Acute or chronic pain is one of the main complaints for seeking medical care. According to the International Association for the Study of Pain, pain is defined as “unpleasant sensory and emotional experience associated with, or resembling that associated with, actual or potential tissue damage” Raja et al. (2020). Pain is a protective reaction restricting physical functions with various physiological parameters, such as heart rate, respiratory rate, and arterial blood pressure, which are potential indicators of pain intensity Arbour et al. (2014); Cowen et al. (2015); Peters and Schmidt (1991). The specific changes can be observed by the reactivity and reflexivity of the autonomic nervous, cardiovascular, and respirator systems Kyle and McNeil (2014).

Adults and children suffer the pain associated with different medical conditions, undergo different painful procedures, or are referred with acute pain to the emergency department Dahlhamer et al. (2018); Keating and Smith (2011); Mura et al. (2017); Othow et al. (2022). Data suggest that 7 out of 10 patients come to the emergency department due to pain Todd et al. (2007). Meanwhile, chronic pain affects about 20.5% of adults in the United States Yong et al. (2022) and about 10%–30% in Europe Breivik et al. (2006); Reid et al. (2011). The meta-analysis showed that the prevalence of chronic pain ranges between 0% and 24% globally Mansfield et al. (2016).

Different medications, such as non-steroidal anti-inflammatory drugs, opioids, or others, are used daily against pain. Despite various pain medications and strategies, pain treatment faces many adversities Fishman (2007), such as severe side effects, the use of illicit drugs, opioid crisis St. Marie and Broglio (2020), peptic ulcers Tai and McAlindon (2021), and others. All the more, pain management is a fundamental human right Enright and Goucke (2016); Fishman (2007).

Vagal nerve stimulation (VNS)—as a pain neuromodulation technique, as reviewed in Kaniusas et al. (2019a)—has been investigated in humans and animals. VNS can affect the autonomic nervous system and is an approved treatment for pharmacoresistant depression and drug-resistant epilepsy Nemeroff et al. (2006); O’Reardon et al. (2006). Non-invasive transcutaneous modalities of VNS emerge Busch et al. (2013); Nesbitt et al. (2015), such as the electrical stimulation of the external surface of the ear innervated by the afferent auricular branch of the vagus nerve, known as auricular vagus nerve stimulation (aVNS). aVNS is performed using miniature electrodes tightly fixed inside the auricular concha. The current intensity is individually adjusted at the beginning of the aVNS session to a level without evoking pain. However, the initial personalization of the intensity of current alone does not ensure adequate vagus nerve stimulation for the relatively long treatment duration (from days to weeks). Here habituation effects, varying physiology, and deterioration of the electrode-tissue interface contribute to this uncertainty in the treatment Bouton (2017); Kaniusas (2019); Kaniusas et al. (2019b). Therefore, aVNS can be hypothesized to avoid under- or over-stimulation, reduce side effects, and save stimulation energy when based on individual physiological biofeedback.

Biofeedback can be assessed using data from internal (in-the-ear) and external (outside-the-ear) sensors, i.e., auricular and non-auricular biosignals. However, it is not known which biosignals and extracted parameters help estimate the balance between the stimulated parasympathetic system and the complementary sympathetic system, which is generally predominant in chronic ailments such as pain. In the ideal case, this balance should be provided to the aVNS stimulator to avoid the disadvantages of the non-personalized aVNS. Thus, easy-to-access biosignals are of high interest which could estimate this balance in favor of the efficiency of aVNS therapy.

The present study proposes a cold stressor as a sympathetically driven stimulus (usually accompanied by acute pain) and deep breathing as a mainly parasympathetically driven stimulus (with relaxing effects) to manipulate the sympathovagal balance from sympathetic to parasympathetic dominance while recording a set of auricular and non-auricular biosignals. The aim of this study is to assess the significance of diverse parameters of biosignals with respect to their deflection from cold stressor to deep breathing and their suitability for use as biofeedback in aVNS stimulator.

Out of 78 recorded data sets, two ECG, thirteen PPG, twenty-four Z, and six BP traces were eliminated from the analysis due to poor quality. Eight participants retreated earlier and did not finish the CPT #1 phase, i.e., two women (age ≤30 years), two women (age > 30 years), one man (age ≤30 years), and one man (age > 30 years). Two other participants did not finish CPT #2, i.e., one woman and one man, both aged > 30 years.

The present study investigates the physiological footprints of auricular and non-auricular biosignals in response to a cold stressor and deep breathing. While a cold stressor is a sympathetically driven stimulus (accompanied by acute pain), deep breathing is a mainly parasympathetically driven stimulus (with relaxing effects). Thus, it was investigated how the opposing sympathetic and parasympathetic stimuli are reflected by the auricular biosignals, namely, its parameter Z _AC, and by parameters accessible from the auricular biosignals such as RR and PPG _AC. All these three parameters can be used as biofeedback to close the loop in aVNS, i.e., in a targeted stimulation of the parasympathetic system. The closed-loop set-up personalizes aVNS with an expected tendency to avoid over and under-stimulation of the vagus nerve/parasympathetic system. Thus, the closed-loop aVNS may minimize both the energy consumption of the aVNS stimulator and potential side effects (no over-stimulation) while optimizing and personalizing the aVNS therapy (no recruitment of pain fibers), e.g., in chronic pain. Here, the non-auricular biosignals with their parameters BP _S and BP _D serve as a necessary reference to monitor stimuli-related vital functions of the body and as an instructive substrate for their comparison with stimuli-related changes in auricular biosignals.

The auricular biosignal Z, namely its parameter Z _AC (Figure 1), decreases significantly during the sympathetic stimulus (CPT #2) as compared with the parasympathetic one (DB) (Figures 3–5), as well as decreases tendentially with increasing pain perception (Figure 6). This behavior indicates the potential suitability of Z _AC in assessing changes in the balance of the parasympathetic and sympathetic stimuli, or, more generally, in the balance of the parasympathetic and sympathetic systems of the human body (sympathovagal balance). On the other hand, this balance, especially its normalization from a derailed state, is usually a therapeutic target in aVNS when applied to different chronic ailments Kaniusas et al. (2019a). Thus Z _AC can be hypothesized to be reasonable auricular biofeedback for the closed-loop aVNS without using any sensors external to the ear, which may obstruct the patient.

The parameters RR and PPG _AC also reflect sympathovagal balance. RR and PPG _AC decrease significantly during the sympathetic CPT #2 compared to the parasympathetic DB (Figures 3–5). While RR tends to decrease with increasing pain, the level of PPG _AC does not (Figure 6). Therefore, RR and PPG _AC, the former to a larger extent, can also be hypothesized to be reasonable auricular biofeedback for the closed-loop aVNS targeting a derailed sympathovagal balance. Please note that RR could be estimated from the period of the cardiac oscillation of the auricular Z (Figure 3), whereas PPG _AC from the cardiac deflection of PPG from the earlobe Allen (2007). However, limitations in the precision of the estimated RR may apply in the former case due to a rather smooth waveform of Z in contrast to the spiky R peak of ECG. Likewise, limitations in PPG _AC may apply in the latter case due to a rather central connection of the ear perfusion in contrast to the peripheral perfusion of the finger (Figure 1).

The non-auricular parameters BP _S and BP _D reflect the sympathovagal balance as well. Both increase significantly during the sympathetic CPT #2 as compared with the parasympathetic DB (Figures 3–5), while this increase in CPT #2 tends to be larger for stronger pain (Figure 6). The level of BP _D appears to depend even stronger on the stimuli-induced sympathovagal balance with the observed changes of 11.7% (from CPT #2 to DB) in contrast to the associated changes in BP _S of 8.0% (Figure 4). Likewise, the sympathetically-governed vasoconstriction (governing BP _D) may be more dominant than stroke volume changes (governing BP _S) Kaniusas (2012). This leads to a hypothesis that BP _S and BP _D could be used as non-auricular biofeedback for the closed-loop aVNS when external sensors are used outside the ear.

In terms of gender and age, the largest changes from CPT #2 to DB were shown in RR for adult women (≥30 years), PPG _AC for young women ( < 30 years), Z _AC for young women ( < 30 years), BP _S for adult women (≥30 years), BP _D for young women ( < 30 years). In contrast, the minor changes from CPT #2 to DB were shown in RR for young men ( < 30 years), PPG _AC for adult men (≥30 years), Z _AC for young men ( < 30 years), BP _S for adult men (≥30 years), BP _D for adult men (≥30 years). Overall, men seem to show fewer changes from CPT #2 to DB than women. This conclusion is in line with previous studies, which conclude that women are more sensitive to pain Fillingim et al. (2009); Mogil (2012); Popescu et al. (2010), but it depends on the method of pain induction and assessment. In most cases, the study also supports the still controversial claims that older individuals are more tolerant of pain and show fewer physiological effects than younger individuals Edwards et al. (2003); Riley et al. (2010); Rittger et al. (2011). However, these statements are very limited in their validity due to the small sample in this study.

A limitation of the present study is the relatively small database of recordings representing the elder part of the population which has tendentially a larger prevalence of suffering pain. Collecting and analyzing a more representative database is planned as a future research direction in the research of the aVNS stimulator. Since the warm water phase immediately preceding CPT #1 strongly affected the results in CPT #1, we focused our investigations on the comparison of CPR #2 and DB, both preceded by rest phases. Another limitation is that the order of the different phases of the protocol were not randomized, especially the order of CPT and DB. Therefore, the results may have been influenced by other factors such as expectation, adaptation, prolonged exposure.

Lastly, it should be noted that the recorded pain level, in contrast to nociception with its physiological encoding and processing of nociceptive stimuli, is a subjective feeling connected with the emotional experience to impeding or actual harm Loeser and Treede (2008) but also altering autonomic nervous system Woo et al. (2017); Adamczyk et al. (2020); Abdallah and Geha (2017). Thus, the investigated objective characteristics of the autonomic system may be useful for a continuous and objective personalization of aVNS in chronic ailments such as pain.

The three parameters RR, PPG _AC, and Z _AC accessible from auricular biosignals reflect the artificially-induced stimuli with sympathetic or parasympathetic dominance and thus the sympathovagal balance derailed in pain and other chronic conditions. Therefore, auricular biosignals can be used as biofeedback to close the loop in auricular vagus nerve stimulation to personalize the strength and timing of the stimulation in favor of therapy, patient compliance, and resourceful energy use.