Left and right hemispheric lateralization of the amygdala in pain.

Pain is an unpleasant sensory or emotional experience associated with actual or potential tissue damage, or described in terms of such damage, as defined by the International Association for the Study of Pain ( Loeser and Treede, 2008 ). The amygdala, due to its well-defined role in emotional processing and more recently described involvement in nociceptive processing, is uniquely positioned to influence the experience of pain ( Andreoli et al. , 2017 ; Carrasquillo and Gereau IV, 2008 ; Ji and Neugebauer, 2009 ; Neugebauer and Li, 2003 ; Veinante et al. , 2013 ). In the context of human neuroimaging, many studies have identified the amygdala as an important brain region activated in the context of a variety of pain modalities and demonstrated that the amygdala undergoes changes as a result of both experimentally induced and chronic pain ( As-Sanie et al. , 2012 ; Simons et al. , 2014 ). As described above, the central nucleus of the amygdala (CeA) is the main output nucleus of the amygdala; the CeA receives and integrates both nociceptive input and polymodal emotional input from a variety of brain regions, modulating both pain and affect ( Figure 1 ). Although other amygdala nuclei (lateral and basolateral amygdala, as well as the bed nucleus of the stria terminalis, referred to as the “extended amygdala”) are also involved in pain, this review focuses on the role of the CeA in pain modulation. It is important to note that while it is fairly easy to discuss animal studies that specifically investigate the CeA, human neuroimaging studies are more complicated. The resolution of neuroimaging techniques is not fine enough to pinpoint different amygdala subnuclei. Therefore, we can only narrow down human imaging data presented to the level of “amygdala”, which may include basolateral, lateral, and central amygdala. As the studies described investigate pain, it is reasonable to assume results discussed apply to the CeA, even if the spatial resolution is not CeA specific. As a general rule, the right amygdala seems to play a more pro-nociceptive function with the left amygdala having no role, a reduced pro-nociceptive function, or in some studies a novel anti-nociceptive role in the context of acute and chronic injury. As we describe and analyze the existing literature below, we will pay special attention to studies that do not fit well into this lateralization model. As we will see, some of the inconsistencies that have been observed appear to be associated with different types of pain, such as somatic versus visceral pain, or the targeting of different cell types within the amygdala. Bilateral brain regions are often assumed to have the same function. As a result, experiments often utilize either bilateral or side-randomized manipulation to explore the role of these areas. Both of these approaches prevent accounting for the possibility that the two brain regions may actually serve different purposes or not contribute to the experimental outcome equally. For example, rats with neuropathic pain have increased amygdala cell volume and cell number two months after injury, and animals exhibit increased depression-like behavior, analogous to the increased depression humans with prolonged pain develop ( Dworkin and Gitlin, 1991 ; Gonçalves et al. , 2008 ). However, in that study, the authors neglected to separate data by side, so it is impossible to determine if the increased amygdalar volume is unique to the amygdala hemisphere that is contralateral to the neuropathic injury or if it also seen in the ipsilateral amygdala hemisphere with respect to the injury. If and how the left and right amygdala contribute equally to pain remains relatively understudied and poorly understood. Here, we review existing literature that investigates the role of the left and/or right CeA in the context of pain in order to gain insight into how these hemispheric brain regions interact in order to modulate pain. We have organized this information based on the pain model used. As described above, the most direct path for nociceptive information from sensory neurons to reach the amygdala is through the spino-parabrachio-amygdaloid pathway. Parabrachial nucleus (PBN) projection neurons to the CeA are of great interest when investigating pain modulation by the CeA, as this input contains relatively pure nociceptive information that has not been highly processed by other cortical and thalamic brain regions ( Figure 1 ). This PBN circuit is critical to fear conditioning studies in rodents using nociceptive unconditioned stimuli ( Sato et al. , 2015 ). Tracing studies in rats suggest that the left and right CeA receive equal, primarily ipsilateral input from the PBN ( Bernard et al. , 1993 ). During stimulation of both the left and right periphery, nociceptive signals reach both the left and right PBN and left and right CeA. Newer evidence suggests that even unilateral stimulation of the periphery activates both the left and right PBN in rat ( Miyazawa et al. , 2018 ). The potential for each PBN to send nociceptive information to the ipsilateral CeA makes differences observed in left and right CeA activity increasingly more interesting. Unilateral injection of formalin in the face of rats increases neuronal activation ( c-Fos ) bilaterally in the PBN, but the right CeA has consistently higher neuronal activation than the left CeA, regardless of the side of inflammation ( Miyazawa et al. , 2018 ). Neuroimaging in mice has also revealed that right amygdala activation is increased following formalin in the left paw, and chemogenetic inhibition of the right amygdala decreases formalin-induced activation of multiple bilateral brain regions, suggesting that right amygdala activation in necessary for bilateral inflammatory pain-induced changes in activation ( Arimura et al. , 2019 ). The right CeA shows disproportionately higher activation of extracellular signal-regulated kinase 1/2 (ERK1/2) after formalin as well, regardless of the side of formalin-induced paw injury in mice ( Carrasquillo and Gereau, 2007 ). Pharmacologic inhibition of ERK 1/2 phosphorylation in the right CeA, but not the left CeA, of mice is able to reverse inflammation-induced mechanical hypersensitivity in both paws ( Carrasquillo and Gereau IV, 2008 ). Interestingly, inflammation-induced pain causes molecular changes in the central amygdala as well as behavioral effects that vary depending on age; middle-aged and old mice not only exhibit more nociceptive spontaneous behaviors in response to formalin injected in the right rear paw and develop persistent pain, but they also exhibit increased levels of activated ERK 1 and metabotropic glutamate receptor 5 (mGluR5) in the right but not left CeA compared to young mice ( Sadler et al. , 2017a ). Inherent differences in expression of various proteins and receptors likely contribute to the right CeA’s dominant role in the modulation of inflammatory pain, but little has been done to compare gene expression differences across the right and left CeA. mGluR5 represents a notable example of the potential for baseline expression differences to drive lateralized responses. mGluR5 is intrinsically more highly expressed in the right CeA, and mGluR5 activation in the right CeA of mice is sufficient to induce ERK activation ( Kolber et al. , 2010 ). Further, mGluR5 inhibition in the right CeA prevents both ERK activation and formalin-induced pain-like hypersensitivity ( Kolber et al. , 2010 ), indicating that activation of mGluR5 is necessary for ERK activation and therefore paw inflammation-induced hypersensitivity in both the left and right hind paw. Inflammation-induced hypersensitivity of the right hind paw is also associated with an increase in kappa opioid receptor (KOR) binding in the ipsilateral CeA in mice, in contrast to other opioid receptors, which are downregulated in the context of injury ( Narita et al. , 2006 ). Although the left and right CeA receive equal input from the PBN, rodent studies suggest that transmission at the PBN-CeA synapse is potentiated only in the right CeA in the context of inflammatory pain, regardless of the side of injury ( Miyazawa et al. , 2018 ; Sugimura et al. , 2016 ). Inflammation-induced potentiation is dependent on the presence and activity of calcitonin gene-related peptide (CGRP) ( Shinohara et al. , 2017 ). CGRP is a neuropeptide involved in pain processing throughout the central and peripheral nervous systems and is highly expressed in neurons along the spino-parabrachio-amygdaloid pathway, including those that project from the PBN to the CeA ( Dobolyi et al. , 2005 ; Lu et al. , 2015 ). In mice, CGRP-positive input from the PBN to CeA is involved in nociceptive transmission in the right CeA ( Lu et al. , 2015 ), and infusion of CGRP into the right CeA is able to increase pain-like behaviors in naïve rats ( Han et al. , 2010 ). Electrophysiology experiments in rodents suggest that CGRP drives potentiation in the right CeA by increasing the amplitude of NMDA-mediated EPSCs ( Han et al. , 2010 ; Okutsu et al. , 2017 ), most likely through CGRP receptor-driven activation of protein kinase A (PKA) ( Han et al. , 2005 ). In another study, pharmacological infusion of CGRP into the left CeA of naïve rats increases mechanical hindlimb threshold (side not specified), suggesting an anti-nociceptive effect ( Xu et al. , 2003 ). It is possible that CGRP has different roles in the left and right CeA, or its role could depend on pain modality or chronicity. The modulation of inflammatory pain may be lateralized to the right CeA due to intrinsic differences in expression of CGRP, CGRP receptors, or some downstream signaling molecule, but these potential differences remain unexplored. Despite overwhelming evidence for the involvement of the right CeA in driving inflammatory pain, an analgesic neural ensemble activated by general anesthesia (GA) has recently been discovered in the bilateral CeA (CeA GA ) of mice ( Hua et al. , 2020 ). In the context of inflammatory pain, activation of right CeA GA neurons decreases formalin-induced pain-like behaviors while inhibition of right CeA GA neurons restores formalin-induced pain-like behavior, regardless of the side of injury ( Hua et al. , 2020 ). Whether or not activation of left CeA GA has similar effect in inflammatory pain has not been explored. However, activation of left or right CeA GA neurons were shown to have analgesic effects at baseline, decreasing mechanical, heat, and cold sensitivity bilaterally in the hind paws ( Hua et al. , 2020 ). Interestingly, activation of right but not left CeA GA neurons decreases baseline facial mechanical sensitivity, suggesting that there is some aspect of hemispheric lateralization to the effects of these CeA GA neurons ( Hua et al. , 2020 ). Both human and animal studies demonstrate changes in the right amygdala as a result of arthritic pain. Structural MRI data from patients with rheumatoid arthritis shows decreased right amygdalar volume compared to healthy patients ( Wartolowska et al. , 2012 ). While rheumatoid arthritis is typically experienced bilaterally, this study neglects to mention if patients are experiencing rheumatoid arthritis bilaterally or only on the left of right side of the body. Functional neuroimaging in men and women with left or right knee osteoarthritis exhibit greater right amygdala activation in response to experimentally evoked pain via thermode application to the arthritic knee ( Kulkarni et al. , 2007 ). In preclinical animal models, neurons in the right CeA of male rats have a much larger receptive field than neurons in the left CeA. These right CeA receptive fields appear to be bilateral compared to contralateral receptive fields of the left CeA neurons. Additionally, arthritis-like injury increases the receptive field of neurons only in the right CeA and does not change the receptive field of left CeA neurons regardless of the side of injury ( Ji and Neugebauer, 2009 ), indicating a right CeA dominance in arthritic pain models. The right CeA also exhibits increases in neuronal activity and enhanced transmission of nociceptive input from the PBN in arthritic rats, no matter which side of the body arthritis is induced on ( Han et al. , 2005 ; Ji and Neugebauer, 2009 ; Neugebauer and Li, 2003 ; Neugebauer et al. , 2003 ). Similar to inflammatory pain, pain-induced plasticity observed in the right CeA following arthritis in rodents is dependent on both NMDA and CGRP receptor function ( Fu et al. , 2008 ; Han et al. , 2005 ). CGRP receptor activation in the right CeA increases NMDA-mediated excitatory currents at the PBN-CeA synapse in mice, but the mechanism driving this facilitation is still unclear ( Okutsu et al. , 2017 ). PKA activation is necessary for arthritis-induced increases in right CeA neuronal activity in rats ( Ji and Neugebauer, 2009 ), so it is likely that CGRP receptor activation facilitates activation of PKA in the right CeA, which has downstream effects on NMDA receptor function to enhance synaptic transmission ( Okutsu et al. , 2017 ). Non-NMDA receptors are likely also important in arthritis-driven increases in right CeA pain plasticity in rats, as they contribute to both nociceptive and non-nociceptive transmission under normal conditions and are enhanced in the arthritic condition ( Li and Neugebauer, 2004a ). Therefore, non-NMDA receptors may be important for mediating enhanced glutamatergic transmission to the right CeA that potentiates both nociceptive and non-nociceptive sensory signals in the arthritic state ( Li and Neugebauer, 2004b ); mGluRs 1 and 5 have increased expression in the right CeA in the arthritic pain state (right knee). mGluR5 in the right CeA contributes to the facilitation of synaptic transmission under control conditions and mGluR1 is involved in right knee arthritis-induced pain-like behaviors ( Han and Neugebauer, 2005 ; Neugebauer et al. , 2003 ). It is likely that mGluR1 is recruited in the arthritic state, contributing to pain-induced facilitation ( Neugebauer et al. , 2003 ). Although the right CeA seems to be dominant in the modulation of arthritis pain, a role for the left CeA in arthritic pain has been found in both humans and rodents. Male and female patients with knee arthritis exhibit increases in left amygdala activity measured via PET imaging during spontaneous arthritis pain compared to the pain free state, regardless of the side of arthritis injury ( Kulkarni et al. , 2007 ). This is in contrast to the increase in right amygdala activity seen in arthritis patients during experimental pain evoked via a thermode. Overall, these data suggest that the hemispheres may respond differently to distinct stimuli ( Kulkarni et al. , 2007 ). In rats, PKA inhibition reverses left knee arthritis-driven increases in neuronal activity in the right CeA but has no effect in the left CeA ( Ji and Neugebauer, 2009 ). Elevations in cAMP increases neuronal activity in both the left and right CeA, though, suggesting that the left CeA may have the ability to contribute to arthritis-induced pain modulation via cAMP signaling but has developed mechanisms to prevent arthritis-induced PKA activation ( Ji and Neugebauer, 2009 ). The role of the left CeA and the extent to which it is involved in arthritic pain is still unclear, and much more left CeA specific investigation must be conducted in order to elucidate its contribution to the modulation of arthritis pain. The right CeA has been noted for changes in activation and functional connectivity in humans and animals as a result of visceral pain. fMRI studies in healthy male and female patients reveal increases in right amygdala activation during noxious gastric distention ( Lu et al. , 2004 ). Interestingly, women with chronic irritable bowel syndrome (IBS) exhibit deactivation of the right amygdala during anticipation of rectal distention but stable activation during actual distention ( Naliboff et al. , 2006 ). Researchers have posited that amygdalar deactivation is a coping strategy for aversive but unavoidable stimuli, even if the activation during pain remains constant. Alternatively, differences in amygdalar activation and deactivation may be due to the subjects being in an acute versus chronic pain state. An fMRI study of female twins with and without bladder pain found that the women with bladder pain exhibit elevated connection between the right amygdala and the periaqueductal grey (PAG) as well as other brain regions during distention compared to the healthy twin ( Kleinhans et al. , 2016 ). Preclinical animal studies also find right amygdala involvement in visceral pain modulation. Rats undergoing colorectal stimulation exhibit increased neuronal activity in the right CeA measured via quantity of immediate early gene c-fos ( Lazovic et al. , 2005 ). fMRI studies also demonstrate increased right amygdala activation in response to noxious colorectal stimulation in female rats with estrogen dependent stress-induced visceral pain ( Hubbard et al. , 2016 ). Additionally, there is increased neuronal excitability and transmission at the PBN-CeA synapse in the right CeA of rats with zymosan-induced colitis ( Han and Neugebauer, 2004 ). Interestingly, the changes in intrinsic membrane properties that lead to this potentiation in the right CeA as a result of colitis in rats are different than those of arthritis pain ( Han and Neugebauer, 2004 ), indicating that pain-induced potentiation observed in the right CeA is a result of different mechanisms depending on the source of nociceptive information. Manipulation of neuronal activity in the right CeA is also able to influence visceral pain behaviorally and physiologically in rodents; mGluR5, already established as an important driver in the potentiation of pain transmission in the right CeA, is also vital to modulating visceral pain, as pharmacological blockade of right CeA mGluR5 is able to reduce bladder pain-like responses in mice ( Crock et al. , 2012 ). Pain-related changes in the left amygdala are also seen in visceral pain conditions; the fMRI study of twin women with bladder pain referenced above found that women with chronic bladder pain have increased connectivity between the left amygdala and PAG at baseline compared to the healthy twin, and that this connectivity then decreases after bladder distention ( Kleinhans et al. , 2016 ). Similarly, female patients with IBS exhibit increased activity in the left amygdala during rectal distention compared to healthy controls ( Berman et al. , 2008 ), and women with urologic chronic pelvic pain and endometriosis have increased gray matter volume in the left amygdala compared to healthy patients ( As-Sanie et al. , 2012 ). This increase is correlated with shorter symptom duration in the long term ( Bagarinao et al. , 2014 ), suggesting that these left amygdala gray matter changes are somehow involved in the symptomology of chronic pelvic pain. Further, the left amygdala of female patients with urologic chronic pelvic pain has increased functional connectivity with the posterior cingulate cortex ( Martucci et al. , 2015 ), a connection that is positively associated with urologic symptom as well as pain severity ( Martucci et al. , 2015 ). It is likely that the left amygdala contributes significantly to the affective components of visceral pain. Preclinically, acid-induced visceral pain has been found to increase c-fos activation in the left amygdala of rats compared to control animals ( Nakagawa et al. , 2003 ). Unfortunately, activation of c-fos in the right CeA was not measured in this study, so it is impossible to determine if this is a unilateral or bilateral effect and how this increase compares to the right CeA. Increases in left amygdala activation have also been found in response to noxious colorectal distention in female rats via fMRI ( Hubbard et al. , 2016 ). There is a scarcity of left CeA specific studies in visceral pain, likely due to the right CeA’s well documented role in other pain models, leading to an a priori bias to focus just on the right side. Human studies clearly indicate involvement of the left amygdala in visceral pain, though it’s role may be involved more in modulating the affective components and symptom severity. It is difficult to tease apart visceral hypersensitivity and affective symptomology in animal studies, so it remains challenging to determine the possible affective influence of the left CeA on visceral pain. Both the left and right amygdala contribute to visceral pain modulation, but a definitive role for each and how they work together remains largely unknown and unexplored. While increases in neuronal activity in the left and right CeA are likely indicative of pain induced changes contributing to modulating and maintaining pain, it is still unclear how these changes are influencing the pain experience. fMRI imaging reveals that the right amygdala of women with bladder pain sees an increase in functional connectivity with the PAG during distention, but the left amygdala exhibits a decrease in this same connectivity following distention ( Kleinhans et al. , 2016 ). A collection of neuroimaging studies beyond visceral pain supports the idea that the left and right amygdala have distinct roles in pain processing. A meta-analysis of 40 neuroimaging studies found that the right amygdala was consistently more active in experimental pain studies but the left amygdala was more active in clinical pain patients ( Simons et al. , 2014 ). It is possible that the left and right amygdala’s opposing patterns of dominance to experimental and clinical pain are the result of distinct contributions to either acute or chronic pain; it may be that the right amygdala is more involved in acute pain (i.e. experimental pain) while the left amygdala is dominant in chronic pain (i.e. clinical pain). Unilateral side specific manipulations of neuronal activity in the left or right CeA of animals results in unique and opposing differences in pain-like behavior and physiology. For example, nonspecific optogenetic activation of neurons in the left or right CeA reveals divergent functions of the two nuclei in the context of peripheral bladder stimulation. Activation of the right CeA increases physiological responses to urinary bladder distention and increases mechanical abdominal sensitivity in mice ( Sadler et al. , 2017b ) ( Crock et al. , 2012 ). This right CeA effect could be replicated with pharmacological application of pituitary adenylate cyclase-activating peptide (PACAP) prior to bladder stimulation. The PACAP effect was independent of the side of the body for which physiological measurements were made. PACAP had no effect on bladder pain-like changes when injected in the left CeA. However, physiological responses to urinary bladder distention also increased during optogenetic inhibition of the left CeA, suggesting an ongoing, analgesic output of the left CeA in naïve mice independent of PACAP ( Sadler et al. , 2017b ). Following bladder sensitization, optogenetic activation of the left CeA reduces abdominal mechanical hypersensitivity, a measure of referred bladder pain ( Sadler et al. , 2017b ). These data suggest that the analgesic capability of the left CeA could be harnessed to reduce pain ( Sadler et al. , 2017b ). Whether this is also the case in chronic pain is interesting to consider, as the shift from acute injury to chronic pain may also shift amygdala output and affect neuronal activity of the left and right CeA differently. Other studies have failed to find differences in pain modulation with left versus right modulation, though; bilateral activation of the CeA with cortisol pellets in rats is sufficient to increase reflexive responses to colorectal distention, but neither right nor left amygdala unilateral stimulation is sufficient to cause visceral hypersensitivity ( Tran and Greenwood-Van Meerveld, 2012 ). Similar to other pain models, neuropathic pain increases neuronal activity in the right CeA by facilitating transmission at the PBN-CeA synapse and potentiates pain-like behaviors ( Ikeda et al. , 2007 ). Mice with left leg neuropathic pain show increases in expression of corticotrophin-releasing factor (CRF) in the CeA (side not a variable) ( Andreoli et al. , 2017 ). CRF-expressing neurons are a main output population of the CeA ( Pomrenze et al. , 2015 ), and overactivation of CRF-expressing cells in the CeA is thought to be an important contributing factor in pain chronification ( Andreoli et al. , 2017 ). CRF neurons in the CeA are largely under inhibitory control of enkephalinergic (Enk) interneurons, which modulate amygdala projections and help modulate the descending regulation of pain via the PAG ( Paretkar and Dimitrov, 2019 ). However, the laterality of these effects and side-specific influence of the amygdala remains unexplored. Interestingly, the right CeA regulates neuropathic pain via differential activation of two non-overlapping cell populations. PKCd-expressing cells in the right CeA are sensitized after left-sided neuropathic injury, but somatostatin-expressing cells are inhibited by injury ( Wilson et al. , 2019 ). Chemogenetic activation of right CeA PKCd cells increases mechanical sensitivity, and silencing these cells in animals with neuropathic injury reduces pain, suggesting a pro-nociceptive function for these neurons and a role in sustaining chronic pain ( Wilson et al. , 2019 ). Furthermore, chemogenetic activation of somatostatin cells in the right CeA decreases left paw mechanical sensitivity while inhibition of right CeA somatostatin in naïve animals increases pain ( Wilson et al. , 2019 ). Whether this cell-type specific bidirectional modulation of pain extends to left amygdala remains unknown. Interestingly, the analgesic neural ensemble of CeA GA neurons recently discovered in mice is a heterogenous population that includes PKCd-expressing neurons ( Hua et al. , 2020 ). The extent of overlap between right CeA PKCd pain driving neurons in Wilson et al 2019 and the PKCd CeA GA neurons remains to be determined. Activation of right CeA GA neurons decreases bilateral mechanical sensitivity and spontaneous pain-like behaviors in mice with right side neuropathic injury ( Hua et al. , 2020 ). Whether or not left CeA GA neurons have a similar effect in neuropathic pain was not tested. The right CeA’s role in modulating classical behavioral symptoms of neuropathic pain is also highly dependent on the side of injury and sensory modality ( Cooper et al. , 2018 ). While the right CeA is vital for the development of mechanical allodynia, hyperalgesia, and cold hypersensitivity when the injury is in the contralateral (Ieft) leg, the right CeA is only necessary for modulation of hyperalgesia when the injury is in the ipsilateral leg; inactivation of the right CeA with lidocaine increases mechanical thresholds to von Frey as well as reduces withdrawal duration to pinprick and acetone only when the injury is in the left leg ( Cooper et al. , 2018 ). When the injury is in the right leg, though, inactivation of the right CeA only reduces pinprick withdrawal duration and does not affect von Frey or acetone responses ( Cooper et al. , 2018 ). Unlike in other pain models, group I mGluRs, including mGluR1 and mGluR5, in the right CeA of rats have little to no effect on left leg neuropathic pain-induced pain-like behavior. mGluRs 1 and 5 do, however, mediate emotional pain-like behaviors associated with nerve injury, including aversive place conditioning ( Ansah et al. , 2010 ). Similarly, kappa opioid receptors (KOR) in the right CeA are important for mediating the aversive components of left side neuropathic pain, determined via conditioned place preference, but have no effect on mechanical hypersensitivity (von Frey) in rats ( Navratilova et al. , 2019 ). Pain-like behaviors associated with left side nerve injury are accompanied by increases in right CeA neuronal activity at the PBN-CeA synapse in rats ( Gonçalves and Dickenson, 2012 ; Ikeda et al. , 2007 ); however, the mechanisms driving PBN-CeA potentiation may be different than other types of pain. In contrast to inflammatory, visceral, and arthritic pain, this potentiation does not seem to be NMDA receptor dependent, suggesting that the molecular mechanism responsible for increasing potentiation in the right CeA in the context of neuropathic pain is distinct from other types of pain ( Ikeda et al. , 2007 ). It is possible that NMDA receptors play a role in the maintenance of neuropathic pain through non PBN-CeA pathways, including input from the BLA projections or CeA interneurons ( Ansah et al. , 2010 ). Further investigation is needed in order to elucidate the specific mechanisms modulating different neuropathic pain symptoms in the right CeA, as they seem to be distinct from other types of pain. Though the left CeA is less often the main focus, it is also fundamental to the modulation of neuropathic pain. Inactivation of the left amygdala is required for reduction in mechanical allodynia induced by contralateral (right) leg neuropathic pain in rats ( Cooper et al. , 2018 ), and mGluR5 activation in the left amygdala is sufficient to induce right side only mechanical sensitivity in uninjured mice ( Kolber et al. , 2010 ). Ablation of left but not right CeA CRF projection neurons to the locus coeruleus decreases mechanical sensitivity following left leg neuropathic injury ( Andreoli et al. , 2017 ). Furthermore, the left amygdala exhibits an increase in spontaneous and evoked neuronal activity compared to the right amygdala following left side neuropathic injury, although this increase is not long lasting (2–6 days)( Gonçalves and Dickenson, 2012 ). Interestingly, in that same study neuronal activity in the right amygdala is increased compared to the left amygdala 14 days after nerve injury ( Gonçalves and Dickenson, 2012 ), suggesting that time-dependent differences in activation are important in the development of the chronic state after neuropathic injury. The time staggered activation of the left and right amygdala may be indicative of different functions as pain shifts from acute to persistent to chronic.

Pain, aversion, anxiety, and fear are processed in the amygdala and there is considerable overlap in the cell types involved in each. At the same time, the mechanisms involved in modulating different types of pain are not always consistent. The amygdala’s influence on pain modulation is by no means limited to the cell types or mechanisms discussed here, and substantial gaps in the circuitry, connections, and mechanisms involved in pain processing in the amygdala remain. Further, hemispheric lateralization of the amygdala in the context of pain is a new but quickly emerging area of study, and there is significant progress still to be made. Currently, amygdala lateralization is most clearly observed in conditions where the pain is not limited to one side of the body, such as in visceral pain conditions like bladder or colorectal pain. Human and animal studies of visceral pain indicate differential activation patterns ( Kleinhans et al. , 2016 ; Lazovic et al. , 2005 ; Lu et al. , 2004 ; Naliboff et al. , 2006 ) and suggest imbalanced or even opposing functions of the left and right amygdala ( As-Sanie et al. , 2012 ; Berman et al. , 2008 ; Sadler et al. , 2017b ). Such clear hemispheric lateralization is not so easily observed in other pain models, though. Decussation of sensory afferents complicates the interpretation of lateralization when the pain is confined to a single side of the body, such as inflammatory, neuropathic, or arthritic pain models that are concentrated in one limb. Seemingly contradictory or asymmetrical roles of the CeA in relation to pain modulation may stem from simple differences in cell type or receptor abundance or perhaps variations in downstream pathways that lead to functional differences between hemispheres. It is also likely that the amygdala is not the only pain/affective brain area to show functional lateralization, and such phenomena may be observed upstream or downstream in the pain processing pathway. With all of the evidence for side-specific amygdalar differences in neuronal activity and plasticity, lateralized studies are especially necessary to fully elucidate the changes that happen in the amygdala during pain and how specific cell types in each hemisphere change and influence each other. When evaluating any example of functional lateralization, it is valuable to consider the purpose of such lateralization. Lateralization has been hypothesized to (1) deal with functional incompatibility when interacting with the environment (i.e. attending to both the changing and non-changing parts of the environment) (2) provide opportunities for multi-tasking (3) improve success or performance on tasks and (4) help influence whether approach or avoidance behaviors are initiated. Here, we present some speculative ideas about lateralization of the amygdala in the context of pain. First and foremost, we consider lateralization in the context of acute pain and injury ( Figure 2 ). During acute injury, neuronal activity and excitability in the right CeA increases, driving pain and therefore motivating evolutionarily beneficial escape or harm reduction behaviors. As the pain persists, overactivation of neurons in the right CeA fades and returns to baseline spontaneous activity, and neuronal excitability in the left CeA increases to drive pain relief. Such an effect may be driven by differences in temporal activation of specific “pro-“ and “anti-nociceptive” cell types, such as PKCd and somatostatin ( Wilson et al. , 2019 ), or it may be the result of similar cell types but different mechanisms, including activation of different downstream signaling cascades. This balancing act between activation of the left and right CeA is therefore able to modulate acute pain, which is important for detecting possible damage; nociceptive signals warn of potential harm and protect damaged tissue during healing, but they eventually cease when the stimulus is removed or healing is complete. In some circumstances, such a lateralized system might be engaged to actually drive left amygdala analgesia. This might occur during a stressful or life-threatening situation in which suppression of the pro-nociceptive right CeA would provide the opportunity to attend to survival over pain. Here, the overarching hypothesis is that lateralization in the amygdala provides for fine-tuning the immediate response to injury and how that response changes over time. This hypothesis harkens back to explanations for functional lateralization in language areas in the brain providing for specialization over redundancy. In chronic pain conditions, the balance of this system is likely disturbed, resulting in persistent pain. This imbalance may be due to 1) maintenance of neuronal hyperexcitability in the right CeA that is able to surpass the pain blocking effect of the left CeA, or 2) lack of activation of left CeA neurons to counteract the right CeA’s pain-driving effect. Amygdala lateralization in the context of pain is only beginning to be acknowledged and explicitly explored, and there are a multitude of different elements to consider that will vary greatly based on pain modality. We acknowledge that this hypothesis remains relatively broad, but the specific intricacies necessary to fully understand how amygdala lateralization functions across pain models will only be clarified with further studies that specifically address side as a variable.

“Nous parlons avec l’hémisphère gauche.” These immortal words concluded Paul Broca’s mid-19 th century treatise on the lateralization of language function in humans. Broca found that patients who suffered from left brain injuries demonstrated interruptions in their speech production (i.e. motor output) ( Ocklenburg and Gunturkun, 2017 ). However, other patients who suffered from the same injuries on the right side of the brain did not show the same speech deficits ( Ocklenburg and Gunturkun, 2017 ). In another series of famous studies, German neurologist Carl Wernicke identified a distinct left temporal area responsible for speech fluency and syntax ( Rogers et al. , 2013 ). These experiments provided a foundation for neuroscientists and psychologists to begin to explore anatomical specialization within the brain and how lateralization can influence diverse behaviors and functions. There is a direct line from these early observations in humans to the explosion of lateralization data in the mid 20 th century ( Ocklenburg and Gunturkun, 2017 ) in a variety of organisms and to this review. We focus on exciting new data that has emerged in the last 20+ years showing that left-right hemispheric lateralization of the amygdala in the mammalian brain may have a critical impact on both acute and chronic pain. Evolution through natural selection appears to be a major driver of asymmetrical left-right lateralization within nervous systems. Lateralization is present across diverse taxa, from invertebrates to humans ( Ocklenburg and Gunturkun, 2017 ; Rogers et al. , 2013 ). It is likely that nearly all organisms with bilateral body plans (phylum Bilateria ) carry the potential for asymmetry in the nervous system. Evidence for left/right side lateralization has been found in primates, zebrafish ( Rogers and Vallortigara, 2017 ), nematodes ( Rogers and Vallortigara, 2017 ), insects ( Rogers and Vallortigara, 2017 ) and songbirds ( Rogers et al. , 2013 ). This lateralization comes in the form of anatomical lateralization ( Ocklenburg and Gunturkun, 2017 ), gene expression lateralization ( Varlet and Robertson, 1997 ), and finally functional lateralization. Often these different “types” of lateralization are intertwined. That is, changes in gene expression on one side of the nervous system can lead to changes in local growth (anatomy), and ultimately functional output. Upon this foundation, well-known asymmetric specialization, such as the functional lateralization of language and visual processing in humans, likely evolved over top of existing anatomical asymmetry ( Levy, 1977 ). Although brain lateralization has been known to neuroscience and psychology since 1861, the advent of modern neuroscience techniques, including electrophysiology and functional imaging, has greatly expanded our understanding of lateralization and its impact to the normal and diseased nervous system. Findings from human imaging studies suggest that lateralization within the brain may actually be the norm rather than an exception ( Sergerie et al. , 2008 ; van Elst et al. , 2000 ). That is, a multitude of imaging studies find some evidence of lateralized differences in activation, structure, and/or functional connectivity. These data, in turn, are providing new hypotheses for the pathophysiology of neurological disease. By comparing imaging data from individuals with acute disease to those in the chronic state, neuroscientists are learning about the plasticity that drives chronicity ( Simons et al. , 2014 ). An area where evidence of lateralization is expected to make a large impact is the field of pain. The peripheral nervous system, where noxious stimuli are detected, shows evidence of anatomical lateralization. Pain and temperature sensory information is transmitted directly to the brain via the lateral spinothalamic tract, though various other indirect (spinomedullary and spinobulbar projections) pathways exist ( Wall et al. , 1999 ). The majority of nociceptive information is transmitted from sensory fibers to superficial laminae I and II of the dorsal horn of the spinal cord, but a small percentage synapse into deeper laminae (IV, V, VII, and VIII) as well ( Wall et al. , 1999 ). After local processing of the afferent signal in internal dorsal horn circuits, secondary projection neurons decussate (approximately 90%) and carry the nociceptive signal out of the contralateral side of the spinal cord and terminate in the lateral parabrachial nucleus (PBN), periaqueductal grey (PAG), nucleus of the solitary tract, and various nuclei of the thalamus ( Wall et al. , 1999 ). Thus, nociceptive information sensed on the left side of the body is largely processed in the right hemisphere of the brain and vice versa ( Hudspeth et al. , 2013 ). This anatomical lateralization of pain inputs then provides the theoretical foundation for functional lateralization in the brainstem and brain. Although numerous areas within the brain show preliminary evidence for functional lateralization in the context of pain, no area has shown greater capacity for left and right differences in input and output than the amygdala. The amygdala was first described in the 19 th century, named because of the almond shape of one of its many nuclei ( LeDoux, 2007 ). The amygdala is known for playing a role in emotional processing ( Phelps and LeDoux, 2005 ; Sergerie et al. , 2008 ), fear ( LeDoux, 2007 ), stress ( Adamec et al. , 2005 ; Li et al. , 2017 ; Nation et al. , 2018 ; Rauch et al. , 2000 ), and other affective disorders, such as anxiety and depression ( Myers and Greenwood-Van Meerveld, 2010 ; Narita et al. , 2006 ; Tran and Greenwood-Van Meerveld, 2012 ). The amygdala consists of the lateral, basolateral, central, and medial nuclei ( Janak and Tye, 2015 ). In the context of lateralization, the central nucleus of the amygdala (CeA) is of great interest. The CeA has been implicated in modulating emotional processes and it is known to have a role in fear and threat-conditioning ( Baker and Kim, 2004 ; Haubensak et al. , 2010 ; Li et al. , 2013 ). The CeA itself is largely GABAergic and has different functional subnuclei, including the central medial amygdala (CeM), the lateral central amygdala (CeL) and the capsular central amygdala (CeC) ( Figure 1 ) ( Cassell et al. , 1999 ; Paxinos and Franklin, 2019 ; Veinante et al. , 2013 ). Because scientific nomenclature evolves and changes over time, some of the literature discussed throughout this review uses different names or abbreviations to refer to the part of the CeA investigated. For example, the CeC is a relatively new, independent subdivision of the CeA; previously, the CeC was included as a part of the CeL or referred to as the CeLC. For the purpose of this review, we will utilize the most recent abbreviations for the CeA (CeM, CeL, and CeC) as defined by the fifth edition of Paxinos and Franklin’s mouse brain atlas. Because the CeC receives direct input from the PBN via the spino-parabrachio-amygdaloid pathway, it has been referred to as the “nociceptive amygdala” and has been thought to be responsible for attaching emotional valence to incoming nociceptive information ( Neugebauer et al. , 2004 ). Most studies investigating the amygdala in the context of pain are therefore looking at neurons located in the CeC, even though the nomenclature used in the literature likely varies depending on year of publication. The CeC includes both nociceptive specific ( Ji and Neugebauer, 2009 ) as well as wide-dynamic range neurons ( Ji and Neugebauer, 2009 ) and has both local connections as well as external projections. Local connections include the lateral central amygdala (CeL), the medial central amygdala (CeM) and the basolateral amygdala (BLA) ( Veinante et al. , 2013 ). External projections include the thalamus, the PBN, hypothalamus, the hippocampus and the PAG ( Veinante et al. , 2013 ) ( Figure 1 ). These amygdala circuits have recently been extensively reviewed ( Thompson and Neugebauer, 2018 ; Veinante et al. , 2013 ). Many studies of the amygdala, especially in rodent models, have assumed equal input to and output from the left and right CeA. Typically, researchers have manipulated and recorded from both hemispheres or focused on one side (typically the right) assuming an equivalent effect on the opposite side ( Carrasquillo and Gereau IV, 2008 ; Ji and Neugebauer, 2009 ). Data collected over the last 20 years has suggested that the left and right amygdala may show functional lateralization in the context of pain and nociception. When the left and right amygdala have been distinctly evaluated, the general trend is that the right CeA is pronociceptive while the left has a minimal role in pain ( Carrasquillo and Gereau IV, 2008 ; Carrasquillo and Gereau, 2007 ; Han and Neugebauer, 2004 ; Ji and Neugebauer, 2009 ; Neugebauer and Li, 2003 ). However, recently the left CeA has been found to have an anti-nociceptive effect in the context of bladder pain ( Sadler et al. , 2017b ), making the idea of amygdala lateralization even more compelling. In this review, we will present all of the current evidence for amygdala lateralization. This review represents the first time that these data on amygdala lateralization have been systemically cataloged and reviewed. We will begin by focusing on early evidence from the literature of amygdala lateralization in the fear/threat conditioning field to provide context for more recent data from the pain field. Next, we will present data from unilateral studies (i.e. right or left amygdala only studies) in different pain models as well as the cadre of newer studies that have separately manipulated or measured the left and right amygdala in single studies of pain. Overall, our goal is to present the data to show both evidence for lateralization of the amygdala in pain and to highlight the existing controversies, mysteries, and questions that remain in the field.