What are the mechanisms of action of auriculotherapy, auricular acupuncture or auricular neuromodulation?

How can stimulating a specific area of the ear pavilion have a therapeutic action, such as relieving pain or stress? Here are some answers. Warning! These elements do not refer to "traditional" theories such as those found in auricular acupuncture or auriculotherapy teachings. The focus will be on neurophysiology.

The Vagus Nerve Pathway

Why the vagus nerve?

Everything can arise from a misunderstanding! At the end of the 19th century, James Corning, an American neurologist, designed a device to stimulate the vagus nerve in its retro-carotid portion in order to treat epilepsy [1]. This device resembled a two-pronged fork, each prong being applied opposite the carotids, and delivering a continuous current. At that time, epilepsy was considered to be the result of excessive cerebral blood flow, which could be reduced by slowing the heart rate.... by stimulating the vagus nerve. If the effectiveness was not quite there in humans, invasive stimulation of the vagus nerve was effective in animal models developed in the 1950s. This research led in the 1990s to the first so-called "invasive" vagal stimulation by implantable devices. This invasive stimulation has shown its effectiveness in the treatment of resistant epilepsy, thus validating Corning's hypothesis, but with a completely different rationale [2]! In addition to resistant epilepsy, severe depression can also be treated by invasive vagal stimulation [3].

From invasive vagal stimulation to non-invasive transauricular vagal stimulation

The problem with invasive vagal stimulation is its invasive nature! The device must be implanted, and the immediate post-operative complications are 9.1% in adults (hematoma, infection, vocal cord paralysis) [4]. At six months, more than one in two patients has dysphonia [5], not to mention the cost of the device and the limited number of teams that can implant it.

The next step was to develop vagal stimulation in a non-invasive way. The first stimulators were devices that the patient had to apply to the carotid pathway of the vagus nerve, such as the gammaCore®. This device is used in the treatment of migraine attacks [6] and has received approval from the Food and Drug Administration in this indication. Its main drawback is its ergonomics, as the patient must hold the device permanently.

Now, the concha is innervated by the vagus nerve, and especially its upper part (upper hemiconcha) [7]. Why not design a non-invasive device that would stimulate the vagus nerve via the upper hemiconcha? This was the next step. This type of stimulation has shown its effectiveness in the two main indications of invasive vagal stimulation, drug-resistant epilepsy [8], and major depression [9], but also in many other indications such as insomnia, functional bowel disorders, post-stroke rehabilitation, etc. (the list is long).

But the vagus nerve is not alone

If we look at the anatomy of the auricle, it is innervated by 3 main contingents: in addition to the vagus nerve via its auricular branch, there is the trigeminal nerve via one of its branches, the auriculo-temporal nerve, which innervates the upper and posterior part of the auricle, and the superficial cervical plexus via the occipital nerves, which innervates mainly the lobe of the ear. Other nerves contribute more secondarily to the innervation of the auricle, such as fibers from the facial nerve (7th cranial pair) for the edge of the auditory canal or the glosso-pharyngeal nerve (9th cranial pair) for part of the tragus [10]. Could it be that the stimulation of these other nerves has an action as interesting as that induced by the vagus nerve?

How does cutaneous stimulation have a neuromodulatory action?

Classical anatomical data do not explain the modulatory action of cutaneous stimulation, nociceptive or not, of the auricle. Taking only the vagus nerve as an example, its afferent fibers all belong to the general sensitivity called "somatic" ("GSA" for General Somatic Afferent) and all terminate in the spinal nucleus of the trigeminal nerve [11]. Therefore, in principle, stimulation of the auricle should only lead to sensory relays via general somatic sensitivity.

Fortunately, studies in retrograde tracing and functional MRI allow us to go further.

Retrograde tracing is the first breach in classical anatomy

In animals, by retrograde tracing, the afferent information conveyed by the three main contingents is projected onto structures other than the spinal nucleus of the trigeminal nerve [10]. Vagal afferents project as expected onto the trigeminal nucleus, but also the nucleus of the solitary tract (NTS), which is the relay of all non-painful interoceptive afferents ("GVA" for general visceral afferent) [12]... mediated by the vagus nerve.

The afferents of the small and large occipital nerves (the branches of the superficial cervical plexus) project as expected into the cell bodies in C2-C3 and the trigeminal nucleus, but also onto structures not provided for by the anatomical model, which are a) the superior cervical ganglion (with sympathetic valence), b) the cervical spinal cord on layers I to V (recall that layers II, III and IV are places of encounters with cortical afferents involved in the inhibitory control of pain for example), c) the cuneate nucleus (where the first synaptic relay of general sensory afferents is located. This nucleus was formerly called the Goll nucleus, whose counterpart for the lower limbs is the Burdach nucleus, currently named the gracile nucleus), and d) the nucleus of the solitary tract.

The afferents of the auriculotemporal nerve (the branch of the trigeminal - V3 -) arrive as expected on the trigeminal nucleus, but also the cuneate nucleus, like the branches of the superficial cervical plexus [10]. Finally, these three contingents all project onto the III and IV layers of the upper cervical spinal cord, which are also relays of the sympathetic pathways. In the case of vagal afferents, these projections do not exist if the tracing is done directly from the trunk of the vagus nerve and not its auricular branch [10]. Interesting, isn't it?

Functional MRI adds further insights into the mechanisms at play.

And in humans? The repetitive electrical stimulation of the concha confirms the results obtained in animals, namely that it does indeed activate the relays of the vagus nerve related to general visceral sensitivity ("GVA", Cf supra) [11]. In addition, there is an activation of other higher structures, such as the locus coeruleus, the amygdala, the nucleus accumbens or the parabrachial area, and this in a bilateral way. The activation of these structures is different depending on the area stimulated on the auricle, even if the available data concern only four different areas, as shown by Yakunina et al. [12].

The clinic is where the action is !

Let's go back to the clinic: three reproducible principal observations make it easier to understand the pathophysiology: induced allodynia, somatotopy and the analgesic efficacy after stimulation.

Allodynia

When nociceptive information emanates from a dermatome, hypersensitivity to a usually non-painful stimulus (called allodynia) will appear on a circumscribed area of the auricle. The most striking example is to experimentally pinch the pulp of the 2nd phalanx of the thumb with a drawing pin, then, 1 minute later, to look for a hypersensitive area at the level of the auricle (we give you the trick; it is at the top of the scaphoid fossa, but it will require a tip with a foam edge of 1 to 3 mm in diameter to properly search for this area. There are dedicated tools available commercially). This experiment is very easy to do and remains reproducible.

Somatotopy

The distribution of these hypersensitive areas follows a particular topographic organization, like a map. This somatotopy was published for the first time by the founder of auriculotherapy, Paul Nogier, in the 1950s, in the form of a cartography [13]. This somatotopy is easy to describe for exteroceptive nociceptive afferents, but a little less so for interoceptive afferents, whether nociceptive or not. This is why many "auricular cartographies" are currently circulating.

Somatotopy is not unique to the ear! It already exists in many brain structures. The oldest one described is the somatotopic organization of the motor cortex on the ascending frontal gyrus, called the Penfield homunculus, named after its discoverer. Many other topographic organizations exist in other brain structures, such as the thalamus, the insula... and the ambiguous nucleus and the dorsal motor nucleus of the vagus, which are the two motor nuclei of the vagus nerve.

Analgesia after stimulation of the allodynic zone

Finally, vigorous stimulation of the allodynic zone of the auricle will significantly reduce somatic pain. Magic? Not at all. There are at least two models that can explain this phenomenon: the first is the gate control theory, the second is the interoceptive paradigm known in manual medicine.

• The gate control theory, enunciated by Melzach and Wall in the 1960s, is a typical example of neuromodulation at the spinal level [14]: it consists of competing non-nociceptive influxes, carried by large diameter fibers, with nociceptive influxes carried by small diameter fibers: at the spinal level, the two pieces of information arrive on relays that only allow one piece of information to pass, the non-nociceptive one. This theory has been largely amended since then [15], but is the basis of treatment by transcutaneous electrical nerve stimulation, better known by the English acronym TENS (transcutaneous Electrical Nerve Stimulation).
• The interoceptive paradigm requires a little development. Pathological interoceptive information generates a state of sensitization on certain anatomical areas. In manual medicine, this accounts for both so-called "referred" pain, such as Maigne's syndrome, but also metameric hypersensitivity on different tissues: muscle (muscle cord, trigger zone), the integument (the integumentary infiltration felt during the palpate-roll maneuver, called "cellulalgia"), but also the tendon and the periosteum. This is the "cellulo-periosto-teno-myalgic syndrome". However, an action (often mechanical) on the sensitized anatomical areas has an action on the autonomic nervous system, as has been demonstrated in manual medicine but also after stimulation of the auricle [16-18]. This is how, by strongly pinching the integument, stimulating a muscle trigger zone or performing a high-velocity but low-amplitude joint mobilization (a "manipulation"), an analgesic effect is expected... exactly as by stimulating the auricle on an allodynic zone. To explain this mechanism, supraspinal relays are needed. That's good, there are some!
  • At the bulbar level, let us mention the locus coeruleus, which acts as a hub. The latter could explain the parasympathetic activity induced by an osteopathic treatment [17] and is itself controlled by supra- and sub-tentorial efferents involved in the descending control of pain (periaqueductal gray matter and ventral tegmental area).
  • At the cortical level, two structures are important: the insular cortex, especially in its posterior part and the anterior cingulate cortex, which projects its efferents onto the periaqueductal gray matter, which itself is a major effector in the descending inhibitory control of pain.
  • It will be noted that all these structures can be activated in fMRI after stimulation of the auricle [11,12].

In summary

Did you make it this far? Well done! The key points are as follows:

• Stimulation of the auricle for therapeutic purposes, whether called auricular neuromodulation, auriculotherapy or auricular acupuncture, generates an activation or inhibition of brain structures, therefore of neural networks, at the subcortical and cortical level (fMRI data).
• This activation or inhibition differs depending on the stimulation site on the auricle (Yakunina et al.)
• Stimulation of the auricle in areas not innervated by vagal afferents also generates activation or inhibition of connected brain structures depending on visceral vagal afferents (NTS and locus coeruleus).
• The search for areas to be stimulated on the auricle is facilitated by local hypersensitivity, as at the somatic level (cellulo-periosto-teno-myalgic syndrome).
• The modification of interoceptive information could account for the pathophysiological mechanism.

‍