Transcutaneous Auricular Vagus Nerve Stimulation Attenuates Early Increases in Heart Rate Associated With the Cold Pressor Test

Abstract

Introduction:

Transcutaneous auricular vagus nerve stimulation (taVNS) may be useful in treating disorders characterized by chronic parasympathetic disinhibition. Acute taVNS decreases resting heart rate in healthy individuals, but little is known regarding the effects of taVNS on the cardiac response to an acute stressor. To investigate effects on the acute stress response, we investigated how taVNS affected heart rate changes during a cold pressor test (CPT), a validated stress induction technique that reliably elicits a sympathetic stress response with marked increases in heart rate, anxiety, stress, and pain.

Materials and Methods:

We recruited 24 healthy adults (ten women, mean age = 29 years) to participate in this randomized, crossover, exploratory trial. Each subject completed two taVNS treatments (one active, one sham) paired with CPTs in the same session. Order of active versus sham stimulation was randomized. Heart rate, along with ratings of anxiety, stress, and pain, was collected before, during, and after each round of taVNS/sham + CPT.

Results:

In both stimulation conditions, heart rate was elevated from baseline in response to the CPT. Analyses also revealed a difference between active and sham taVNS during the first 40 seconds of the CPT (Δ heart rate [HR] = 12.75 ± 7.85 in the active condition; Δ HR = 16.09 ± 11.43 in the sham condition, p = 0.044). There were no significant differences in subjective ratings between active and sham taVNS.

Conclusions:

In this randomized, sham-controlled study, taVNS attenuated initial increases in HR in response to the CPT. Future studies are needed to investigate the effects of various taVNS doses and parameters on the CPT, in addition to other forms of stress induction.

Clinical Trial Registration:

The Clinicaltrials.gov registration number for the study is NCT00113453.

INTRODUCTION

Chronic stress system activation is present in a range of mental health disorders.¹ One of the hallmarks of chronic stress system activation is lowered vagal tone, which reflects the functional balance between sympathetic and parasympathetic activity in the cardiovascular system.² Stress can lead to poor inhibition of sympathetic activity and inflexible parasympathetic activity, resulting in measurable changes in vagal tone indices (commonly reduced heart rate variability).³ Although chronically lowered vagal tone, an indicator of vagus nerve (cranial nerve X and fundamental component of the parasympathetic) activity, is an important contributor to various mental health disorders, interventions targeting the vagus nerve may improve physiological and psychologic response to stress.

Pharmacologic blockade of the vagus nerve causes increased heart rate, suggesting that cardiac activity is regulated by inhibitory vagal control.^4,5 Transcutaneous auricular vagus nerve stimulation (taVNS), meanwhile, is a noninvasive method for targeting the vagus nerve using electrical stimulation rather than medication, and it also may be useful in treating psychiatric disorders associated with stress and low vagal tone. Preclinical and clinical trials show that implanted vagus nerve stimulation (VNS) may be an effective anxiolytic treatment and may improve outcomes in anxiety disorders.^6,7 In addition, acute taVNS decreases physiological markers of stress, such as heart rate, in healthy individuals at rest.⁸ Because chronic stress system activation is a component of many psychiatric disorders, investigating the effects of taVNS during a stress induction technique may yield insight into ways to develop the modality as a treatment for clinical populations.

One commonly used technique for studying stress responding is the cold pressor test (CPT), which involves immersing an extremity (ie, hand, foot) in an ice bath for 1 to 2 minutes. The CPT was first introduced in the 1930s by Hines and Brown and is a valid and reliable test that allows researchers to elicit a sympathetic stress response in a safe and monitored environment.⁹ CPT exposure leads to profound changes in cardiovascular parameters, most notably an increase in blood pressure through peripheral vasoconstriction and to a lesser extent, cardiac output resulting from an increase in both vascular alpha-adrenergic and cardiac beta-adrenergic drive.^10,11 The CPT also produces significant changes in subjective experiences of the test, including increases in anxiety, distress, and pain.¹²

To our knowledge, no study has yet investigated the effect of taVNS on a stereotypical, sympathetic stress response. Therefore, we conducted an exploratory trial exploring the effect of taVNS on heart rate (HR) during the CPT. We hypothesized that active compared with sham taVNS would attenuate typical increases in HR seen during the CPT. In addition, we hypothesized that subjective ratings of anxiety, distress, and pain would be less pronounced in subjects receiving active than in those receiving sham taVNS.

MATERIALS AND METHODS

Overview

Participants attended a single study visit comprising two sessions of CPT with concurrent active or sham taVNS. Study visits were conducted at the Medical University of South Carolina (MUSC) Brain Stimulation Lab. This study was approved by the MUSC institutional review board and is registered on ClinicalTrials.org (NCT00113453). All participants signed written informed consent before participating in the study.

Participants and Inclusion Criteria

In total, 24 healthy individuals (ten women) were enrolled after meeting the following inclusion criteria: 1) aged 18 to 65 years, 2) English-speaking, 3) non–treatment seeking community members, 4) no diagnosis of COVID-19 in the past 14 days, 5) no facial or ear pain, 6) no recent ear trauma, 7) not taking any medications currently, 8) no significant medical history, and 9) signed inform consent. Subjects with any significant cardiac, pulmonary, psychiatric, or neurologic history were excluded from the trial. In addition, pregnant subjects were excluded from the trial.

Transcutaneous Auricular Vagus Nerve Stimulation Setup

We delivered taVNS using a constant current electrical nerve stimulator (Digitimer DS7AH, Digitimer LTD, Hertfordshire, United Kingdom) connected to 1 × 3/8-inch hydrocolloid electrodes (Micro NeoLead^® ECG Electrodes, Neotech Products LLC, Valencia, CA). Identically to previous studies,^13,14 electrodes targeted the left cymba conchae (anode) and the tragus (cathode). Stimulation was triggered manually by a blinded investigator (more detailed description of blinding found below). Electrical stimulation was delivered using a 500 μs pulse width and 25 Hz stimulation frequency. Stimulation was continuous for 2 minutes (no off cycle) at 200% of the participant’s perceptual threshold. In a previous study by our group, these parameters had the largest effect on HR in healthy individuals at rest.⁸

Assessment of a subject’s taVNS perceptual threshold (PT) followed procedures as described in other studies by this group.¹⁴ The procedure entails delivering single pulses of stimulation to the target area and obtaining verbal confirmation of stimulation perception while modulating the current intensity through a parametric estimation by sequential testing method.¹⁵ This allows determination of the minimum current intensity perceived by the participant. A short train of stimulation at two times the PT was delivered to assess for tolerability and pain. If the subject found the train painful or could not tolerate the train, the dose of current was decreased to a tolerable level.

The study was intended to be double blind, with the participant and primary investigator not aware to which arm the participant was assigned. However, owing to the perceptual difference between the active and sham conditions (participant could feel stimulation in the active condition and did not feel stimulation in the sham condition), the study functioned more as a “pseudo-blinding” for the participants, and the investigator was fully blinded. The blinding procedure comprised a nonblinded research coordinator adjusting the output-enable switch on the Digitimer before beginning each round of concurrent taVNS and CPT. After the PT was determined, the nonblinded coordinator would enter the study space and adjust the output-enable switch to either ON (allowing the unit to generate output pulses for the active condition) or OFF (sham). A cover was then placed over the Digitimer to shield the primary investigator and participant from the device settings. The nonblinded coordinator left the research space after adjusting the taVNS switch, and active stimulation + CPT sessions were conducted by the primary investigator. The nonblinded coordinator was always the same person, and they had no other contact with subjects or the data or analysis.

In the active taVNS condition, current flowed from the Digitimer to the electrodes on the subject’s ear. The stimulator’s output-enable switch was turned OFF during the sham condition;thus, no electrical current was delivered to the electrodes targeting the subject’s ear.

Cold Pressor Test

Since its first introduction in the 1930s,⁸ many studies have confirmed the validity and reliability of the CPT and its ability to induce profound changes in cardiovascular parameters.^10,12,16,17 This study used a variation of the CPT that involved participants immersing both feet in ice water. Both feet were immersed in the tub of ice water for 2 minutes duration. The temperature of the water was initially measured between 1 and 3 °C (measurement at the beginning of CPT, mere seconds before participants placed their feet in the water). A script was used when communicating with study participants once the visit began. Participants were asked to remain as still and quiet as possible throughout the study procedures but were encouraged to voice concerns about any discomfort.

Outcomes

The primary outcome in this study was HR. HR was measured using a BIOPAC MP150 + ECG100C system and AcqKnowledge© Data Acquisition and Analysis Software. HR was measured online using disposable Ag/AgCl electrodes placed on the right collarbone and at the base of each rib cage. Data were sampled online at 1000 Hz and filtered online using a bandpass filter from 0.05 to 35 Hz. After data collection, HR was determined using the AcqKnowledge QRS detection method. In analysis windows, HR was downsampled into 10-second bins.

In addition, secondary measures of subjective anxiety, pain, and distress were collected periodically during each study visit. Participants were asked to rate their subjective experience of anxiety, distress, and pain on a scale of 0 to 10 (where 0 represents the feeling was not present at all and 10 represents the most extreme experience possible) at three different time points during the visit: 1) at the beginning of each round (Rest), 2) immediately after each CPT (asked about their subjective experience over the past 2 minutes; CPT), and 3) at the end of each round (Decline).

Timeline

After subjects were screened and signed the informed consent, they were asked to remove their shoes and socks and sit upright in the designated chair. Physiological recording electrodes were placed at three different positions on the subject’s thorax and abdomen (Fig. 1).

Figure 1.

Electrode placement for taVNS and recording physiological parameters. EKG electrodes were attached to the subject’s chest in the locations marked on the diagram. taVNS electrodes were attached to the subject’s left cymba conchae (anode) and tragus (cathode).

Once physiological recordings were assessed for quality, taVNS electrodes were placed on the left cymba conchae and tragus as previously described (Fig. 1). The PT was then determined. The nonblinded research coordinator entered the room and arranged the stimulator’s settings according to the randomized blinding, and a cardboard box was placed over the stimulator to shield its settings from view.

After arrangement of the recording and taVNS electrodes, a script was used to convey the CPT procedures. Each round of concurrent taVNS and CPT comprised 10 minutes of resting physiological measures (Rest/Anticipation), 2 minutes of concurrent taVNS and CPT (CPT), and another 10 minutes of recovery measures (Decline). Subjects were notified that the task would begin 2 minutes before the start of CPT (Anticipation). At the start of the CPT, participants were asked to place their bilateral feet in the tub of ice water, at which time taVNS was turned “ON” (with current flowing from device to electrodes in the active condition but no current delivered in the sham condition). After 2 minutes, they were asked to remove their feet from the tub. As described above, participants were asked to rate subjective measures at Rest, CPT offset, and Decline for each CPT round. Half the participants were randomized to receive active taVNS first, then sham, and the other half received sham, then active, taVNS.

Statistical Analysis

In total, the data of 21 of 24 enrolled participants were used in the final analyses (two were not used owing to poor data quality, and one was not used owing to inability to complete both phases of the trial). A two (Condition: Active vs Sham)-by-three (Time: Rest, CPT, Decline) repeated measures ANOVA with Geisser-Greenhouse sphericity correction was conducted using SPSS version 28 (IBM Corporation, Armonk, NY) to determine main and interaction effects. Significant effects were subsequently examined using Bonferroni adjusted pairwise comparisons.

RESULTS

Participants

A total of 24 (ten female) healthy (mean age 28.8 ± 5.3 years SD) participants were enrolled in this study. Of the 24 participants enrolled, 23 completed the study (one participant dropped out owing to inability to maintain bilateral feet in the ice bath during the first round of CPT). The data from 21 participants were used in the final analysis of HR because two sets of data were of insufficient quality to include in the analysis (electro-cardiogram [EKG] electrodes displaced from their bodies owing to arm/torso movement in response to the CPT). Two of the three dropouts were randomized to receive active taVNS first. There were no significant differences between the groups randomized to active taVNS or sham stimulation first in terms of age, sex, or race.

Adverse Events

A single participant completed one of two rounds of the CPT owing to inability to maintain bilateral feet in the ice bath during the first round of the CPT. This subject’s data were not included in the final analysis. Otherwise, no reportable adverse events were produced by this protocol. taVNS and CPT were well tolerated by the remaining subjects, without any significant or persistent side effects.

TaVNS was well tolerated by all participants in this protocol. There were no significant differences in PT between the groups receiving active or sham stimulation first (Table 1).

Table 1.

PTs and Stimulation Intensities.

XXX	Active taVNS	Sham stimulation
PT (mA)	0.75 ± 0.22	0.76 ± 0.30
Stimulation Intensity (mA)	1.46 ± 0.43	N/A

PTs were determined before each round of concurrent stimulation/CPT, even though current was not administered in the sham condition. There were no significant differences in the PTs between the active taVNS and sham stimulation conditions.

Effect of Stimulation on HR

The overall pattern of HR over time followed a similar trajectory to the typical CPT stress response curve (Fig. 2). In both active and sham taVNS, a sharp increase in HR at the beginning of the CPT is followed by a period of sustained, elevated HR and then a decrease in HR toward the end of the CPT. Of note, two small elevations in HR are seen during the “anticipation” period (2 minutes before start of CPT), which aligns with the timing of participants being told CPT would start in 2 minutes and the ice bath being moved closer to participants just before the start of CPT. These elevations were not significantly increased from baseline, and there was not a statistically significant difference in HR between the active and sham conditions during the anticipation period. The increases in HR seen on either side of the CPT (at the beginning of CPT and at the end of CPT) also were likely influenced by movement, when the participants placed their feet into the ice bath and took them out of the bath.

Figure 2.

Δ HR during concurrent taVNS and CPT. Change in HR from baseline during the cold pressor test. The blue rectangle represents the 2-minute duration of concurrent stimulation and CPT. Heart rate is plotted in 10-second averages over time during four blocks: Anticipation (2 minutes before CPT), S1 (first 40-second block of CPT), S2 (second 40-second block of CPT), S3 (last 40 seconds of CPT), and Decline (2 minutes after CPT). Baseline HR was measured as an average of the subject’s HR during the 2 minutes before the anticipation period. Error bars are standard error of the mean. bpm, beats per minute. *S1 had between-group significance (p < 0.05).

The active and sham groups have differently shaped HR curves, especially during the early portion of the CPT. A Shapiro-Wilk test of normality did not show any significant deviation at p < 0.05; thus, the HR data were normatively distributed. Baseline HR was calculated as the average HR over the 2 minutes before the Anticipation period in each stimulation condition (from 4 minutes before CPT until 2 minutes before CPT). Mean changes from Baseline to the Anticipation and CPT phases (averaged over 40-second blocks for the analysis, consistent with statistical analyses in previous CPT studies) can be seen in Table 2. Δ HR from Baseline (change from Baseline to Anticipation, CPT, and Recovery) was analyzed in a repeated measures ANOVA and revealed a significant difference between the active taVNS and sham stimulation groups during the first one-third of the CPT (first 40 seconds of the 2-minute CPT, p = 0.044). There was no significant difference in Δ HR between active and sham during the middle or late portions of the CPT.

Table 2.

Mean Changes in HR During Concurrent taVNS and CPT.

Stimulation condition	Anticipation Δ HR	S1 Δ HR (1st 40-s block of CPT)	S2 Δ HR (2nd 40-s block of CPT)	S3 Δ HR (3rd 40-s block of CPT)
Active taVNS	6.31 ± 6.43	12.75 ± 7.85*	10.28 ± 12.11	9.03 ± 11.58
Sham stimulation	8.32 ± 9.57	16.09 ± 11.43*	10.93 ± 12.3	8.87 ± 11.23

Mean changes in HR from baseline during the anticipation period (2 minutes before taVNS/CPT), first 40 seconds of taVNS/CPT (S1), second 40 seconds (S2), and final 40 seconds (S3). There was a significant between-group difference in the change in HR during the first 40 seconds of CPT (S1, p < 0.05).

Effect of Stimulation on Subjective Ratings

Subjective ratings of anxiety, distress, and pain before, during, and after concurrent taVNS and CPT are seen in Figure 3. There was a main effect of time on anxiety, distress, and pain such that there were significant increases in scores during the CPT compared with baseline. There were no main effects of condition or interaction effects on subjective ratings throughout the protocol.

Figure 3.

Effect on anxiety, pain, and distress. Change in subjective ratings of anxiety, pain, and distress from baseline to concurrent stimulation/CPT. There were no significant differences between the active taVNS and sham stimulation groups regarding subjective ratings.

Integrity of the Blind

Immediately after the second CPT, participants were asked which of the stimulation conditions they received first. Responses were 95% correct. After participating in both rounds of concurrent taVNS/CPT, most participants reported that they could feel ear stimulation in the active condition and did not perceive any stimulation during sham. However, between the first and second round of stimulation, most participants were not confident which stimulation condition they had received during their first round. Unfortunately, we did not ask for guesses after each condition but only after both were completed.

Discussion

This was a randomized, sham-controlled, crossover trial investigating the effects of taVNS on the canonical increases in HR seen during the cold pressor test. We hypothesized that active taVNS would attenuate the typical increases in HR during the CPT compared with sham stimulation, and that active taVNS also would attenuate the typical increases in anxiety, distress, and pain seen during the CPT compared with sham stimulation. In support of our first hypothesis, active taVNS significantly attenuated HR during the early portion (first 40 seconds) of the CPT. In contrast, regarding our second hypothesis, active taVNS did not have any significant effect on subjective ratings of anxiety, distress, or pain. In addition, and importantly, no adverse effects or serious adverse effects were seen throughout this protocol. taVNS and CPT were well tolerated by the study participants. One participant could not maintain both feet in the ice bath for the 2-minute duration of the CPT and thus was not included in the final analysis. All other participants were able to participate in the full protocol without any issues.

In the repeated measures ANOVA analysis, there was only a significant difference between groups (ie, Active vs Sham taVNS) in change in HR from over the first 40 seconds of the CPT (of 2 minutes). There were no significant differences in Δ-HR (from Baseline) between the active and sham conditions over the remaining 2 minutes of CPT. These results suggest that taVNS attenuates the early portion of typical elevations in HR associated with the CPT. As previously indicated, the CPT is a stress-provocation protocol that is sympathetically mediated. The CPT is associated with increases in circulating sympathetic hormones such as epinephrine and norepinephrine, which ultimately cause increases in sympathetic biomarkers, including HR, mean arterial pressure, and cardiac output.¹⁸ In contrast, the vagus nerve is part of the parasympathetic nervous system with an estimated 20% efferent fibers that have end-organ effects, including modulation of cardiac function.¹⁹ In previous studies, our group showed that taVNS has parameter-specific effects on HR and found that certain parameters produce larger decreases in HRs in healthy individuals at rest. In the current study, we showed that taVNS evokes the same effect on HR, but during a sympathetic stress induction task.

Low vagal tone and exaggerated stress responding are common in many stress-related disorders.²⁰ The CPT has a long history and is a validated stress induction task that allows researchers to reliably reproduce a stereotypical stress response. In this study, taVNS attenuated typical increases in HR seen during the CPT, suggesting that taVNS may be useful as a tool for modulating response to an acute stressor. These results indicate that taVNS may be useful in modulating autonomic arousal and hyperactivity. The ability of taVNS to modulate the autonomic nervous system may have implications for clinical populations, given many psychiatric disorders are associated with acute and chronic stress. Furthermore, clinical trials investigating the efficacy of taVNS as a treatment for depression and other psychiatric and neurological disorders have shown that there is a cumulative effect of taVNS over time.²¹ Thus, multiple taVNS treatments (over days to weeks) may have a larger effect on stress than does the brief 2-minute treatment delivered in this trial.

The current study had several strengths. It successfully used the CPT, inducing the typical stress response seen in other trials using the CPT. This allowed us to investigate the effects of taVNS on the stress response induced by the CPT. Although taVNS only mildly suppressed HR during the early portion of the stress response, these results are consistent with other studies revealing that auricular VNS modulates cardiac activity²² and offer cause for further exploration in studies that investigate effects of various taVNS parameters and doses on stress and disorders associated with low vagal tone. There are several ways to manipulate the parameters of electrical stimulation in taVNS treatments, including current intensity, frequency of pulses, pulse width, duty cycle (on/off time), and duration/number of treatments. In this trial, we delivered 2 minutes of continuous stimulation (at two times the PT, 25 Hz frequency, and 500 microsecond pulse width). It is possible that the effects seen early in the stress response were not durable owing to habituation; a duty cycle (short trains separated by periods without stimulation) is one potential direction that may enhance the cardiac modulatory effects of taVNS over longer periods. There is still much to be understood about the modulatory effects of various parameters.

This trial was limited by several factors. First, the study recruited a small sample. A larger sample may reveal information about ways certain subject traits influence the effect of taVNS on the stress response. Owing to lack of previous investigation in this area, the taVNS parameters and dosing chosen for this study were based on previous trials in healthy individuals at rest. However, it is possible that a different set of parameters or a different dose (current intensity, starting the taVNS before starting the CPT, etc) would be more efficacious in suppressing the stress response induced by the CPT. It has been proposed that stimulation of the right vagus nerve may have stronger effects on HR than stimulation of the left vagus nerve owing to its skewed innervation toward the sinoatrial node. In this study, taVNS was applied to the left ear’s cymba conchae and tragus, the electrode montage that maximally activates the auricular branch of the vagus nerve.²³ Future investigations should consider comparing these results with stimulation of the right ear. It is important to note that the mechanism of parasympathetic nervous system activation is likely indirect, through afferent central targets (ear to brain). Mechanisms outside the vagus nerve (trigeminal nerve, ortho-sympathetic modulation) also may be playing a role in the results of this study.

We did not assess blinding integrity after each CPT and did so only after both were completed. We assume, but did not document, that for at least the initial CPT, subjects were fully blinded. Raters and investigators were blinded throughout each session until the end. We considered several factors when choosing the sham for this study. We also considered an active, but less intense, sham; however, we did not want this to blunt any effect we would see in the results. We chose a blinded switch as the best alternative (setting of stimulation switch ON or OFF was blinded from investigator and participant); however, there are limitations to this sham method. Future studies should focus on developing sham techniques that are consistently indifferentiable from active taVNS, given in this study, there was a perceptual difference between the two conditions.

The CPT was chosen as the method to induce a stress response for this study on the basis of taVNS being a potential treatment option for disorders associated with acute and chronic stress (accompanied by autonomic arousal and hyperactivity). In this study, we successfully induced a sympathetic stress response using the CPT, and taVNS had a significant effect on the early portion of that response. This may be due to the factors previously described (taVNS dosing/parameters or electrode location may play a role). The CPT produced a highly painful and shocking response in this study, as evidenced by the significant and large increases in anxiety, distress, and pain during the CPT. This response is comparable to a panic attack or to the anxiety seen at the height of a fear hierarchy in social anxiety disorder/specific phobias. Alternative methods of stress induction may offer clues to the potential of taVNS to treat anxiety and related disorders. It is possible that the significant amounts of pain and anxiety provoked by this test obscured our ability to see more of a physiological or subjective response to taVNS during the CPT. In addition, heightened pain induced by the CPT may have masked responses on other subjective measures (anxiety, distress) during concurrent taVNS and CPT. Disorders associated with low vagal tone are often associated with high levels of anxiety and stress; however, high levels of pain would not typically be seen in these populations. Other stress induction tests may offer further clues when investigating the potential of taVNS as a treatment for these disorders moving forward.

Conclusions

In this randomized, sham-controlled, crossover study, taVNS significantly attenuated increases in HR during the early portion of the CPT. The study was limited by a small sample size. Future studies should investigate the effects of various taVNS doses and parameters on the CPT and consider using other forms of stress induction that are less painful.