately 80% of the subjects did not perceive the sinusoidal stimulation at 0.0125 mA amplitudes, intensities of 0.025 mA were not felt by approximately 20%, and an amplitude of 0.5 mA evoked in almost all subjects slight pain (Fig. 4). A further increase of the sinusoidal amplitude significantly enhanced evoked pain (ANOVA, F [5, 105] = 81.3, P < 0.0001).

Figure 4.

Individual NRS recording during sinusoidal frequency-amplitude–dependent stimulation. Pain NRS (numeric rating scale, 0–10) of n = 22 subjects to 30 transcutaneous electrical pulses of 0.0125 to 0.4 mA amplitudes applied at frequencies of 3 to 20 Hz (A–F). The horizontal line indicates the mean NRS of all subjects. Values below the dotted line (NRS 0) indicate that the stimulus was not felt. Most subjects perceived the stimuli at 0.05 mA amplitudes as slightly painful (NRS > 0) and increasing the stimulus amplitude increased the pain NRS. NRS, numeric rating scale.

Pain intensity was significantly dependent on the delivered frequency of sinusoidal pulses (ANOVA, F [5, 105] = 10.1, P < 0.0001), with a strong interaction between the factors “sinusoidal frequency” and “stimulus amplitude” (2-way ANOVA F [25, 525] = 2.02, P < 0.003). Of note, lower pain ratings were reported at suprathreshold stimulus amplitudes of 0.1 to 0.4 mA, when pulses were delivered with frequencies exceeding 6 Hz (Fig. 5A). For example, pain NRS to 0.4 mA was significantly lower at sinusoidal pulses of 20 Hz when compared with pain NRS at a frequency of 3, 4, or 6 Hz (NRS 4.3 ± 0.5 vs NRS 4.7 ± 0.4, Bonferroni post hoc P < 0.05, Fig. 4A). Lower pain ratings at higher sinusoidal frequency was also apparent after normalization of the pain ratings to the NRS grand mean values (Fig. 5B). At sinusoidal amplitudes of 0.1 to 0.4 mA, pain ratings at sinusoidal frequencies of 3 to 6 Hz were clearly above the grand mean values, whereas pain ratings at frequencies of 8 to 20 Hz were below the grand mean.

Figure 5.

Psycho-physics of sinusoidal frequency-amplitude–dependent pain. Magnitude pain NRS (A) and delta grand mean NRS (B) to 30 sinusoidal stimuli delivered at frequencies of 3 to 20 Hz and with amplitudes of 0.0125 to 0.4 mA (n = 22 subjects). Pain intensity was significantly dependent on the applied sinusoidal frequency and amplitude (2-way analysis of variance interaction P < 0.003) (A), particularly at 20 Hz when compared with 3–6 Hz (marked by asterisks, Bonferroni post hoc P < 0.05). Normalization of pain NRS to the delta grand mean NRS (dotted line) (B) revealed significantly higher pain ratings to suprathreshold stimuli (amplitudes 0.1 to 0.4 mA) at frequencies of 3 to 6 Hz and lower pain NRS at a frequency of 20 Hz. Data are presented as mean ± SEM. NRS, numeric rating scale.

3.2.2. Session 2

3.2.2.1. Full-field laser perfusion imaging

Similar to session 1, we delivered 10-second sinusoidal pulses at frequencies from 3 to 20 Hz. For each frequency, stimuli were applied with increasing amplitudes from lowest (0.0125 mA) to highest (0.2 mA), to identify the minimum stimulus amplitude required for an axon reflex flare. In few subjects, an amplitude of 0.2 mA did not evoke an apparent flare reaction (n = 9), and thus, 0.4 mA was finally administered. Increasing the sinusoidal amplitude evoked an enhanced flare development at any stimulation frequency (ANOVA F [4, 424] = 141.8, P < 0.0001). A vasodilation area exceeding 1 cm² was observed at sinusoidal amplitudes of 0.1 mA and higher (Fig. 6A). The flare reaction persisted for at least 3 minutes and covered an area of 2.7 ± 0.3 cm² at 0.2 mA (n = 13) and 3.8 ± 0.2 cm² at 0.4 mA (n = 9). We observed a trend that high sinusoidal frequency stimulation caused smaller flare responses when compared with stimuli of lower frequencies (Fig. 6B), and identified a significant interaction between the factors “sinusoidal frequency” and “stimulation amplitude” (2-way ANOVA F [20, 404] = 1.63, P < 0.05). In particular, at an amplitude of 0.2 mA, the flare reaction at 20 Hz (1.7 ± 0.2 cm²) was smaller as compared with 3 Hz stimulation (3.2 ± 0.2 cm², Bonferroni post hoc, P < 0.03). In addition, vasodilation after 0.4 mA stimulation was smaller at 20 Hz compared with 3 Hz stimuli (3.2 ± 0.6 cm² vs 5.2 ± 0.7 cm², n = 9), but this difference was not significant (ANOVA F [5, 53] = 0.82, P > 0.5).

Figure 6.

Full-field laser perfusion imaging of the axon reflex flare area. Average values of the axon reflex flare area (cm²) recorded with full-field laser perfusion imaging across 40 seconds and in response to a 10-second sinusoidal stimulation period. Sinusoidal stimuli delivered at frequencies of 3 to 20 Hz evoked a stimulus amplitude-dependent increase of vasodilation (analysis of variance, P < 0.0001) (A) and a significant interaction between “sinusoidal frequency” and “stimulation amplitude” (2-way analysis of variance, P < 0.05). Electrical stimulation with a frequency of 20 Hz evoked smaller flare responses, particularly as compared with 3 Hz stimuli (Bonferroni post hoc, P < 0.03) (B). Note that a sinusoidal amplitude of 0.4 mA was applied in n = 9 subjects only. Data are presented as mean ± SEM.

3.2.3. Correlation of psycho- physics with full- field laser perfusion imaging axon reflex flare

During session 2, we also assessed pain levels in response to the sinusoidal stimulation and correlated the NRS values with the evoked axon reflex area assessed by FLPI (Fig. 7). In line with the results reported above, increasing the stimulation amplitude enhances the evoked pain (see session 1) and also increases the corresponding flare response. Pain and flare significantly correlated for each delivered sinusoidal frequency, with the strongest interaction found at 8 and 10 Hz (Spearman rank correlation r = 0.72, P < 0.05) and weaker correlation at 4 Hz stimulation (r = 0.58, P < 0.05).

Figure 7.

Correlation between pain NRS and FLPI axon reflex flare area. Spearman rank correlation between electrical sinusoidal pain NRS (max. 10) and the corresponding FLPI axon reflex flare (cm²) evoked by stimulation amplitudes of 0.0125 to 0.4 mA. A significant correlation was identified at each stimulus frequency of 3 to 20 Hz (r = 0.58 to 0.72, P < 0.05). Note that a sinusoidal amplitude of 0.4 mA was applied in n = 9 subjects only. Numeric rating scale values below 0 indicate that subjects did not perceive the stimulus, NRS values of 0 indicate that the stimulus was felt but not painful. FLPI, full-field laser perfusion images; NRS, numeric rating scale.

4. Discussion

In this study, we identified the optimum sinusoidal frequencies for electrical activation of C-nociceptors in pig in vivo. In addition, we aimed to outline sinusoidal stimulation parameters best suited to activate C-nociceptors in humans, and recorded therefore pain and axon reflex flare responses. As rationale, this knowledge may be transferred into the clinics for a facilitated nociceptor assessment in pain patients.

4.1. Single nerve fiber recordings in pig

During the single nerve fiber recordings of C-nociceptor classes, it seemed that “polymodal” (HT-) fibers show lower electrical sinusoidal activation thresholds than “very high threshold” (VHT-) and “silent” C-fibers. This finding is in line with our previous result observed on transdermal (pig) and intradermal (human) electrical sinusoidal stimulation at 4 Hz.¹² Obviously, the threshold amplitude of the electrical pulse required to stimulate C-nociceptors depends on its proximity to the nerve fiber ending or the axon. Thus, it could be argued that HT-fibers are located more superficially than VHT- or “silent” nociceptors, and should therefore be activated more easily by transdermal electrical stimuli (particularly considering that current spread to deeper tissue declines exponentially²⁴). However, this aspect cannot be solved as long as there are no specific markers available to stain for C-nociceptor subtypes to validate potential differences in depth of their epidermal branches.

The activation of C-nociceptors was dependent on the delivered sinusoidal frequency, with higher amplitudes required for action potential initiation at higher frequencies. This increase of activation threshold was independent of the recorded C-nociceptor class. By contrast, on stimulation with intradermally delivered electrical rectangular pulses, a distinct following-frequency stimulation pattern was seen between C-nociceptor classes in pig³² and can be used for a distinguished discrimination of C-fibers. This stimulation frequency-dependent excitation profile is most likely related to unique channel activation/inactivation kinetics of voltage-gated sodium channel subtypes expressed on C-nociceptors.³³ Whether this also holds true for our electrical sinusoidal stimulation remains open. With higher frequency of sinusoidal stimulation at same amplitudes, the profile of each cycle comes closer to the shape of a rectangular pulse. It therefore would be interesting to analyze at what phase of the sinusoidal cycle the action potential occurs, which determines the time of charge accumulation and action potential generation,²¹ to identify whether this varies also depending on the sinusoidal frequency.

Frequencies from 3 to 8 Hz allow accumulation of charge for approximately 60 to 160 milliseconds. Within this time window, passive and active components may facilitate C-fiber activation. A passive component would be driven by the time constant of the C-fiber membrane, which is around 100 milliseconds and does not vary between mechano-sensitive and -insensitive human nociceptors.² An active component that would facilitate C-fiber activation in a time window between 60 and 160 milliseconds includes the expression of voltage-dependent sodium channel NaV1.7 that can produce a ramp current on slow depolarization <100 milliseconds.³ Another contribution could be mediated by the hyperpolarizing part of our sine wave stimulation causing an inward current (I_h) generated by hyperpolarization-activated cyclic nucleotide-gated channels.⁶ These channels are known to provide pacemaker depolarizations during rhythmic-oscillatory activity, for instance in the thalamus at around 0.5 to 4 Hz,²² and could be activated by the sinusoidal hyperpolarizing cycle, thereby facilitating the ensuing depolarization phase. However, as we did not include pharmacological interventions during the recordings, we cannot clarify the underlying mechanism of an optimized sinusoidal frequency stimulation.

4.2. Psycho-physics and axon reflex flare assessment in humans

We did not observe a clear difference of pain thresholds between the stimulation frequencies as subjects reported slight pain to an amplitude of 0.05 mA (pain threshold) and amplitudes of 0.1 mA were clearly suprathreshold painful for all frequencies tested. However, at higher amplitudes, sinusoidal frequencies exceeding 8 Hz were perceived as less painful (most prominent at 20 Hz) when compared with lower frequencies of 3 to 6 Hz. One might be tempted to attribute this finding to C-nociceptors being less capable to follow high-frequency discharges,²⁹ particularly for “silent” C-fibers that have a characteristic conduction block for prolonged stimulation of 5 Hz,³² but this obviously cannot be validated psycho-physically.

Our axon reflex flare measures objectively assessed the activation of “silent” C-nociceptors.²⁷ We demonstrated previously a profound skin blood flow erythema development on electrical sinusoidal stimulation at 4 Hz (50 cycles) and 0.2 mA amplitudes,¹² but found here with FLPI recordings, which enables instantaneous vasodilation assessment with high temporal resolution,⁴ an axon reflex flare development occurring already after 30 cycles of 0.1 mA amplitude. Increasing the stimulation amplitude to 0.2 mA enhanced the flare area, and this vasodilation persisted for several minutes indicating the release of calcitonin gene-related peptide from terminal branches of activated nociceptor endings.³¹ At this amplitude, the flare reaction was significantly smaller at 20 Hz stimulation when compared with a lower sinusoidal frequency of 3 Hz, which may reflect reduced recruitment of “silent” C-nociceptors at higher frequencies, also observed in our single nerve fiber recordings in the pig.

Correlation analyses indicated a strong link between FLPI flare responses and corresponding pain ratings for each sinusoidal frequency. This result suggests on one hand that the increasing efficiency of nociceptor activation at any frequency of sinusoidal stimulation is similarly effective for “silent” nociceptors underlying the flare response, and on the other hand, that “polymodal” nociceptors are potentially dominating the pain response. These findings are also in accordance with our neurophysiologic recordings in which we identified that the optimum sinusoidal stimulation profile for C-nociceptors is 3 to 8 Hz, independently of the fiber class.

4.3. Clinical implication

In recent studies, electrical sinusoidal stimuli of 4 Hz were applied using small bipolar electrodes in patients with chronic pain^1,12,16 and itch.²⁵ Using these localized stimuli, the amplitudes of 0.2 to 0.4 mA were felt extraordinary painful or itchy in approximately 50% of the patients. Our present study confirmed that the optimum sinusoidal frequency for the activation of C-nociceptors in healthy human skin is between 3 and 8 Hz. This frequency range also fits to earlier attempts using transcutaneous sine wave stimulation of 5 Hz to assess C-fiber function in peripheral neuropathy.^14,19 In their study, current detection threshold at 5 Hz was several milliampere and correlated with thermal thresholds, but the use of large contact electrodes (2 cm²) prevented a selective C-fiber stimulation. Notably, we reported recently that electrical sinusoidal pain can still be evoked when heat pain sensitivity has been abolished by topical capsaicin,³⁰ indicating that the electrical sinusoidal stimulus plus electrode configuration is effective to activate nociceptive axons even when their heat transduction site is blocked. Thus, optimized sinusoidal electrical stimulation may be used as additional tool to assess excitability of C-fiber axons in chronic pain patients.

Disclosures

The authors have no conflict of interest to declare.