Keywords

1 Neuroanatomy of Pain: A General View

Wilder Penfield and colleagues lead the way to our current understanding regarding the principles of the cortical representation of somatosensory input. Their spectacular discoveries led to the famous somatosensory homunculus, a distorted scaled model of the human body neurally arranged at the postcentral gyrus [1, p. 1721]. Later on, cortical motor, sensory, and speech areas were discovered based on electrical stimulation of respective brain areas within the context of presurgical examination in epileptic patients [2]. Interestingly, those pioneering works revealed no distinct pain responses. Penfield and Jasper noted that some patients expressed feeling sensations best described as prickling or tingling and slightly unpleasant, but not painful at all [3].

With the advent of noninvasive neuroimaging techniques, this spectacular initial work was extended and resulted in extensive knowledge incorporating the whole signaling cascades from peripheral somatosensory transduction mechanisms to their repository in different cortical and subcortical areas. Over the decades, profound explorations regarding neuroanatomical and functional aspects have led to the identification of a pain-associated neural network often denoted as a “pain matrix” [4, p. 883] or in more recent publications as a “neurologic signature of physical pain” [5, p. 1072]. This network can be summarized as a confluence across a huge amount of clinical and basic research with a main focus on pain. Fig. 6.1 and Table 6.1 provide a schematic according to brain areas observed to be involved in coding the whole experience of pain.

Fig. 6.1
figure 1figure 1

Schematic illustration of cortical and subcortical areas found to be incorporated in the processing of experimental and chronic pain (see corresponding Table 6.1 for anatomical and functional description)

Full size image
Table 6.1 Corresponding anatomical and functional descriptions of areas within the “cortical functional pain circuit” delineated in Fig. 6.1
Full size table

For quite some time, a rather deductive approach leads clinicians and scientists to believe that pain was primarily a nociceptive phenomenon and thus assignable to just a few distinct cortical areas across the pain matrix, with the primary somatosensory cortex playing the major role. This obvious simplification was on the one hand attributable to methodological constraints but also due to a primary “sensory-guided” view of pain processing. After all, sensory input is sensory by nature and thus should be processed within somatosensory areas. Thus, it logically followed that pain, as the strong(est) sensory sensation, must be localized and most pronounced within those sensory regions [6, p. 913, 7, p. 1145].

Since then, our knowledge regarding the underlying principles of brain-related pain processing have broadened substantially. In particular, the importance of a multifaceted perspective on the topic became evident. Today, even a simple definition of pain has become challenging. Currently, the International Association for the Study of Pain (IASP) suggests the following: An unpleasant sensory and emotional experience associated with actual or potential tissue damage, or described in terms of such damage (http://www.iasp-pain.org/terminology). Of course, there are lots of additional contents listed, but the key element is evident by focusing on the term experience.

Indeed, pain has to be interpreted as a “global experience,” being hardly definable by anyone single, distinct neural constituent. Current views and explanations regarding human pain take into account multiple, complexly intertwined systems including sensory, motor, attentional, and cognitive neural processes [8, p. 1874, 9, p. 887]. Further, there is little doubt that pain has a strong and unique attentional activation quality that channels feelings, emotions, and thoughts in a specific direction, preserving the negative thoughts related to the possible events and consequences surrounding a specific pain experience [10, p. 1820]. But despite pain being a mostly unpleasant experience, it is probably the most important aspect of somato-sensation [11, p.1667, 12, p. 958]. We need pain in order to be protected from injury and tissue damage. Without an intact, functioning pain system, it’s challenging for an individual to stay healthy and free from injury. Indeed, reports on people suffering from congenital insensitivity to pain quite remarkably demonstrate the consequences of this statement. Several cases describe affected individuals who learned to live with their handicap and long-term outcomes were worsened by severe orthopedic complications mostly as a result of untreated skeletal injuries sustained in childhood [7, p. 1145, 13, p. 2017, 14, p. 2018].

Based on the above reasoning, when one considers the areas delineated in Fig. 6.1/Table 6.1, it now makes sense why there are numerous brain regions that participate in the experience of pain. There is a complex interplay of sensory, vegetative, emotional, motor, and cognitive aspects of pain, and it is now evident that the brain as a whole is challenged to adequately deal with such a global experience.

2 Acute Versus Chronic Pain

A healthy somatosensory system is one that is optimally equipped to accurately process pain, meaning that temporally, the pain diminishes either simply through the passage of time or by administration of medication. But some people are less “lucky”; in some cases, the pain remains and turns into a disruptive chronic entity, sometimes accompanied by severe comorbidities that become very difficult to treat [15, p. 613, 16, p. 1803, 17, p. 2021]. There are several open questions regarding this maladaptive development, and the associated risk factors remain poorly understood [18, p. 2027]. The rather simplified assumption was that chronic pain results from either constant nociceptive activation or central/peripheral somatosensory system damage which was generally summarized as neuropathic pain [19, p. 2028]. Others suggested that the chronicity derives from structural reorganization within the spinal cord and associated brain regions due to either the stimulation of long-lasting intense pain or severe psychologically/environmentally coincident stress experienced when the injury was sustained [20, p. 980, 21, p. 1091, 22, p. 1157]. More recent views suggest a complex intermingling of structural alterations in combination with disturbances of default mode networks (DMN) and connectivity patterns, which are probably interdependent on each other [23, p. 1820, 24, p. 1917]. Additionally, cortically localized “risk factors” have been theorized to exist in the form of deviant circuits incorrectly connecting relevant cerebral areas that could underlie a person’s vulnerability to develop the manifestations of chronic pain. Additionally, brain regions not classically considered to be associated with pain processing, structures located within the corticolimbic system (dorsal medial prefrontal cortex-amygdala-nucleus accumbens, ventral medial prefrontal cortex-amygdala, orbitofrontal cortex-amygdala), have been hypothesized to be key regions involved in the development of an individuals increased risks or susceptibility to chronic pain [18, p. 2027].

2.1 Example

A very interesting investigation by Mutso et al. [25, p. 2019] took a look at the structural and functional mechanisms underlying the transition from acute to chronic back pain. Based on the results of previous animal research, they assumed that the hippocampus would be the locus where structural and connective deviations would lead to the generation of the maladaptive circuitry, which is ultimately responsible for the switch from subacute to chronic pain. Indeed, they observed a significant involvement of hippocampal and prefrontal areas during the transformation from subacute (1–4 months) to chronic back pain (>10 years). The most severe alterations were observed in the structural reorganization of the hippocampus itself in addition to unbalanced connectivity patterns between the hippocampus/amygdala and prefrontal areas. Strikingly, experimentally driven acute pain studies very rarely report the hippocampal area as being activated.

However, another hypothesis favors the existence of a dynamic pain connectome, which exists as a spatiotemporal neural signature involving a variety of brain networks that communicate in a distinct fashion to integrate all the aspects of the pain experience. This model seems to represent the most accurate view we have to date (adapted from [10, p. 1820]). Table 6.2 summarizes the suggested networks and assigned functions to the areas involved in this theory.

Table 6.2 A possible dynamic connectome regarding cortical pain processing including suggested networks, supposed functions, and associated areas
Full size table

It is not entirely clear which underlying mechanisms drive the alterations in network connectivity and increases or decreases in neural center activity. There are indications of aberrant DMN (default mode network) characteristics in several chronic pain states, but caution is advised in terms of conclusive causal interpretations of the reported observations. Also, alterations in connectivity strengths between respective areas/networks are discussed in both directions (amplification or mitigation) again in several chronic pain states [25, p. 2019, 26, p. 266, 27, p. 135, 28, p. 2118]. The underlying basic systems—sensorimotor, default mode, salience, and nociceptive—can further be interpreted as being involved in other daily behavior regulation processes we only begin to understand in detail [29, p. 1472]. Thus, despite the concept of a dynamic pain connectome representing an intelligible approach for explaining the proneness to pain chronification and its neural manifestations, it is still challenging to allocate the different facets of pain specifically to a single areas connectivity changes in resting state network architecture or structural changes in areas belonging to the neural signature of pain.

Thus, to summarize this short general sketch, based on what we know from brain studies, both human and animal, chronic pain seems rather maintained by cortical areas, networks, and circuits whose functions are assigned to all sensory, vegetative, as well as emotional and cognitive processes. This might be a result of a maladaptive learning and association process [30, p. 987] or the amplification within pre-vulnerable systems possibly manifested in aberrant fronto-limbic structures and related processing [18, p. 2027].

3 Neuroanatomy of TMJ-Related Pain

The neural arrangement of the temporomandibular joint is—from a non-nociceptive somatosensory perspective—quite unequivocal: the locus is within the face area of the primary somatosensory cortex, tightly adjoining the hand area. Aside from the hand/finger area, the face area is the largest representation in this part of the brain, correlated with the associated peripheral receptor densities reflected in the sensitivity of the tongue, teeth, lips, nose, eye, and the skin of the human face in general. Thus, this system constitutes the main afferent pathways for all somatosensory processing regarding these structures, including the mandibular joint as well as associated muscles and tendons. Considering the pain-related neural signature delineated in Fig. 6.1, it’s important to note that trigeminally mediated nociceptive input is underrepresented compared to pain evoked at other body sites [31, p. 1506, 32, p. 1950]. This is quite astonishing as trigeminal pain and associated burden, suffering, and costs would justify enormous effort toward examination of underlying brain processing. In principle, there are two options to pursue to unravel this paradox, either one can study chronic TMJ conditions or focus on studying healthy volunteers in experimental orofacial pain models.

3.1 Experimental Approaches

Investigating healthy human subjects in experimental orofacial pain models in a standardized setting represents an important branch of research aimed to elucidate fundamental mechanisms of associated cortical pain processing. In this vein, toothache can be utilized as an acute form of orofacial pain, thus providing the researcher with an ideal experimental orofacial pain paradigm to evaluate trigeminally mediated cortical activation and response patterns. Indeed, several reports have been published applying either painful or painless stimuli to a least one tooth, while concomitantly recording brain responses. The modalities range from tactile/vibrotactile, electric, and air stimuli. An overview is given in Table 6.3.

Table 6.3 Studies combining tooth stimulation and neuroimaging
Full size table

To summarize, brain response patterns in response to tooth stimulation strongly resemble those from experimental pain applied to extra-trigeminal sites, especially stimulation at painful levels. This means that the associated neurological signature encompasses the areas illustrated in Fig. 6.1, however, with a number of obvious peculiarities.

Focusing firstly on somatotopic cortical organization aspects, the study by Jantsch et al. [34, p. 683] demonstrated S1 activity contralateral to hand pain compared to bilateral activity during tooth pain. This bilateral activity pattern might be related to the fact that the stimulated left incisor tooth is close to the body midline, whereas the hand is clearly more distal. It is critical to be cautious about this result as ideally, both incisor teeth should be stimulated to conclusively prove this assumption. The work conducted by Brügger et al. [31, p. 1506] addresses this issue by stimulation of the left/right maxillary canines and central incisors. A direct comparison between central incisor and canine stimulation revealed a more prominent tendency toward contralateral S1 activity for canines compared to central incisors. This finding implies a certain cortical lateralization scheme related to the distance from the body midline. Both findings support the concept of somatotopic organization within S1 also for teeth, however, with the limitation that this somatotopic pattern was induced by pain and not by painless somatosensory stimulation.

Utilizing MEG (Magnetic Encephalography) as an alternative method, the work by Kubo et al. [36, p. 1074] compared painless stimulation of the right maxillary first premolar with the median nerve of the right wrist. The findings also revealed bilateral activity in a region the authors termed “parieto-temporal” area. But when looking at the sources of this activity, a contralateral main focus located in the central sulcus (S1) was observed for tooth and wrist stimulation with a slight posterior/superior shift of tooth stimulation. Further fMRI-based evidence investigating possible somatotopy of the intraoral area is further supported by results of Miyamoto et al. [35, p. 1075]; their protocol also involves applying painless stimuli to the right upper central incisor tooth, the lower lip, and the tongue. They used a rubber tip and administered the stimuli manually. Clearly overlapping S1 activity was observed with a lip-tooth-tongue gradient from superior to inferior in rostral S1 subareas of their activity cluster. Important to note only the contralateral S1 region was investigated; therefore it is not possible to compare their results with the Jantsch/Brügger/Kubo reports as they observed a general bilateral activation pattern but with lateralization tendencies. Importantly, the report by Brügger et al. [31, p. 1506] demonstrated a robust bilateral activation pattern with further contralateral tendencies in the thalamus, in the posterior insula, around the parietal operculum (BA 43), and surprisingly in the amygdala. The finding of increased contralateral activation of the amygdala has to be specifically brought to attention as it is the only experimental human pain report demonstrating such a pattern. For example, a review by Baas et al. [46, p. 1102], summarizing 54 studies with amygdala activity, revealed no clear lateralization effect and highlighted the main functional contribution of the amygdala in processing primarily negative affective states such as fear and anxiety but no somatosensory encoding properties. On the other hand, Neugebauer [47, p. 1104] and Neugebauer [48, p. 1103] found evidence in rats that this structure consists of a so-called nociceptive amygdala located in the latero-capsular division of the central nucleus which directly processes sensory input. Yet, in humans, this has yet to be clarified although the study by Brügger and colleagues opened the window toward the amygdala’s possible direct involvement in decoding somatotopic information. A possible explanation of this finding might be that tooth pain induces higher levels of threat/anxiety than pain originating from other parts of the human body, requiring the aberrant recruitment of additional brain structures to somato-topically encode the afferent sensory signals. A recent report by Meier et al. [49, p. 1436] substantiated this presumption by demonstrating enhanced conditioned fear induced by a short tooth pain stimulus compared to pain administered to the tibia. However, it must not be forgotten that functional measurements of such small subareas require specific imaging strategies especially when the amygdala is targeted. Mainly, this is due to its central localization, the surrounding vasculature and bordering cerebrospinal fluid. Those facts are accompanied by strong phase-encoding susceptibility inferences leading to false positive and negative activation patterns unrelated to a specific stimulation or task [50, p. 1101, 51, p. 1578].

Besides the information of “where” does it hurt and the “how much” does it hurt is—at least—of comparable importance. There is one report addressing this question directly by applying five different stimulus strengths to a right maxillary canine, whereas two were painless, and the remaining three were painful [41, p. 84]. Also, the study of Jantsch et al. [34, p. 683] can be interpreted as intensity coding as they applied “weak” and “strong” pain, however, no painless stimuli were used. An alternative approach was used by Trulsson et al. [38, p. 1073] also focusing on “intensity coding” by applying tactile stimulation of different frequencies to the left maxillary incisor, but no painful stimulation specifically. Cortical correlates of somatosensory intensity coding have been demonstrated across the literature, most particularly in the subareas of the insular and cingulate cortices (i.e., [52, p. 916, 53, p. 1144, 54, p. 919]). Beginning with the study of Brügger et al. [41, p. 84] which applied the whole range of perception from painless to painful, the anterior insula together with two cingulate cortex subareas, namely, the anterior mid- and pregenual anterior cingulate cortex, demonstrated a significant linear relation between applied stimulus strengths and activity levels. The observation that the insular cortex plays a crucial role in intensity coding was also demonstrated by the Jantsch study, however; the insula also showed a stronger activity pattern during the hand stimulation, which may be due to the varying modalities used in their paradigm (electrical for tooth and mechanic for the index finger). Interestingly, the Trulsson study revealed generally stronger activity when applying higher tactile stimulus frequencies (100 Hz) and additional insular-opercular subarea activity, but only contralateral to the stimulation site. The same pattern was observed in the right cerebellar cortex. Generally, these results fit well into known cortical response patterns to stimuli of extra-trigeminal origin. But recent elaborations of cortical systems coding specifically for different strengths of somatosensory input suggest a multisensory magnitude—instead of a specific pain-related assessment module, particularly within the insular cortex [21, p. 1091, 55, p. 417, 56, p. 920, 57, p. 1415, 58, p. 206]. In our opinion, this line of reasoning is understandable, as somatosensory stimulation of, for example, the human back induces a multitude of perceptions due to a variety of different receptor types transmitting the whole range of sensory modalities. On the other hand, teeth are unique “organs” consisting of hard mineralized material surrounding densely innervated and vascularized soft tissue located within the tooth pulp. The pulp itself is predominantly innervated by C and A-delta fibers, implying that neural inputs are (1) of mostly nociceptive characteristic and (2) of rather homogenous perceptive quality [31, p. 1506, 59, p. 897]. This physiological specificity makes the tooth an ideal stimulation target in order to investigate more thoroughly the “pure pain perception” and probably also the processing of the intensity level of pain. Two studies can be considered in this vein: [42, p. 46] and [44, p. 1504]. The first investigated patients suffering from dental hypersensitivity in response to an application of an air stimulus sufficiently strong to evoke pain. A sensitive as well as an insensitive tooth were stimulated and patients were required to focus selectively on their intensity perception. Surprisingly, intensity coding related activity was observed in a multitude of areas, including anterior insular and mid-cingulate subareas (see results section for details). Those two regions were also found to specifically code for the sensitive tooth with clearly stronger activity, providing the evidence that these areas seem to have pain specific functions beside the intensity coding properties. To substantiate this finding, a follow-up study by the same group addressed this issue with an elegant approach [44, p. 1504]. Using electric stimuli at a constant intensity applied to the left mandibular canine, they injected an anesthetic drug (articaine) to block afferent signaling transmission while the stimulation continued at the same intensity level. Over time a gradual pain decrease was perceived despite the ongoing painful stimulation. The specific brain response pattern in reaction to the articaine-induced dental pain relief was observed in a small portion of the left posterior insula, reiterating the critical role of this brain area in coding pain specificity and related intensity coding.

From a neurochemistry perspective, there is strong evidence subserving the idea of the insula is key structure within the cortical dental pain circuitry derived from two studies using fMRS (functional magnetic resonance spectroscopy) [40, p. 882, 43, p. 1152]. In the first attempt, the whole left insular cortex was measured during continuous stimulation of the right maxillary canine. Significant increases were measureable in the levels of glutamine (Gln) and the glutamine-glutamate complex (Glx) together with a significant drop in myo-inositol (mI). The second report investigated the insular cortex bilaterally, and they subclassified the insula into an anterior and posterior portion using the same paradigm. Comparable to the first study, Glx, Gln, and also glutamate (Glu) showed a significant increase during the pain stimulation phase, whereas mI significantly dropped. This pattern was observed in all four subareas. An interesting effect was found in significant differences between left and right insular subareas irrespective of stimulation or rest. As the insular cortex is incorporated in a manifold of different cortical functions, this fundamental disparity might support the suggestion that there are inherent functional differences between the subareas as pointed out by several investigations [20, p. 980, 60, p. 976] or [54, p. 919]. Neurochemical alterations within subareas of the right insular cortex have been shown by applying heat pain to an inner left forearm area. However, they measured a dorsal-anterior area, thus a rather evaluative-cognitive region than sensory encoding. Regardless, they revealed partly comparable reaction patterns with respect to increased Glu levels. None of the other metabolites demonstrated a pattern related with the stimulation. The stimulus related measurement of neurochemistry (event-related MRS or functional MRS) is complementary to fMRI in brain imaging. To date, only four studies have applied experimental pain while measuring changes in neurochemical compositions, thus, a lot of ambiguity remains that has to be investigated in more details.

The group of De Matos et al. [45] attempted a closer investigation into the fundamental nociceptive processing within the CNS. Applying the paradigm used in other studies by this group, the brainstem trigeminal nuclear complex (BTNC) was targeted while administering painful electrical stimuli to the right maxillary canine and neurochemical alterations during pain vs baseline were assessed. The BTNC constitutes the first CNS relay along the peripheral-central signaling cascade and therefore enables the investigation of pain-related processing at a very initial level. As the main result, a significant decrease in NAA and GABA during experimental orofacial stimulation was found. To date, a conclusive summary regarding this neurochemical pattern is not advocated by the authors as the results need clarification by further investigations. While of particular interest regarding early nociceptive processing, the study demonstrated above all, the possibility to measure the human brainstem neurochemistry with high accuracy, thus paving the way to a better understanding of this important brain area in the context of acute pain processing as well as pain related chronification mechanisms.

4 Summary

Pain is a multidimensional experience incorporating sensory, motor, affective and cognitive components. This applies to pain in general, as well as to orofacial pain in particular. This chapter provided an overview of neural signatures based on experimental acute orofacial pain up to explanatory approaches possibly underlying chronification mechanisms and chronic pain.

Summarized, brain responses of acute orofacial pain are quite well characterized based on sophisticated experimental models combined with imaging methods such as fMRI, fMRS, MEG and EEG. This strategy is important as such experiments allow standardized and controlled application of pain stimuli. In this way it is possible to understand the fundamental neural processes of experiencing pain, which in turn is the prerequisite for understanding the much more complex chronic pain. Concluded, these experiments revealed that several brain areas, often termed “Pain Matrix” or “Neurological Signature of Pain”, representing the neural framework regarding the multidimensional facets of an acute pain experience.

Still a challenge to understand are neural underpinnings of chronic pain and associated mechanisms that facilitates the transition from acute to subacute and finally chronic pain. Recent investigations suggest that chronic pain involves additional areas not known as classic pain areas (hippocampus) or propose a highly “Dynamic Pain Connectome” linking attention and pain related brain areas such as Salience-, Default Mode and Antinociceptive networks.

Recent years of intensive basic and clinical research has not yet brought the solution, but we are on good terms to comprehend the basic processes of pain chronification better and better. Including multimodal approaches that measure and quantify different facets of brain function, together with improved analytical methods (i.e. deep learning and big data management), a much better understanding of this highly significant global health problem is closer than ever before.