Pain Pathways and Nervous System Plasticity: Learning and Memory in Pain

Abstract

Introduction

Chronic pain is one of the primary reasons patients seek health care in the United States, with a significant proportion of physician visits attributable to some type of chronic pain complaint [1]. The International Association for the Study of Pain (IASP) provides a definition of chronic pain that accounts for both the duration and complexity of pain (ie, persistent or recurrent pain lasting longer than three months) [2]. In addition, the IASP, in cooperation with the World Health Organization, has proposed a rational classification for chronic pain based on etiology, underlying pathophysiological mechanisms, and body site [2].

On a community level, the prevalence and cost burden of chronic pain are considerable, and the individual distress caused can also be substantial, with cognitive impairment, depression, and loss of function [3,4]. According to an Institute of Medicine US report, chronic pain affects more than 100 million American adults—more than those affected by heart disease, cancer, and diabetes combined—and costs up to $635 billion each year in medical treatment and lost productivity [5].

Traditionally, chronic pain has been categorized as chronic cancer pain and chronic noncancer pain. However, this distinction is largely philosophical, with little, if any, medical or clinical relevance [6], and it does not help clinicians to better understand the mechanisms underlying pain chronification or provide guidance for appropriate treatment approaches. Understanding the mechanisms that underlie chronic pain and the processes associated with its development can help inform treatment. Thus, the primary goal of this article is to explore the association between learning, memory, and the development and perpetuation of chronic pain. Because the topics covered in this review are complex, this article begins with a brief overview of the transmission of pain in the peripheral and central nervous system (CNS) and the structural and functional changes in pain chronification, which are important for understanding how these processes relate to learning and memory.

Pain Pathways and Mechanisms: From Nociception to Central Processing

Pain pathways represent a complex sensory system, with cognitive, emotional, and behavioral elements having evolved to detect and integrate a protective response to noxious stimuli [7]. In humans, this system involves both primitive spinal reflexes and complex conscious and subconscious supraspinal responses. The intensity and spatial and temporal patterns of noxious stimuli are transduced into a signal through nociceptors and carried by thinly myelinated or unmyelinated nerve fibers (Aδ and C fibers, respectively), which are then conveyed to the CNS and ultimately to higher CNS centers (vide infra) (Figure 1 ) [9–11]. Along with other touch sensory neurons, the cell bodies of primary afferent Aδ and C fibers are located in the dorsal root ganglion [8,12] and synapse onto several cell types in the spinal dorsal horn, releasing excitatory neurotransmitters including glutamate, substance P, calcitonin gene–related peptide (CGRP), and other peptides [9,10]. In addition to synapsing with interneurons involved in local reflexes, primary afferents synapse onto nociceptive-specific neurons involved in signaling the presence and location of pain and inhibitory and excitatory interneurons in the Rexed lamina I and II (substantia gelatinosa) [9,12]. Primary afferent neurons also synapse onto projection neurons located more deeply in the dorsal horn, termed “wide dynamic range” neurons, which signal the intensity of pain, and their response characteristics can be modified by inputs [13–15].

Figure 1

Open in new tabDownload slide

Ascending and descending pain pathways (adapted from Zhuo [8]). ACC = anterior cingulate cortex; DRG = dorsal root ganglion; GABA = gamma aminobutyric acid; Glu = glutamate; HC = hippocampus; IC = insular cortex; NRM = nucleus raphe magnus; PAG = periaqueductal gray; PFC = prefrontal cortex; S1 = primary somatosensory cortex; S2 = secondary somatosensory cortex; SDH = spinal dorsal horn; VPL = ventral posterolateral nucleus of the thalamus.

Projection neurons from the dorsal horn decussate at the ventral commissure, ascend in the lateral spinothalamic tract, and end in the ventral posterolateral nuclei of the thalamus, with relays to the somatosensory cortex and periaqueductal gray matter (PAG) [9,12]. Ascending spinoreticular and spinomesencephalic tracts carry pain and touch information to areas that play a role in the memory and affective components of pain such as the amygdala, hypothalamus, and PAG [9,16].

Supraspinal involvement is fundamental to the experiential aspects of pain. Multiple regions are involved, including the hypothalamus, amygdala, and cortex—specifically, the anterior cingulate cortex, insular cortex, primary cortex, and secondary somatosensory cortex—as well as the nucleus accumbens (NAc) and the PAG (Figure 1) [8,9,12,17–21].

A larger system involves the function of various subsystems, including the descending pain modulatory system encompassing the PAG and rostral ventromedial medulla (RVM), as well as the reward-motivation network, including the NAc, the prefrontal cortex, and the ventral tegmental area [9,18,19]. These descending tracts are monoaminergic, and their activation from the RVM and PAG can release norepinephrine and serotonin into the spinal dorsal horn [22,23].

Glial cells also play a role in the release of neurotransmitters and molecules involved in pain processing (eg, CGRP, substance P, and glutamate) and are critically important in the processing, expression, and transmission of pain, as well as in the development of chronic pain [24,25]. In addition, glia can release inflammatory cytokines known to be involved in pain [26–29]. The multifaceted interaction between glia and neurons depends on glial cell types, the location of the regulatory process (peripheral nerve, spinal cord, or brain), and the type of the pain (eg, acute, chronic) [30,31].

The concept of simple transmission of pain signals in the spinal cord fails to account for the complex interactions that occur in the dorsal horn and elsewhere. Central sensitization of dorsal horn neurons constitutes a form of long-term plasticity in the CNS and often occurs following injury and inflammation. It is characterized by increased excitability of dorsal horn neurons, an increase in spontaneous activity, enlarged receptive field areas, and an increase in responses evoked by large- and small-caliber primary afferent fibers (Figure 2 ) [32–35]. The mechanism underlying central sensitization involves glutamate released by C fibers onto postsynaptic N-methyl-D-aspartate (NMDA) receptors, causing calcium influx [36,37]. The release of glutamate and substance P is mediated through presynaptic neuronal N-type voltage-gated calcium channels [38]. In addition to sensitization of the dorsal horn, calcium flux through NMDA receptors is critical to synaptic plasticity [33]. Indeed, central sensitization involves changes in the spinal cord gray matter, which mimic the supraspinal plasticity that accompanies both learning and memory [39], which is discussed further in the next section.

Figure 2

Open in new tabDownload slide

Central sensitization in the dorsal horn of the spinal cord (adapted from Kuner [32]). Nociceptor afferent may increase the strength of synapses between these fibers and dorsal horn neurons. These changes can result in stronger responses to a given stimulus (sensitization) and expansion of the area on the body surface to which a given dorsal horn cell responds (ie, its receptive field).

“Windup”—which refers to progressive increases in the magnitude of C-fiber-evoked responses during a continuous stimulus in experimental settings—shares some features with central sensitization and provides a theoretical foundation for understanding the role of neural plasticity in the development of chronic pain [40,41]. Although the analysis of the cellular and molecular mechanisms underlying central sensitization and synaptic plasticity that accompanies learning and memory suggests some subtle differences, there are several mechanistic similarities, including similar pharmacology (ie, specific glutamate receptors and molecules involved in calcium signaling) and structural changes, which are further discussed in the next section.

Anatomic and Physiologic Overlap with Pain, Memory, and Learning

Considerable progress has been made toward understanding the neurologic underpinnings of memory, learning, and chronic pain, thereby permitting greater insight into common anatomical systems, and neurochemical substrates.

Memory is defined as explicit (declarative; eg, remembering a list of dates) or implicit (nondeclarative; eg, tying shoelaces) [42]. Implicit memory includes associative and nonassociative learning [43]. Encoding is a critical initial step in memory formation that is initiated by attention to an event (regulated by the thalamus and frontal lobe) causing more frequent neuronal firing [12,42,44,45]. Attention is increased by emotional content and processed on an unconscious level in the amygdala. These sensations are then translated experientially in sensory areas of the cortex before being combined in the hippocampus. The hippocampus and amygdala interact synergistically to form long-term memories of significantly emotional events [46,47], and the amygdala is a key site of synaptic plasticity associated with fear conditioning (a form of associative learning) [48].

Explicit and implicit forms of memory are distinct from pain memory—the neuroplasticity leading to memory-like maladaptive changes in response to painful stimuli [49]. Nevertheless, neuroplasticity is one of the neurochemical bases of learning and memory, as well as the development and persistence of chronic pain (Table 1, Figure 3) [33,50]. This involves long-term potentiation (LTP), which refers to a persistent increase in the efficacy of synaptic transmission when the same neurons are active at the same time and become sensitized to each other [33,45,50]. Windup is an enduring form of synaptic plasticity that shares essential biochemical features with some forms of LTP associated with memory consolidation, including the activation of NMDA receptors [37,50,51,68]. Central sensitization in the spinal cord is characterized by alterations in glutamate transmission through NMDA receptors and alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors, similar to changes underlying LTP in learning and memory [33,48,52,69].

Figure 3

Open in new tabDownload slide

Long-term potentiation (adapted from Lamprecht and LeDoux [50]). As in central sensitization, specific stimulation results in strengthening of synapses that may involve multiple mechanisms. This strengthening results in a greater response of the postsynaptic cell to a given presynaptic input. LTP = long-term potentiation.

Animal studies suggest that while LTP plays a role in encoding memory, long-term depression (LTD; persistent reduction in synaptic strength) can inactivate memory [70]. Similarly, LTD has been observed in nociceptive pathways and is associated with reduced pain sensitivity or hyperalgesia [49,71].

The transience of short-term memory reflects the transitory nature of underlying biochemical changes, whereas long-term memory involves new neuronal processes and synapses arising from alterations in protein synthesis and gene regulation [42,45]. Indeed, synaptogenesis plays a central role in learning and memory [53,72], and this process is important in the development of chronic pain [54]. Long-term potentiation represents post-translational plasticity, which persists when it is followed by gene transcription, protein translation, and protein synthesis, which act to consolidate changes at synapses [73]. In memory consolidation, additional connections and pathways are created, with new experiences leading to rerouting of connections and reorganization [55], reinforced by the synthesis of new proteins to remodel synapses [74]. In this manner, the brain undergoes reorganization in response to experiences, creating new memories that are triggered and shaped by experience. Eventually, memory is stored in other areas of the brain, including the striatum (memories for skills and habits) [75,76], the neocortex (priming), and the cerebellum (simple forms of associative learning) [77,78].

Although neurons have traditionally been the focus of neuroplasticity in memory and learning, glia also play a critical role in regulating synaptic and structural plasticity in areas involved in not only learning and memory, but also pain [30,31,79–81]. Oligodendrocytes also support the production of myelin [82], which is essential to the fast impulse conduction tracts required for memory formation [83,84]. Thus, the involvement of glia represents another functional similarity between the plasticity underlying learning and memory and chronic pain.

Chronic pain and learning and memory share anatomical sites of synaptic plasticity or functional changes (Table 1). For example, in addition to their roles in learning and memory, the hippocampus and/or amygdala also likely play a role in chronic pain and are subject to pain-induced plasticity. In murine models, chronic pain has been associated with alterations in hippocampus-dependent behavior, and cellular and molecular fluctuations in the hippocampus and/or amygdala—including changes in volume, neurogenesis, and altered synaptic plasticity—have been detected [56–60]. Chronic pain–related changes have been observed in the anterior cingulate cortex and NAc as well [85–87]. Clinical experience suggests an anatomical overlap between chronic pain and learning and memory [61,88]. It has been recognized for some time that pain is exacerbated by anxiety, and individuals differ in their sensitivity to pain expectancy. Specifically, activity in the hippocampal network is associated with exacerbation of pain by anxiety, whereas individual sensitivity to pain expectancy is related to differential activation of the hippocampus and amygdala [61,88].

Another mechanistic similarity between chronic pain and learning and memory is the involvement of neuronal growth factors in synaptic plasticity. For example, the persistent synaptic plasticity underlying the formation of object–place and spatial reference memory in mice requires brain-derived neurotrophic factor (BDNF) [89]. Similarly, spinal BDNF signaling contributes to the synaptic plasticity underlying neuropathic pain in a rat model of nerve injury, and supraspinal BDNF facilitates pain in a model of persistent inflammatory pain [90].

On a neurochemical level, the presence of serotonin markers in the brain areas involved in memory and learning is well documented [91–96]. Serotonin is projected widely throughout the brain [97], and a large body of evidence indicates the interplay between serotonergic transmission and other neurotransmitters (eg, dopamine, glutamate, and gamma-amino butyric acid [GABA]) in the neurobiological control of learning and memory [62,63,98]. Dysregulation of serotonin signaling also plays an important role in chronic pain [9,62].

Mechanistically, serotonin is involved in short-term memory and learning by modulating neuronal membrane channels, including those that directly regulate neurotransmitter release (ie, Ca²⁺ channels in the presynaptic neuron), neuronal excitability, and synaptic responses in postsynaptic neurons [39,45]. Specifically, serotonin binds to neuron receptors coupled to either the diacylglycerol-protein kinase C system or the cyclic adenosine monophosphate-protein kinase A system [39,45]. Through phosphorylation, these protein kinases regulate the properties of membrane channels and other processes involved in neurotransmitter release. Serotonin also leads to changes in the properties of the postsynaptic motor neuron, including an increase in the number of glutamate receptors [9,62].

Overall, there is mounting evidence of anatomic and physiologic overlap between pain and memory on several levels. Anatomically, preclinical and clinical evidence indicates that several neuroanatomical substrates are common in these processes. The hippocampus and amygdala and anterior cortex are central to the encoding and consolidation of memory, including the emotional content, and these structures are implicated in the experiential aspects of pain. Mechanistic overlap between pain and memory is plausible because several neurotransmitters and the process of LTP are common to both. These anatomic and mechanistic correlates strengthen the hypothesis that these phenomena do not exist in isolation, but instead intimately interact on several levels.

Chronic Pain, Memory, and Learning

Evidence from Neuroimaging Studies

Clinical studies using neuroimaging techniques, including magnetic resonance imaging (MRI), functional MRI (fMRI), and magnetic resonance spectroscopy (MRS), have greatly advanced our understanding of the interaction between pain and memory/learning on neuroanatomical and functional levels. More than a decade ago, neuroimaging demonstrated that cortical gray matter density decreased regionally in chronic back pain [99]. In this early report, patients with chronic back pain had 5% to 11% less neocortical gray matter volume than control subjects—estimated to be equivalent to the gray matter volume lost in 10 to 20 years of normal aging [100,101].

Since this initial finding, mounting clinical evidence has documented brain morphological changes in various chronic pain conditions (eg, back pain, fibromyalgia, complex regional pain syndrome, knee osteoarthritis, irritable bowel syndrome, headache) [102–112]. Using voxel-based morphometry (VBM) and other techniques, conspicuous and varied patterns of altered brain morphology have been observed in a number of chronic pain conditions. For example, gray matter reductions in the amygdala and several cortical areas were observed in patients with chronic migraine vs healthy controls [108], and increases were seen in the striatum and orbitofrontal cortex in patients with fibromyalgia [111]. Of interest, a recent MRI study found significantly less bilateral hippocampal volume compared with controls in patients with some chronic pain conditions [59]. These findings are consistent with the notion that the regions involved in memory, particularly the hippocampus and amygdala, are altered morphologically in patients with chronic pain.

In general, cross-sectional fMRI studies indicate preferential involvement of the limbic and emotional systems in encoding pain when the pain is chronic [64, 113–121]. For example, in patients with trigeminal neuralgia, a recent structural (VBM) and functional (MRS) neuroimaging study showed reductions in gray matter volume compared with controls [119]. This was seen in the thalamus, putamen, primary somatosensory cortex, anterior insular cortex, and NAc. A coincident increase in gray matter volume in the posterior insula was also observed. In the same study, MRS revealed reductions in N-acetylaspartate/creatine ratio (NAA/Cr)—a biochemical marker of neural viability [122]—in the region of thalamic volume loss. The NAA/Cr changes correlated with pain duration and intensity. Results from other studies using various neuroimaging approaches further support the structural and functional plasticity in limbic and cortical sites that accompanies chronic pain [123, 124].

In addition to morphological and functional alterations, chronic pain also appears to modify brain dynamics by altering brain resting state interactions between networks implicated in attention, salience, and reward [125–129]. Together, findings such as these support a mechanistic model for pain chronification in which nociceptive inputs contribute to structural and functional plasticity in the limbic–cortical circuitry. Learning mechanisms within the limbic circuitry may contribute to the transition from acute to chronic pain.

Within the context of this model, two recent longitudinal brain imaging studies have explored the reorganization of brain activity that accompanies and perhaps mediates the transition to chronicity [64,65]. In one study, patients with subacute back pain (SBP), defined as pain lasting four to 16 weeks, were assessed over one year by T1-weighted MRI, VBM, and fMRI as patients transitioned to persistent pain or recovery, thereby allowing comparisons with healthy and chronic back pain subjects [65]. Persistent or recovering patients with SBP were subdivided based on a self-reported 20% change in pain intensity from baseline to one year. Notably, over the study period, patients with persistent SBP exhibited regional brain morphologic changes including decreased gray matter density in the insular cortex, primary somatosensory and motor cortex, and NAc, a key part of the mesolimbic circuitry underlying reinforcement learning. In contrast, across all measures and time points, patients who recovered from SBP resembled healthy controls [65]. Thus, in this study, gray matter density decreased in patients with persistent SBP, but not in those who recovered from SBP.

Related alterations in functional connectivity were also observed in those with persistent pain, which collectively correlated with back pain intensity in the second study. Within the one-year study period, back pain–related brain activity in patients with persistent SBP transitioned from the insular cortex, anterior cingulate cortex, thalamus, and basal ganglia to the medial prefrontal cortex, amygdala, and basal ganglia [64]. Although the transition was slow, it was preceded by functional connectivity differences detectable as soon as the first brain scan [64]. Moreover, the baseline medial prefrontal cortex–NAc functional connectivity strength predicted the development of chronic back pain with 81% accuracy [65], suggesting that enhanced mesolimbic circuitry drives brain reorganization [130].

This temporal profile of brain reorganization during pain chronification and the assumption that early vs chronic stages of back pain are preferentially associated with acute pain or emotion circuitry, respectively, were subsequently tested by the same investigators. Patients with early, acute/SBP (defined as pain lasting approximately two months with no prior history of back pain for one year) were compared with the patients with chronic back pain (defined as pain lasting >10 years). Persistent or recovering SBP patients were defined as in the previous study. Investigators found that brain activity for back pain in patients in the combined early, acute/subacute group was limited to the regions involved in acute pain, whereas in the chronic back pain group, activity was confined to emotion-related circuitry. Among the subset followed longitudinally for one year, patients with recovering and persistent SBP showed activation within acute pain regions at the first two follow-up visits encompassing the bilateral insula, thalamus, and anterior cingulate cortex. However, patients recovering from SBP showed no significant activity for the final two visits, whereas patients with persistent SBP exhibited increased activation in the medial prefrontal cortex and amygdala at the final visit (Figure 4 ) [64]. This suggests that spatiotemporal reorganization of brain activity accompanies the transition to chronic pain, during which the representation of back pain gradually shifts from sensory regions toward emotional and limbic structures.

Figure 4

Open in new tabDownload slide

Longitudinal changes in brain activity underlying spontaneous pain when patients transition from an acute to a chronic back pain state (used with permission from Hashmi et al. [64]). At visits 1 and 2, recovering and persistent subacute back pain (SBP) groups show activation within acute pain regions including the bilateral insula, thalamus, and anterior cingulate cortex. At visits 3 and 4, patients recovering from SBP show no significant activity, whereas increased activation in the medial prefrontal cortex and amygdala is observed in patients with persistent SBP at visit 4.

Limbic and Cortical Involvement

The interaction of limbic–cortical circuits identified by neuroimaging studies is consistent with the concept of close interaction between pain, memory, and learning. In this model, chronic pain is pain that fails to extinguish its memory trace, and instead reflects a state of continuous learning in which the interaction between the prefrontal cortex and limbic learning circuitry is central to the transition to chronicity (Figure 5 ) [130].

Figure 5

Open in new tabDownload slide

A model of the brain circuitry involved in the transition from acute to chronic pain (used with permission from Mansour et al. [130]). ACC, anterior cingulate cortex; Amyg = amygdala; Hippo = hippocampus; lPFC = lateral prefrontal cortex; mPFC = medial prefrontal cortex; NAc = nucleus accumbens.

Transient nociceptive signals mainly evoke acute pain perception through activation of the anterior cingulate cortex and insular cortex. Mansour et al. [130] proposed a model in which limbic–cortical plasticity underlies the shift to chronic pain. In this model, normally any learned associations mediated by the limbic circuitry would be gradually extinguished or unlearned with time [130]. However, the limbic circuitry could be preferentially activated (including the amygdala, hippocampus, and NAc) if the nociceptive signal was persistent and/or intense. These structures are integral to learning and memory. In turn, these pathways shift cortical activity from a predominately nociceptive state to a more emotional one by interacting with the prefrontal cortical circuitry, with pain transitioning to a more emotional state [130]. The limbic circuitry also provides modulatory signals to the cortex, causing functional and anatomical alterations. According to this hypothesis, the persistence of pain-related perceptions is driven by implicit (nondeclarative) and explicit (declarative) memories interacting with subconscious signals. Thus, chronic pain is either “unlearned” or maintained depending on the reaction of the mesolimbic emotional learning circuitry.

Interactions also occur at the spinal level. That is, the spinal cord also exhibits plasticity, and nociceptive inputs that are altered by peripheral and spinal cord sensitization processes impinge on limbic circuitry. Conversely, the altered learning processes in the limbic–cortical circuitry impact descending modulatory pathways, thereby influencing spinal cord responses to nociceptive inputs [130]. Thus, the chronification of pain involves interactions at multiple levels and offers several possible therapeutic targets. Increased understanding of the neurophysiology of chronic pain has provided insight into potential prophylactic and treatment strategies that prevent the establishment of chronic pain pathways or reverse the reorganization that accompanies chronic pain including regions involved in memory and learning.

Prevention and Treatment of Chronic Pain

Ideally, chronic pain should be prevented early by pharmacological and/or psychological interventions to prevent the establishment of the pain-induced plasticity in the nervous system analogous to memory mechanisms. Acute pain, when poorly treated, may lead to chronic pain syndromes; therefore, preemptive analgesia represents a possible approach to preventing chronic pain by blocking peripheral and central nociception [131]. Some evidence favors pharmacologic agents that interfere with molecules involved in CNS plasticity (eg, perioperative nerve block) as effective preemptive analgesics [132–136]; however, controversy remains regarding this approach, and the body of literature focused on preemptive analgesia is limited.

Regarding management of established chronic pain, several clinical neuroimaging studies have shown that effective interventions can restore, at least partially, normal brain structure and function [137, 138]. The reversal of chronic pain may be possible with pharmacologic intervention that prevents or reverses the reorganization that accompanies chronic pain, and a very wide range of targets has been and continues to be evaluated. These targets include specific ion channels, neurotransmitters, and receptors associated with pain signaling pathways [136–141]. In addition, trophic factors that may contribute to changes in the properties of individual neurons and the sprouting of their axons are under investigation [69].

Pharmacologic agents more likely to be encountered in the primary care setting include gabapentin and pregabalin, which bind selectively to α₂δ subunits of presynaptic voltage-gated calcium channels and act to inhibit intense neuronal activity by decreasing the calcium-mediated release of glutamate and other neurotransmitters [142–146]. The α₂δ calcium channel subunit is thought to mediate the hyperexcitability underlying neuropathic pain [54, 66]. These findings are of interest given the established role of voltage-gated calcium channels in synaptic plasticity in the spinal cord [147] and amygdala [51]. Systematic reviews provide reasonably good evidence for the efficacy of gabapentin and/or pregabalin in chronic pain conditions, including diabetic neuropathy, post-herpetic neuralgia, central neuropathic pain, and fibromyalgia [148, 149]. In addition, mirogabalin, a novel preferentially selective α₂δ ligand in development, was well tolerated and demonstrated efficacy for diabetic peripheral neuropathic pain [150, 151].

NMDA receptor antagonists, such as dextromethorphan and ketamine, may have some preemptive analgesic benefit [152]. Ketamine has seen increasing use as an analgesic for acute pain [152], and the literature supports its effectiveness in several forms of chronic pain [153]. In addition to its analgesic properties, likely due to NMDA receptor antagonism, ketamine has rapid-acting antidepressant effects, possibly mediated through an increase in AMPA receptor signaling [97]. Low-dose ketamine infusion has successfully treated refractory neuropathic pain syndromes, which may involve significant depressive comorbidities, most notably complex regional pain syndrome [154, 155]. Interestingly, chronic administration of ketamine appears to lead to memory impairments due to reduced hippocampal activation [156].

Currently, two classes of antidepressants, serotonin and norepinephrine reuptake inhibitors (SNRIs) and tricyclic antidepressants (TCAs), are recommended for the treatment of a variety of chronic pain conditions [157, 158]. SNRIs and TCAs block monoamine uptake, thereby increasing extracellular levels of monoamine neurotransmitters [35]. For SNRIs, activation of the descending modulatory pathway from the RVM and PAG to the dorsal horn neurons and altered norepinephrine signaling and synaptic transmission in the brain are all possible mechanisms [23, 97, 159]. Of note, the antidepressant effect of SNRIs may arise from their effect on synaptic plasticity in the cortex, hippocampus, and striatum, all regions implicated in memory and learning. SNRIs have been successful in treating a range of chronic pain conditions, including diabetic neuropathy, fibromyalgia, and chronic back pain [160–162]. The pharmacologic mechanisms for SNRIs and TCAs are similar [97], and over the last two decades, TCAs have been used to treat a number of chronic pain conditions, most notably neuropathic pain conditions [163, 164].

Psychological approaches that focus on factors that exacerbate or maintain chronic pain, such as maladaptive learned brain responses and failure to extinguish (unlearn) pain responses, are possible with operant behavioral therapy, cognitive behavioral therapy (CBT), and motivational interviewing [165, 166]. Several functional neuroimaging studies have demonstrated that CBT can alter the brain network connectivity that occurs with repeated pain experience [167, 168]. In one randomized fMRI study of patients with chronic pain, functional connectivity between the anterior Default Mode Network and the amygdala/periaqueductal gray region decreased following CBT compared with controls not receiving CBT [168]. Another investigation indicated that CBT in patients with chronic pain resulted in increased gray matter volume in areas that also play a role in learning and memory, including the prefrontal and somatosensory brain regions and hippocampus [169]. Furthermore, increased dorsolateral prefrontal volume was associated with reduced pain catastrophizing [169]. These results add to mounting evidence that CBT can be a valuable treatment option for chronic pain. In addition, techniques such as imagery, mirror training, or the use of virtual reality may help address the maladaptive brain plasticity that accompanies chronic pain [170–173]. Integrative approaches to chronic pain management that consider the patient’s biopsychosocial history include physical activity and exercise, nutrition and supplementation, and yoga [174].

Nonpharmacologic interventions for pain can also play a role in memory. For example, vagal nerve stimulation relieves pain in patients with migraine [175] and some patients with fibromyalgia [176], and can also enhance recognition memory [177]. Acupuncture improves osteoarthritis pain and improves working memory [178, 179].

Exercise (eg, resistance or flexibility exercise, yoga, and tai chi) relieves pain in some patients [180] and also improves memory [181]. The effect of exercise, in particular, on chronic pain is an active area of research. Exercise is thought to exert an analgesic effect through the activation of endogenous systems, particularly the opioidergic, nitrergic, serotoninergic, noradrenergic, and endocannabinoidergic systems, as well as through the release of anti-inflammatory cytokines [182–184]. Recent studies have unraveled the potential of exercise to not only relieve but prevent chronic pain via immune system modulation [185, 186]. Six weeks of voluntary wheel running prevented the full development of allodynia due to chronic constriction injury in rats and caused an increase in levels of the anti-inflammatory cytokine interleukin (IL)-10 compared with sedentary controls [186]. In mice, eight weeks of voluntary wheel running caused an increase in the proportion of macrophages that secrete anti-inflammatory cytokines compared with inflammatory macrophages in the gastrocnemius muscle. Physically active mice were also resistant to acid injection–induced hyperalgesia, and this analgesic effect could be prevented by blockade of IL-10 [185].

Clinical studies show promise for exercise in alleviating chronic pain due to osteoarthritis of the hip and knee [187–189], fibromyalgia [190], intermittent claudication [191], pain of the low back [192–195], neck [196, 197], patella, and femur [198], and pain associated with type 2 diabetes [199]. However, there is some conflicting evidence suggesting that physical activity has no significant effect on pain of the low back [200] or pain due to fibromyalgia [201]. Overall, increasing the sample size of clinical studies could allow stronger conclusions to be drawn about the benefit of exercise for chronic pain patients [180]. It will also be important to study which type of exercise has a positive effect for the specific type of pain condition.

Overall, nonpharmacological interventions can be easily integrated as part of a multidisciplinary, multimodal approach that combines pharmacological, physical, and psychological components tailored to each patient’s needs. Specific noninvasive brain stimulation techniques that induce electrical stimulation to directly alter brain activity, such as repetitive transcranial magnetic stimulation, may have some short-term benefit in chronic pain, although additional evidence is required [170]. More studies are needed to investigate the underlying physiological changes in learning, memory, or pain-related areas of the CNS that accompany the analgesic effects of nonpharmacologic approaches.

Conclusions

In sum, chronic pain can result from a persistent pain stimulus due to injury or disease, but it can also persist after the original injury is healed. Nonetheless, chronic pain is accompanied by the persistence of pain plasticity mechanisms analogous to memory and/or the failure to terminate pain plasticity induced by the original inciting injury. This conceptual framework is consistent with neuroimaging evidence of structural and functional neuroplasticity associated with chronic pain, as well as the large body of clinical and experimental evidence demonstrating anatomical and physiologic overlap with learning and memory. In this context, chronic pain can be conceptualized as the consequence of plastic changes in limbic–cortical circuitry, leading to new learning and reinforced maladaptive plasticity analogous to memory mechanisms that cannot be extinguished due to emotional associations with painful stimuli. Further studies are needed to fully understand the key factors influencing the shift from acute to chronic pain and the changes in learning, memory, or pain-related areas of the CNS that accompany analgesic effects of nonpharmacologic approaches. A better understanding of these mechanisms will aid in the development of novel therapeutic approaches that reverse the pain plasticity and relieve chronic pain.

Pharmacological and psychological approaches that rationally target the processes and mechanisms underlying chronic pain are currently available to clinicians. Additional targets and future approaches to treatment might include novel agents that interfere with receptors and molecules involved in LTP in both subcortical and cortical targets (eg, NMDA receptor, voltage-gated calcium channels and their components, BDNF), as well as nonpharmacological approaches that focus on intensive intervention to prevent consolidation of acute into chronic pain and extinguishing memory-like plasticity associated with chronic pain.

Authors’ Contributions

The authors wrote and revised this manuscript and approved the final draft.

Acknowledgments

Editorial support was provided by Claire Daniele, PhD, and was funded by Daiichi Sankyo, Inc.

References