Chronic pain has numerous definitions, but is generally characterized as pain that persists for longer than 3 months. This is a highly prevalent condition that is associated with significant disability and distress in those who are affected. In Canada, approximately 19% of the population report this painful condition.1 Although chronic pain may be reported in the neck area, in various muscles or as severe headaches, chronic low back pain (CLBP) is one of the more prevalent forms of persistent pain. Besides imposing a large social burden and negatively impacting the quality of life of sufferers, chronic pain imposes an enormous economic burden. The cost of chronic pain has been estimated to be approximately US$600 billion annually.2
Advancements in neuroimaging techniques have become increasingly useful in enhancing our understanding of morphologic and functional brain alterations that accompany painful experiences and those that are observed in chronic pain diseases.3–8 Although numerous imaging modalities have been used to gain insight into pain mechanisms, including positron emission tomography and magnetic resonance imaging (MRI), the MRI methods are favored as they do not use ionizing radiation. In addition, MRI techniques are noninvasive since the administration of a contrast agent is not required in most of the imaging protocols. These advantages, along with the widespread availability of MRI scanners have lead many researchers to experiment with the use of numerous MRI approaches to assess brain manifestations of chronic pain. MRI methods have shown great promise in both identifying neurological pain signatures as well as tracking treatment effects.9–13 Most recently, advanced MRI approaches have been shown to be able to predict the transition from acute to chronic pain.14–17
Although recent research has lead to progress in our understanding of pain mechanisms and our ability to identify objective diagnostic imaging measures associated with chronic pain, translation of our current knowledge into clinical practice is still lacking.18 Chronic pain is generally under-treated and under-managed in the current health care environment in Canada. The history and physical examination, as well as the imaging modalities that are available clinically, do not provide the clinician with a prognostic or predictive quotient to determine which patients will develop chronic pain. In addition, our experience suggests that most clinicians are not aware of the central mechanisms and cortical alterations evident in patients with chronic pain nor are they aware of the capabilities of MRI, beyond routine anatomic scans, in assessing cortical manifestations of chronic pain. This scoping review thus aims at identifying studies that have set out to examine cerebral signatures of chronic pain using objective MRI metrics. We were interested in addressing the following research questions: (1) what MRI neuroimaging techniques are available in the current literature to examine chronic pain? (2) which of these MRI methods have been able to reliably predict which patient with acute pain may develop chronic pain? and (3) for those who have chronic pain, which MRI techniques have been successfully used to track treatment response?
We searched 4 electronic databases: Embase, Medline, PubMed, and Web of Science. Our inclusion criteria were studies published between the year 2000 and May of 2014 that had been conducted on human participants. We chose the start date of 2000 because neuroimaging studies using MRI to assess chronic pain started to surface around this time in the literature. The keywords “chronic pain” and “neuroimaging” and “MRI” were used in the search engines to locate appropriate publications. We were aware of the fact that the term “chronic pain” was general and included pain affecting many bodily regions. However, to get a sense of the breadth of the available studies we decided not to limit our search to a specific body location. We also knew that “MRI” was a large definition, but we wanted to encompass all MRI techniques that were being used currently in the literature to examine cortical manifestations of chronic pain. Using these criteria, the Embase database produced 59 hits, Medline produced 18 publications, PubMed produced 278 publications, and Web of Science produced 61 publications.
Article Selection and Data Charting
The abstracts of the original 416 articles that were identified by the search engines were read to determine relevance to our study questions. Duplicate papers were evident due to the use of multiple databases and only a single copy was included in the final compiled list. We only included studies that were written in English and were original peer-reviewed articles. Some of the publications identified through the database searches were irrelevant, so such studies were not retained. Anecdotal, review, speculative, and editorial articles were excluded, as well as studies that did not explicitly report MRI or those that were carried out on animals. In addition, case studies and studies that reported a small sample size (<4) were excluded. The reference list of relevant publications was hand searched at times to locate studies that had been missed through the database searches. Following these measures, our search strategy generated 78 publications that addressed the research questions and satisfied our inclusion and exclusion criteria. Figure 1 shows a flow chart of our search strategy as well as the number of articles that were obtained.
The full articles of these abstracts were then retained so that they can be included in the review. Each publication was read and the following information was gathered:
* Publication identifiers: publication date, names of the first 3 authors, journal, and volume of publication.
* Chronic pain study population: chronic pain location/pain type.
* Study purpose: identify brain alterations in chronic pain, assess brain response following treatment of chronic pain, or predict transition from acute to chronic pain.
* MRI technique used to investigate chronic pain: structural, chemical, functional, or blood flow imaging method.
* Study results: Cortical regions/circuits identified in chronic pain, or cortical response to chronic pain treatment.
Our search revealed that MRI neuroimaging is being used to examine numerous chronic pain conditions. Table 1 summarizes the literature studies with regards to the pain condition that they address. Most of the published articles that our literature search identified described chronic back pain (CBP). There were also many studies examining fibromyalgia (FM) syndrome. Several articles covered temporomandibular disorder (TMD) and neuropathic pain conditions. A handful of studies addressed irritable bowel syndrome (IBS), complex regional pain syndrome (CRPS), pelvic pain, chronic headaches/migraines, and osteoarthritis (OA). A limited number of articles were found describing multisomatoform disorder, Crohn’s disease, burning mouth syndrome, and unilateral pain from herpes simplex virus.
Most of the identified studies in our literature search were set out to examine brain alterations associated with chronic pain. Several studies used MRI to track brain response following chronic pain treatment4,8–13,30,42,48,65 and a handful of more recent articles used MRI metrics to predict transition from acute to chronic pain.6,15,16,29 The findings of our literature search are presented below with regards to each of these 3 broad areas.
Studies Examining Brain Alterations Associated With Chronic Pain
Literature studies that assess neurological signatures of chronic pain using MRI investigated one (or more) of the following brain alterations: structural changes, changes in brain chemistry, functional changes or cerebral blood flow changes. Most of our identified articles examined structural and/or functional alterations present in the brain of chronic pain patients, with only a handful of studies assessing chemical27,28,41,58 or blood flow changes33,57 associated with chronic pain.
Structural Brain Alterations
For many years, anatomic MRI images have been used to better understand brain reorganization associated with chronic pain conditions. Early MRI studies focused on assessing brain changes using a macroscopic approach, examining brain morphology by quantifying either cortical thickness or measuring subcortical brain volumes in patients with chronic pain and comparing the measurements to those of healthy participants. Voxel-based morphometry (VBM) has been the technique used to quantify regional gray/white matter concentrations on a voxel by voxel basis.36,81,82 More recently, with the advances in MR technology, investigations of brain alterations associated with chronic pain have been carried out on a microscopic level using diffusion tensor imaging (DTI) to probe brain tissue microstructure noninvasively.
VBM to Examine Structural Changes
Our literature review identified many studies that have set out to examine gray matter density changes in chronic pain conditions. The distribution of gray matter density has been noted to differ between healthy participants and patients with CBP.3,25,31,35,36 Observations of brain atrophy in individuals with CBP have been reported3 with reductions in the neocortical gray matter volume ranging from 5% to 11% in patients compared with normal controls.3 Regional analysis has further revealed significant reductions in gray matter density in numerous brain regions associated with pain modulation3,25,31 including the dorsolateral prefrontal cortex (DLPFC) and the thalamus.3,31 Reductions in white matter volumes have also been noted in person with CLBP in the cingulate cortex of the left hemisphere.25 Observations of reduced gray matter density in numerous brain areas in CRPS patients have been made66,68 with an increase in gray matter density in the hypothalamus and left dorsal putamen.66 OA patients also show decreases in gray matter volume in the thalamus10 and the insula and mid anterior cingulate cortex (ACC).23 While a common finding of most studies is a reduction of gray matter density in chronic pain, it has been noted that CRPS, OA, and back pain conditions show gray matter decreases in distinct brain regions in addition to a decrease in the insular cortex common to the 3 pain conditions.23
Investigation of structural alterations in the central nervous system of FM patients has revealed that like many other chronic pain conditions, they show regional changes in gray matter density when compared with healthy controls.37,39,43,49 Although regional reductions in gray matter density have been observed in some studies,37,39,43,49 gray matter increases have been noted in the striatum.49 One study further evaluated gray matter alterations with regards to FM patient age and found that older patients (above 50 y old) showed decreases in regional gray matter volumes compared with healthy controls, whereas an increase in regional gray matter is observed in younger FM sufferers.39 Although most studies report a change (increase or decrease) in gray matter density in FM patients relative to healthy participants, 1 study showed that controlling for affective disorders such as depression or anxiety leads to no observed differences in gray matter volume between healthy controls and individuals with FM43 showing the importance of controlling for other factors within this patient group.
Altered brain morphometry has been noted in TMD51,53,54,58 with reductions in regional gray matter density in areas that are part of the central pain system such as the anterior cingulate gyrus and insular cortex.51 Increases in gray matter density have also been observed in the brainstem of TMD patients who have myofascial pain.54 One study showed no differences in gray matter volume between TMD and healthy controls58 while noting gray matter reductions in neuropathic pain compared with TMD pain.58
Numerous other chronic pain conditions report reductions in gray matter density, including trigeminal neurologia with gray matter decreases in the ACC, thalamus, and DLPFC7 as well as patients with recurring herpes simplex virus infections.78 In chronic pain conditions affecting the pelvis, decreases in gray matter volume have been observed in the thalamus of women reporting chronic pelvic pain,69 whereas another study reported increases in gray matter density in numerous brain regions in women with from chronic vulvar pain.71 With regards to chronic prostatitis/chronic pelvic pain syndrome, no regional gray matter changes have been noted between patients and healthy controls.70 In a number of abdominal disorders associated with chronic pain, including Crohn’s disease80 and IBS,60,64 alterations in brain structure have also been observed. Specifically, reductions in gray matter in the frontal gyrus have been noted in Crohn’s disease.80 In IBS, increases have been identified in the hypothalamus in 1 study,60 whereas another study found reductions of gray matter density in numerous regions.64 A number of groups have also examined morphologic changes that accompany chronic headaches, with results showing decreases in gray matter in chronic tension type headache72 and in chronic migraines74 and no changes in gray matter in medication overuse headaches.72
In summary, VBM studies of various chronic pain conditions have revealed that the distribution of gray matter density differs between healthy participants and patients with chronic pain conditions. Those that are affected by CLBP,3,25,31,35 CRPS,66,68 OA,10,23 FM,37,39,43,49 TMD,51,53,54,58 and trigeminal neurologia7 tend to have a decrease in the gray matter density in various brain regions, reflecting brain atrophy and structural alterations in the central nervous system of individuals with chronic pain.
DTI has been recently used to gain further insight into the microstructural organization of brain tissue of persons with chronic pain. The technique allows for the evaluation of the microstructural integrity of a tissue of interest by computing numerous metrics to describe the self-diffusion of water molecules. The motion of water molecules tends to be isotropic in the absence of hydrophobic barriers such as cell membranes or myelin. However, the presence of subcellular structures or membranes results in anisotropic (ie, restrictive) diffusion, where preferential motion is observed such that the water molecules move more easily along the length of an axon. This gives insight into brain tissue microstructure. DTI allows for the quantification of diffusivity using eigenvalues to characterize the diffusion magnitude along 3 orthogonal directions (λ1, λ2, λ3). From these indices, the mean diffusivity, also known as the apparent diffusion coefficient, can be calculated. The apparent diffusion coefficient measures the arithmetic mean of water diffusivity along the 3 main axes. Fractional anisotropy (FA) is another metric that describes the shape of the diffusion ellipsoid. This scalar quantity ranges between zero (perfect isotropic diffusion) and one (diffusion along an infinite cylinder). Myelin typically has a FA value around 0.5 to 0.7. DTI also allows for the qualitative visualization of white matter tracts using tractography algorithms.
Our literature search identified numerous papers that used DTI to examine white matter integrity in the brain of chronic pain sufferers.6,39,57,59,61,68,70,83 A study examining CRPS patients found decreased FA in the left callosal bundle of patients compared with healthy controls,68 which was suggested to reflect injury to the white matter bundle.68 Further analysis revealed that lower long-range connectivity from this region was evident in CRPS patients and short-range branchings were less space filling when compared with healthy controls.68 Another study examining patients with disabling CLBP also found decreased FA in the patient population in the splenium of the corpus callosum83 further reflecting deficits of white matter integrity in chronic conditions. Examining painful trigeminal neuropathy has also revealed reduction in FA in the primary somatosensory cortex region that represents the lower face, further supporting the notion of altered tissue microstructure with chronic pain.57 Significantly, lower FA was identified in trigeminal nerves as well as clusters of white matter throughout the brain of TMD patients, including the corpus callosum, compared with healthy controls in 1 study.6 Connectivity analysis further revealed sparse connections between the DLPFC and the corpus callosum in TMD patients compared with controls.6 Evidence of microstructural changes, within numerous brain regions, has also been noted in IBS patients where reductions in FA was noted in the thalamus, globus pallidus, and sensory/motor association regions.61 Increases in FA have also been observed in the corpus callosum and prefrontal white matter in the brain of IBS patients compared with controls.61 Most of the studies that have used DTI to examine microstructural changes in the brain of chronic pain sufferers have noted a reduction in the FA in various regions,6,61,68,83 reflecting changes in tissue microvascular structure.
Changes in Brain Chemistry
Magnetic resonance spectroscopy (MRS) is an in vivo MRI method that can be used to gain insight into brain chemistry. The technique allows for regional metabolite concentrations to be determined noninvasively, possibly enabling early detection of disease status before anatomic manifestations become evident. Numerous groups have used proton (1H) MRS to carry out regional biochemical assessments in the brain of chronic pain populations.27,28,41,58 Our literature search identified studies that have been performed in CBP patients,27,28 individuals with FM,41 and those with trigeminal neuropathy or TMD.58 A number of metabolites can be examined using 1H-MRS, including N-acetylaspartate (NAA), choline, creatine (Cr), glucose, myo-inositol, glutamate and glutamine (Glx), and γ-aminobutyric acid. Altered concentrations of chemicals have been identified in brain regions of back pain patients in the DLPFC,27 the ACC, thalamus, and anterior insula (AI).28 Specifically, examination of metabolites from the DLPFC region revealed a 7.8% decrease in NAA and 17.2% reduction of glucose in CBP patients compared with normal controls.27 In addition, NAA reductions were noted in the ACC and AI regions of patients, along with decreases in glutamate and myo-inositol in the ACC region.28 In another study, reductions in the ratio of NAA/Cr were noted in the thalamus of trigeminal neuropathy patients but were not present in the thalamus of those with TMD.58 Assessments of the ventrolateral prefrontal cortex of FM patients also revealed cortical alterations, with elevated ratios of Glx/Cr and Glu/Cr in the patients compared with healthy controls and no changes noted in the thalamus region.41 All of these studies reveal that altered metabolite concentrations are evident in various forms of chronic pain, suggesting that chemical markers, such as NAA or Glu can be useful in understanding chronic pain processing and targeting therapies to treat metabolite imbalances.
Functional Brain Alterations
Assessments of brain activity are commonly done using fMRI studies. These may involve the use of a stimulus to evoke a certain brain response and are referred to as task-based or stimulus-based studies. In such setups, the stimulus may be the administration of heat/cold to a given body region (thermal stimulus) or it may be electrical or mechanical in nature. Stimuli are generally presented in blocks, such that the stimulus is applied between “rest” periods that do not involve stimulus application or use nonpainful stimuli. Following data analysis of the fMRI set, such as B0 inhomogeneity correction, motion correction, slice timing correction, and alignment of the data set with a reference anatomic atlas (for group analysis), statistical analysis is carried out to identify brain activations in response to stimulus application. Statistical assessments may be based on a general linear model, where a model of the time course of the stimulus is convolved with the hemodynamic response function and correlations are then carried out between the measured response and that predicted by the model. This type of hypothesis driven analysis using a predefined stimulus response is what is done in the majority of literature studies examining brain activations in chronic pain patients. A less common approach is the use of data driven approaches such as principal component analysis, independent component analysis or partial least squares. These analysis techniques do not require the use of models to predict stimulus response, but rely on exploring underlying data structure to identify stimulus-related activations.
Although early fMRI studies relied on the use of a stimulus to identify brain responses to various experimental tasks, recently researchers have been interested in examining the on-going neural and metabolic brain activities that occur in the resting (basal) state. Resting state fMRI does not involve the use of a stimulus; instead patients are imaged as they lay with their eyes open without thinking of anything in particular. The most common form of this data analysis involves first defining a “seed” region of interest within a particular brain structure. Subsequent statistical analysis (correlational) is then performed to extract brain voxels that exhibit a similar time course to that of the seed, thereby showing resting-state functional connectivity although the voxels are spatially separated. Synchronized activations have lead to the identification of resting state networks (RSNs). The default mode network (DMN) is a set of brain regions that have been shown to be active when people are not thinking of any particular tasks, but become deactivated during the performance of specific tasks. Although a number of other resting states have been identified, the DMN is one of the more prominent networks that has been examined in a number of diseased populations, including chronic pain patients. More recent approaches to perform the analysis of resting state involves group-wise probabilistic independent component analysis, which avoids any input from the analyst in terms of seed placement.33
Functional MRI has been used to understand functional connectivity as well as brain activation in numerous chronic pain conditions.22,24,38–40,44–47,50,52,56,57,62,63,67,70,76–79 Pain-related RSN, such as the anterior default mode network (aDMN) and the fronto-insular network, have been shown to have altered temporal coherence in chronic pain patients compared with healthy controls.77 Specifically, patients show higher resting-state signal fluctuations in the aDMN and fronto-insular network regions in comparison with healthy participants, reflecting altered neural activity with chronic pain.77 Altered resting-state dynamics have been noted within the pain network of FM patients compared with healthy controls as well.40 Abnormalities have also been identified in the DMN in patients with TMD, such that the mPFC of the DMN exhibited enhanced functional connectivity with the retrosplenial cortex, the posterior cingulate cortex (PCC), and the precuneus (PCu) cortex in the patient group relative to healthy controls.52 Pain rumination, which is the tendency to repeatedly have negative thoughts about pain, has also been shown to be positively related to mPFC functional connectivity to numerous regions in the patient group, including the medial thalamus, the PCC/PCu, and the retrosplenial cortex, which are involved in various aspects of pain perception.52 Hypnotically induced hypoalgesia has been shown to modulate the painful experience in patients with TMD, resulting in the suppression of cortical activity in numerous brain regions while only activating the right insula.50 In 1 study, it was suggested that not all forms of chronic pain are associated with remodeling of the primary somatosensory cortex (S1).57 Although S1 reorganization was evident in neuropathic pain patients, as observed in those with painful trigeminal neuropathy, non-neuropathic chronic pain patients with TMD did not display cortical reorganization.57 Attentional networks have been shown to be modified in patients with diabetic neuropathic pain.55 Reductions in the connectivity of the dorsal attentional network, the ventral attentional network, and the dorsal ACC have been observed in patients compared with healthy controls, suggesting a disruption in attention and salience processing systems in chronic pain.55
Burning mouth syndrome patients report more intense pain during the late afternoon compared with the early morning period.79 It is, thus, not surprising that enhanced functional connectivity has been observed between the mPFC and pain processing regions in the brain of BMS patients in the afternoon compared with the morning scan.79 With regards to chronic migraine, atypical resting-state functional connectivity has been identified in the brain of patients between affective pain processing regions and other regions involved in pain experience.73 This altered connectivity was shown to have an interictal component that persisted between migraine attacks.73 In patients with CRPS of the left hand, application of painful electrical stimulus revealed cortical activations that differed from healthy controls.67 Specifically, the patient group showed stronger activation of the PCC as the painful stimulus was applied to the symptomatic hand compared with healthy controls. This observed activation was also not evident when the asymptomatic hand of CRPS patients was stimulated. These results show that chronic pain changes cerebral pain processing.67
Central pain processing has been shown to be augmented in patients with CLBP.26 Application of the same amount of pressure has been shown to lead to increased activation of only the contralateral secondary somatosensory cortical region in healthy controls, whereas the brain of CLBP patients shows activations in the contralateral primary and secondary cortices, the cerebellum, the inferior parietal lobule, and the ipsilateral somatosensory cortex.26 Application of pressure that results in the same amount of pain has been shown to lead to similar activations in healthy participants and patients with CLBP.26 Comparison of brain activation in response to a mechanical back stimulus in healthy participants and in patients with CLBP further revealed that CLBP sufferers have increased activation in some brain regions (PCCs, right insula) compared with healthy controls.32 Acute experimental pain resulting from thermal stimulation applied to the back has also been shown to elicit cerebral activity different from that of spontaneous pain associated with CBP.20 Resting brain activity in individuals with CBP has also been shown to differ from that of healthy controls as reflected by disruptions of RSNs,21,34 including DMN dynamics.21 Patients with CLBP exhibit a shift in the oscillatory fMRI signal, such that there is an increase in the high-frequency oscillations in the patients compared with the control group.19 The brain regions that show such behavior are the mPFC, the PCC, and the right and left lateral parietal cortex, which are parts of the DMN. This study thus further reveals alterations in the frequency of the DMN in chronic pain.19
In summary, fMRI has been useful in examining altered functional connectivity in the brain of chronic pain sufferers.22,24,38–40,44–47,50,52,56,57,62,63,67,70,76–79 For example, the DMN has been shown to display resting-state signal fluctuations in numerous chronic pain conditions that differ from those of healthy participants.52,77 In patients with diabetic neuropathic pain, attentional networks have been shown to be modified,55 whereas stimulation of the symptomatic hand of patients with CRPS lead to cortical activations that differed from healthy controls.67 In CBP sufferers, resting brain activity has been shown to differ from that of healthy controls as reflected by disruptions of RSNs21,34 and chronic migraine shows atypical resting-state functional connectivity between affective pain processing regions and other regions involved in pain experience.73 These studies and others have shown that cerebral pain processing changes in chronic pain.
Blood Flow Brain Alterations
Arterial spin labeling (ASL) is a noninvasive MRI technique that allows for the quantification of regional cerebral blood flow without the use of an injected contrast agent. There are a host of different ASL approaches, each with their own advantages and disadvantages. For an excellent review on the patient we refer the reader to Peterson et al.84 Briefly, the most basic form of ASL relies on the use of radio frequency pulses to magnetically tag water in blood that then enters the imaging region of interest. Comparison of tagged images with ones that have not been tagged enable the quantification of blood flow. A couple of studies were identified in our literature search that used ASL to examine regional cerebral blood flow in patients with CLBP33 and patients suffering from trigeminal neuropathy or TMD.57 In 1 study, ASL was used to evaluate brain connectivity before and after patients and healthy participants were exposed to maneuvers, such as pelvic tilts, to exacerbate the clinical pain level of patients with low back pain.33 The collected ASL data set was analyzed using probabilistic independent component analysis to extract RSNs. As expected, the DMN was identified as one of the networks associated with clinical pain. The connectivity of the DMN was found to be altered in the patient group compared with healthy participants. Following the pain maneuvers, the change in pain was found to be associated with a change in DMN-right insula connectivity.33 This study thus revealed that DMN connectivity may be used to predict clinical pain. In another study, ASL was used to examine neuropathic and non-neuropathic pain, where authors used trigeminal neuropathy to represent the first form of pain and TMD to represent the latter form.57 The study noted that cerebral blood flow in the area representing pain (the primary somatosensory cortex that represents the lips) was reduced in patients with trigeminal neuropathy compared with healthy participants. However, no differences were identified between patients with TMD and healthy controls. Along with other results from functional and anatomic data, this study suggested that not all forms of chronic pain show the same cortical response, with only neuropathic pain showing changes in the somatosensory cortex and thus, this form of pain may require treatment that aims at correcting cortical alterations.57
In summary, the ASL technique, although used in only a couple of chronic pain conditions33,57 has provided useful insight into cerebral blood flow and the associated alterations observed in persons with chronic pain. It has shown that in CLBP, the DMN connectivity is altered relative to that of healthy participants and may thus be used to predict clinical pain.33 In addition, in patients with trigeminal neuropathy, cerebral blood flow in the primary somatosensory cortex is reduced,57 whereas no changes in blood flow have been noted between patients with TMD and healthy participants.57 This reflects the importance of careful assessments of each chronic pain condition, as some conditions may not require therapies aimed at modifying cortical changes.57
Articles Assessing Brain Response Following Chronic Pain Treatment
Numerous studies have set out to examine whether the neuroplastic changes that accompany chronic pain are reversible after cessation of the agent causing the pain.4,8–10,12,13,30,42,48,65 A couple of studies have assessed gray matter volume using VBM in patients with painful hip OA before and after hip arthroplasty.10,12 Although regional decreases were noted in gray matter volume in the patients preoperatively compared with healthy participants, thalamus gray matter volume reductions were found to be reversible 9 months after surgery in 1 study.10 Another study also detected increases in regional gray matter volume in the DLPFC, the ACC, the brainstem, and the amygdala 4 months after total hip replacement surgery, further showing that pain relief leads to reversible anatomic brain changes.12 Cognitive-behavioral therapy (CBT), which is a nonsurgical intervention used to aid patients with chronic pain cope with their symptoms, has been evaluated in 1 study to assess its impact on structural neuroplasticity.8 VBM was used to analyze brain images collected before and 11 weeks after CBT from chronic pain sufferers and the results revealed that the psychological intervention lead to increases in regional gray matter, reflecting an adaptive response to cope with the pain.8
With regards to CBP, application of Lidocaine treatment to the back for a period of 2 weeks has been shown to result in a decrease in brain activity.4 Regions that had been activated before the treatment due to spontaneous back pain, such as the mPFC, the nucleus accumbens (NAc), the bilateral superior frontal gyrus, and the rostral ACC all showed decreased activity following the treatment as tracked by fMRI.4 In a follow-up study using the Lidocaine patch as well as a placebo patch, decreases in brain activity have been observed following the administration of the treatment in both the Lidocaine-treated group and the placebo-treated group.30 This suggested that it was the placebo effect of the patch that lead to its therapeutic effectiveness.30 The use of other interventions to treat CLBP, such as spine surgery and facet joint blocks has also been shown to lead to reversible structural and functional brain changes.13 In addition, acupuncture use to relieve back pain has revealed that the altered DMN connectivity observed before treatment gets restored to levels similar to those of healthy controls.11 This suggested that resting-state fMRI assessments could be used as objective measures to assess the analgesic effects of acupuncture.11
The effects of a number of FM treatments have also been examined using fMRI to assess the success of the treatments.9,42,48 Following a 12-week behavioral extinction treatment, brain activation was noted to shift from the AI to the posterior insula in response to painful mechanical stimulation, possibly reflecting a reduction in the attention to the painful stimulus after behavioral training.9 Administration of analgesics such as pregabalin42 and milnacipran48 to reduce FM symptoms have also been tracked using MRI neuroimaging, further supporting the role of this imaging modality in examining the impact of analgesics in chronic pain situations.
In summary, a number of studies have looked at whether the neuroplastic changes that accompany chronic pain are reversible following cessation of the agent that causes the pain using objective MRI measures.4,8–10,12,13,30,42,48,65 Gray matter volume assessments using VBM have been carried out in patients with painful hip OA before and after hip arthroplasty10,12 and the results suggest that anatomic brain changes can be reversed following pain relief. In addition, in another study examining the effects of CBT, increases in regional gray matter were noted following the therapy, reflecting that the psychological intervention leads to an adaptive response to cope with the pain.8 CBP treatments, such as the application of a Lidocaine patch,4,30 spine surgery and facet joint blocks,13 and acupuncture,11 have all been shown to lead to reversible brain changes using MRI metrics. Finally, fMRI has been used in patients with FM to examine the success of various treatments such as behavioral extinction treatment,9 and administration of pregabalin42 and milnacipran48 to reduce symptoms.
Studies Using MRI Metrics to Predict Transition From Acute to Chronic Pain
Our literature search identified a number of recent studies that have set out to use MRI to identify neuroimaging metrics that can predict transitioning from acute to chronic pain. All studies were performed in patients with subacute back pain, as well as healthy controls. Patients were followed over a period of 1 year and data were collected at 4 visits during the year.15–17 VBM analysis of anatomic brain data revealed that healthy controls, patients with persisting back pain and those who recover show small regional gray matter changes over the year due to normal aging.15 However, patients with persisting pain have significant longitudinal changes in gray matter density, specifically in the bilateral NAc and insula, as well as the left somatosensory cortex.15 More interestingly though are the results of functional MRI assessments, where patients rated their spontaneous back pain. The observations revealed connectivity differences between patients with persisting pain and those that recover, and these differences were evident on the initial visit soon after pain onset.15 A strong significant positive functional connectivity between the mPFC and NAc was found to accurately predict transition to chronicity in persons with back pain.15 Another study also used fMRI and examined hippocampal processing due to its functional connectivity to mPFC and NAc.17 The connectivity between the hippocampus and the mPFC was different between patients with persisting pain and those who recovered over the year in that decreases were noted in the hippocampus-mPFC connectivity longitudinally in the patients with persisting pain.17 This study thus further revealed that brain functional connectivity may help in identifying whether pain will persist and become chronic or if patients will recover. In another study, DTI parameters, specifically a decrease in FA, was suggested to be involved in the etiology for predisposing individuals to pain chronification.16
To summarize, only a few recent studies have attempted to identify MRI markers that can be used to predict the transition from acute to chronic pain.15–17 VBM as well as fMRI17 are the main imaging metrics that were used to evaluate the longitudinal brain changes present with persisting pain. DTI was also explored,16 with the FA value being used as a marker for pain chronification. Using fMRI, brain functional connectivity, specifically a strong significant positive functional connectivity between the mPFC and NAc, was found to accurately predict transition to chronicity in individuals with back pain.15
Our comprehensive scoping review identified mounting MRI evidence demonstrating structural, functional, chemical, and blood flow abnormalities in the brain of patients with various chronic pain conditions. A common finding to most studies addressing morphologic brain alterations in chronic pain populations is that there are losses in regional gray matter density in chronic pain sufferers in regions associated with pain modulation3,7,23,25,31,37,39,43,49,51,64,66,68,69,72,74,78,80 although a handful of studies have noted increases in gray matter density49,54,60,66,71 and some studies have even noted no observed differences between patients and healthy controls.43,58,70,72 A number of reasons may be behind the variable observed differences. The age of patients examined could play a role, as a study looking at FM patients has noted that older patients (above 50 y old) show decreases in regional gray matter volumes compared with healthy controls, whereas an increase in regional gray matter is observed in younger FM sufferers.39 If researchers control for affective disorders such as depression or anxiety that are present in chronic pain conditions, no differences in gray matter volume may be observed, as 1 study has found when controlling for such factors in individuals with FM.43 Other methodological reasons for the observed differences between research groups include small sample size and nonstandardized data acquisition (ie, MRI acquisition parameters) or analysis approaches.
Although brain analysis using VBM is the approach used by most researchers to quantify regional brain volumes, probing white matter integrity noninvasively with DTI has become possible with advances in MRI technology, thereby providing new insight into the brain of persons with chronic pain.6,39,57,59,61,68,70,83 Most DTI studies have revealed altered tissue microstructure and deficits of white matter integrity in chronic pain conditions, with a common finding of reduced FA in various white matter regions6,57,61,68,83 correlating with injury. It is anticipated that more studies will utilize DTI in the future to further our knowledge of underlying tissue abnormality in various chronic pain populations.
Although not widely applied, 1H-MRS is a promising technique for regional biochemical assessments in the brain of chronic pain patients.27,28,41,58 Impaired metabolite concentrations have been noted in various forms of chronic pain,27,28,41,58 suggesting that chemical markers, such as NAA or Glu can be useful in understanding chronic pain processing and targeting therapies to treat metabolite imbalances. The biggest limitations of 1H-MRS are the scan length (∼7 to 9 min), large voxel size (typically 8 cm3) leading to partial voluming, and the large number of overlapping metabolites within the spectrum. To make this approach more usable for future pain studies higher magnetic fields (ie, 7 T and greater) will be required for increased SNR and hence the ability to reduce voxel size and reduce metabolite spectral overlap. In addition, novel approaches to speeding 1H-MRS acquisitions such as the use of compressed sensing will unquestionably help in making this method something that is more favorable.
Assessment of brain activity using stimulus-based approaches as well as resting-state fMRI have revealed that central pain processing is modified in chronic pain conditions.22,24,38–40,44–47,50,52,56,57,62,63,67,70,76–79 Many chronic pain disorders also show abnormal connectivity of RSNs.40,52,55,73,77 Although there is vast evidence of cortical alterations, both functionally and structurally, in chronic pain sufferers, the neuroplastic changes that accompany chronic pain have been shown to be reversible following interventional procedures in numerous studies as evident from MRI neuroimaging approaches.4,8–10,12,13,30,42,48,65
This comprehensive scoping review aimed at finding evidence to demonstrate structural, functional, chemical, and blood flow abnormalities in the brain of patients with various chronic pain conditions with the aim of being able to make the findings of such studies relevant to the clinician. However, one of the limitations of our study is that neuroimaging is currently still in its infancy with regards to its clinical applicability. Thus, while we have identified numerous studies that have found altered brain connectivity or gray matter density for instance with various chronic pain conditions, further research is necessary before it can be reliably used to diagnose and prognosticate for a patient who have chronic pain. Several major issues prevent the clinical application of neuroimaging to the diagnosis and prognosis of chronic pain, including issues such as inadequate sensitivity and specificity of the various imaging metrics to ensure sufficient positive and negative predictive value, suboptimal reliability of activation associated with pain sensation,85 variable findings between studies that may or may not be explained by factors such as sampling issues, confounders like age, affective disorders, medication use, or hydration status.86 According to Woo and Wager, the desirable characteristics of a neuroimaging-based biomarker include the ability of the biomarker to diagnose or provide appropriate sensitivity and specificity, it must correlate with existing scientific models, thereby making it meaningful, it should be deployable and generalizable.87 These criteria preclude the clinical application of the advanced MRI neuroimaging techniques that were examined in our review in the assessment of chronic pain, as such techniques are still in their infancy and have yet to be further explored before being reliably included in the clinic. That said, the exploration of the use of MRI as an objective method to track treatment response in chronic pain conditions seems to be promising and could be of clinical value. Numerous studies are examining whether the neuroplastic changes that accompany chronic pain are reversible after cessation of the agent that causes the pain using objective MRI measures.4,8–10,12,13,30,42,48,65 Such studies, as well as the more recent studies that are using MRI to identify neuroimaging metrics that can predict transitioning from acute to chronic pain, seem to be the area where future research is heading and could be of great value if such techniques prove to be of diagnostic and prognostic value. However, there needs to be more research to test the reproducibility of the current limited studies available in this area. In addition, while the use of MRI enables numerous assessments to be carried out, such as functional, structural, chemical, and blood flow evaluations, each imaging tool comes with its complexities and limitations. Most of these advanced imaging techniques require significant support from physics/engineering departments for proper data collection and analysis and are thus associated with a significant cost and expertise to obtain accurate imaging metrics. In addition, image acquisition may require extra equipment (in the case of stimulus-based fMRI) and can take longer than a routine sequence. Clinicians have to be aware of these issues and need to realize that there remains many areas that need to be explored before MRI can be implemented into clinical practice to evaluate cortical involvement in chronic pain.
1. Reitsma ML, Tranmer JE, Buchanan DM, et al. The prevalence of chronic pain and pain-related interference in the Canadian population from 1994 to 2008. Chronic Dis Inj Can. 2011;31:157–164.
2. Gaskin DJ, Richard P. The economic costs of pain in the United States. J Pain. 2012;13:715–724.
3. Apkarian AV, SoSa Y, Sonty S, et al. Chronic back pain is associated with decreased prefrontal and thalamic gray matter density. J Neurosci. 2004;24:10410–10415.
4. Baliki MN, Geha PY, Jabakhanji R, et al. A preliminary fMRI study of analgesic treatment in chronic back pain and knee osteoarthritis. Mol Pain. 2008;4:47–59.
5. Han SD, Buchman AS, Arfanaki S, et al. Functional connectivity networks associated with chronic musculoskeletal pain in old age. Int J Geriatr Psychiatry. 2013;28:858–867.
6. Moayedi M, Weissman-Fogel I, Salomons TV, et al. White matter brain and trigeminal nerve abnormalities in temporomandibular disorder. Pain. 2012;153:1467–1477.
7. Obermann M, Rodriguez-Raecke R, Naegel S, et al. Gray matter volume reduction reflects chronic pain in trigeminal neuralgia. NeuroImage. 2013;74:352–358.
8. Seminowicz DA, Shapner M, Keaser ML, et al. Cognitive-behavioral therapy increases prefrontal cortex gray matter in patients with chronic pain. J Pain. 2013;14:1573–1584.
9. Diers M, Yilmaz P, Rance M, et al. Treatment-related changes in brain activation in patients with fibromyalgia syndrome. Exp Brain Res. 2012;218:619–628.
10. Gwilym SE, Filippini N, Douaud G, et al. Thalamic atrophy associated with painful osteoarthritis of the hip is reversible after arthroplasty: a longitudinal voxel-based morphometric study. Arthritis Rheum. 2010;62:2930–2940.
11. Li J, Zhang J-H, Yi T, et al. Acupuncture treatment of chronic low back pain reverses an abnormal brain default mode network in correlation with clinical pain relief. Acupunct Med. 2014;32:102–108.
12. Rodriguez-raecke R, Niemeier A, Ihle K, et al. Brain gray matter decrease in chronic pain is the consequence and not the cause of pain. J Neurosci. 2009;29:13746–13750.
13. Seminowicz DA, Wideman TH, Naso L, et al. Effective treatment of chronic low back pain in humans reverses abnormal brain anatomy and function. J Neurosci. 2011;31:7540–7550.
14. Apkarian A, Baliki MN, Farmer MA. Predicting transition to chronic pain. Curr Opin Neurol. 2013;26:360–367.
15. Baliki MN, Petre B, Torbey S, et al. Corticostriatal functional connectivity predicts transition to chronic back pain. Nat Neurosci. 2012;15:1117–1119.
16. Mansour AR, Baliki MN, Haung L, et al. Brain white matter structural properties predict transition to chronic pain. Pain. 2013;154:2160–2168.
17. Mutso AA, Peter B, Huang L, et al. Reorganization of hippocampal functional connectivity with transition to chronic back pain. J Neurophysiol. 2014;111:1065–1076.
18. Henry JL. The need for knowledge translation in chronic pain. Pain Res Manag. 2008;13:465–476.
19. Baliki MN, Baria AT, Apkarian AV. The cortical rhythms of chronic back pain. J Neurosci. 2011;31:13981–13990.
20. Baliki MN, Chialvo DR, Geha PY, et al. Chronic pain and the emotional brain: specific brain activity associated with spontaneous fluctuations of intensity of chronic back pain. J Neurosci. 2006;26:12165–12173.
21. Baliki MN, Geha PY, Apkarian AV, et al. Beyond feeling: chronic pain hurts the brain, disrupting the default-mode network dynamics. J Neurosci. 2008;28:1398–1403.
22. Baliki MN, Geha PY, Fields HL, et al. Predicting value of pain and analgesia: nucleus accumbens response to noxious stimuli changes in the presence of chronic pain. Neuron. 2010;66:149–160.
23. Baliki MN, Schnitzer TJ, Bauer WR, et al. Brain morphological signatures for chronic pain. PLoS ONE. 2011;6:e26010.
24. Barke A, Baudewig J, Schmidt-Samoa C, et al. Neural correlates of fear of movement in high and low fear-avoidant chronic low back pain patients: an event-related MRI study. Pain. 2012;153:540–552.
25. Buckalew N, Haut MW, Morrow L, et al. Chronic pain is associated with brain volume loss in older adults: preliminary evidence. Pain Med. 2008;9:240–248.
26. Giesecke T, Gracely RH, Grant MAB, et al. Evidence of augmented central pain processing in idiopathic chronic low back pain. Arthritis Rheum. 2004;50:613–623.
27. Grachev ID, Fredrickson BE, Apkarian AV. Abnormal brain chemistry in chronic back pain: an in vivo proton magnetic resonance spectroscopy study. Pain. 2000;89:7–18.
28. Gussew A, Rzanny R, Güllmar D, et al. 1 h-mr spectroscopic detection of metabolic changes in pain processing brain regions in the presence of non-specific chronic low back pain. NeuroImage. 2011;54:1315–1323.
29. Hashmi JA, Baliki MN, Huang L, et al. Shape shifting pain: chronification of back pain shifts brain representation from nociceptive to emotional circuits. Brain. 2013;136(pt 9):2751–2768.
30. Hashmi JA, Baliki MN, Huang L, et al. Lidocaine patch (5%) is no more potent than placebo in treating chronic back pain when tested in a randomised double blind placebo controlled brain imaging study. Mol Pain. 2012;8:29–39.
31. Ivo R, Nicklas A, Dargel J, et al. Brain structural and psychometric alterations in chronic low back pain. Eur Spine J. 2013;22:1958–1964.
32. Kobayashi Y, Kurata J, Sekiguchi M, et al. Augmented cerebral activation by lumbar mechanical stimulus in chronic low back pain patients: an MRI study. Spine. 2009;34:2431–2436.
33. Loggi AML, Kim J, Gollub RL, et al. Default mode network connectivity encodes clinical pain: an arterial spin labeling study. Pain. 2013;154:24–33.
34. Tagliazucchi E, Balenzuela P, Fraiman D, et al. Brain resting state is disrupted in chronic back pain patients. Neurosci Lett. 2010;485:26–31.
35. Ung H, Brown JE, Johnson KA, et al. Multivariate classification of structural MRI data detects chronic low back pain. Cereb Cortex. 2014;24:1037–1044.
36. Valet M, Gündel H, Sprenger T, et al. Patients with pain disorder show gray-matter loss in pain-processing structures: a voxel-based morphometric study. Psychosom Med. 2009;71:49–56.
37. Burgmer M, Gaubitz M, Konrad C, et al. Decreased gray matter volumes in the cingulo-frontal cortex and the amygdala in patients with fibromyalgia. Psychosom Med. 2009;71:566–573.
38. Burgmer M, Pogatzki-Zahn E, Gaubitz M, et al. Fibromyalgia unique temporal brain activation during experimental pain: a controlled fmri study. J Neural Transm. 2010;117:123–131.
39. Ceko M, Bushnell MC, Fitzcharles M-A, et al. Fibromyalgia interacts with age to change the brain. YNICL. 2013;3:249–260.
40. Cifre I, Sitges C, Fraiman D, et al. Disrupted functional connectivity of the pain network in fibromyalgia. Psychosom Med. 2012;74:55–62.
41. Feraco P, Bacci A, Pedrabissi F, et al. Metabolic abnormalities in pain-processing regions of patients with fibromyalgia: a 3t MR spectroscopy study. AJNR Am J Neuroradiol. 2011;32:1585–1590.
42. Harris RE, Napadow V, Huggins JP, et al. Pregabalin rectifies aberrant brain chemistry, connectivity, and functional response in chronic pain patients. Anesthesiology. 2013;119:1453–1464.
43. Hsu MC, Harris RE, Sundgren PC, et al. No consistent difference in gray matter volume between individuals with fibromyalgia and age-matched healthy subjects when controlling for affective disorder. Pain. 2009;143:262–267.
44. Jensen KB, Srinivasan P, Spaeth R, et al. Overlapping structural and functional brain changes in patients with long-term exposure to fibromyalgia pain. Arthritis Rheum. 2013;65:3293–3303.
45. Kim J-Y, Kim S-H, Seo J, et al. Increased power spectral density in resting-state pain-related brain networks in fibromyalgia. Pain. 2013;154:1792–1797.
46. Lee SJ, Jinsong H, Decety J, et al. Do patients with fibromyalgia show abnormal neural responses to the observation of pain in others? Neurosci Res. 2013;75:305–315.
47. Napadow V, Lacount L, Park K, et al. Intrinsic brain connectivity in fibromyalgia is associated with chronic pain intensity. Arthritis Rheum. 2010;62:2545–2555.
48. Petzke F, Jensen KB, Kosek E, et al. Using MRI to evaluate the effects of milnacipran on central pain processing in patients with fibromyalgia. Scand J Pain. 2013;4:65–74.
49. Schmidt-Wilcke T, Luerding R, Weigand T, et al. Striatal grey matter increase in patients suffering from fibromyalgia–a voxel-based morphometry study. Pain. 2007;132(suppl 1):S109–S116.
50. Abrahamsen R, Dietz M, Lodahl S, et al. Effect of hypnotic pain modulation on brain activity in patients with temporomandibular disorder pain. Pain. 2010;151:825–833.
51. Gerstner G, Ichesco E, Quintero A, et al. Changes in regional gray and white matter volume in patients with myofascial-type temporomandibular disorders: a voxel-based morphometry study. J Orofac Pain. 2011;25:99–106.
52. Kucyi A, Moayedi M, Weissman-Fogel I, et al. Enhanced medial prefrontal-default mode network functional connectivity in chronic pain and its association with pain rumination. J Neurosci. 2014;34:3969–3975.
53. Moayedi M, Weissman-Fogel I, Crawley AP, et al. Contribution of chronic pain and neuroticism to abnormal forebrain gray matter in patients with temporomandibular disorder. NeuroImage. 2011;55:277–286.
54. Younger JW, Shen YF, Goddard G, et al. Chronic myofascial temporomandibular pain is associated with neural abnormalities in the trigeminal and limbic systems. Pain. 2010;149:222–228.
55. Cauda F, D’Agata F, Sacco K, et al. Altered resting state attentional networks in diabetic neuropathic pain. J Neurol Neurosurg Psychiatr. 2010;81:806–811.
56. Cauda F, Sacco K, D’Agata F, et al. Low-frequency bold fluctuations demonstrate altered thalamocortical connectivity in diabetic neuropathic pain. BMC Neurosci. 2009;10:138–152.
57. Gustin SM, Peck CC, Cheney LB, et al. Pain and plasticity: is chronic pain always associated with somatosensory cortex activity and reorganization? J Neurosci. 2012;32:14874–14884.
58. Gustin SM, Peck CC, Wilcox SL, et al. Different pain, different brain: thalamic anatomy in neuropathic and non-neuropathic chronic pain syndromes. J Neurosci. 2011;31:5956–5964.
59. Yoon EJ, Kim YK, Shin HI, et al. Cortical and white matter alterations in patients with neuropathic pain after spinal cord injury. Brain Res. 2013;1540:64–73.
60. Blankstein U, Chen J, Diamant NE, et al. Altered brain structure in irritable bowel syndrome: potential contributions of preexisting and disease-driven factors. Gastroenterology. 2010;138:1783–1789.
61. Ellingson BM, Mayer E, Harris RJ, et al. Diffusion tensor imaging detects microstructural reorganization in the brain associated with chronic irritable bowel syndrome. Pain. 2013;154:1528–1541.
62. Hong J-Y, Kilpatrick LA, Labus J, et al. Patients with chronic visceral pain show sex-related alterations in intrinsic oscillations of the resting brain. J Neurosci. 2013;33:11994–12002.
63. Kwan CL, Diamant NE, Pop EG, et al. Abnormal forebrain activity in functional bowel disorder patients with chronic pain. Neurology. 2005;65:1268–1277.
64. Seminowicz DA, Labus JS, Bueller JA, et al. Regional gray matter density changes in brains of patients with irritable bowel syndrome. Gastroenterology. 2010;139:48–57.e2.
65. Apkarian AV, Thomas PS, Krauss BR, et al. Prefrontal cortical hyperactivity in patients with sympathetically mediated chronic pain. Neurosci Lett. 2001;311:193–197.
66. Barad MJ, Ueno T, Younger J, et al. Complex regional pain syndrome is associated with structural abnormalities in pain-related regions of the human brain. J Pain. 2014;15:197–203.
67. Freund W, Wunderlich AP, Stuber G, et al. Different activation of opercular and posterior cingulate cortex (pcc) in patients with complex regional pain syndrome (crps i) compared with healthy controls during perception of electrically induced pain: a functional MRI study. Clin J Pain. 2010;26:339–347.
68. Geha PY, Baliki MN, Harden RN, et al. The brain in chronic CRPS pain: abnormal gray-white matter interactions in emotional and autonomic regions. Neuron. 2008;60:570–581.
69. As-Sanie S, Harris RE, Napadow V, et al. Changes in regional gray matter volume in women with chronic pelvic pain: a voxel-based morphometry study. Pain. 2012;153:1006–1014.
70. Farmer MA, Chanda ML, Parks EL, et al. Brain functional and anatomical changes in chronic prostatitis/chronic pelvic pain syndrome. J Urol. 2011;186:117–124.
71. Schweinhardt P, Kuchinad A, Pukall CF, et al. Increased gray matter density in young women with chronic vulvar pain. Pain. 2008;140:411–419.
72. Schmidt-Wilcke T, Leinisch E, Straube A, et al. Gray matter decrease in patients with chronic tension type headache. Neurology. 2005;65:1483–1486.
73. Schwedt TJ, Schlaggar BL, Mar S, et al. Atypical resting-state functional connectivity of affective pain regions in chronic migraine. Headache. 2013;53:737–751.
74. Valfrè W, Rainero I, Bergui M, et al. Voxel-based morphometry reveals gray matter abnormalities in migraine. Headache. 2008;48:109–117.
75. Parks EL, Geha PY, Baliki MN, et al. Brain activity for chronic knee osteoarthritis: dissociating evoked pain from spontaneous pain. Eur J Pain. 2011;15:843.e1–14.
76. Noll-Hussong M, Otti A, Wohlschlaeger AM, et al. Neural correlates of deficits in pain-related affective meaning construction in patients with chronic pain disorder. Psychosom Med. 2013;75:124–136.
77. Otti A, Guendel H, Wohlschläger A, et al. Frequency shifts in the anterior default mode network and the salience network in chronic pain disorder. BMC Psychiatry. 2013;13:84–93.
78. Vartiainen N, Kallio-laine K, Hlushchuk Y, et al. Changes in brain function and morphology in patients with recurring herpes simplex virus infections and chronic pain. Pain. 2009;144:200–208.
79. Khan SA, Keaser ML, Meiller TF, et al. Altered structure and function in the hippocampus and medial prefrontal cortex in patients with burning mouth syndrome. Pain. 2014;155:1472–1480.
80. Agostini A, Benuzzi F, Filippini N, et al. New insights into the brain involvement in patients with crohn’s disease: a voxel-based morphometry study. Neurogastroenterol Motil. 2013;25:147–e82.
81. Ashburner J, Friston KJ. Voxel-based morphometry—the methods. NeuroImage. 2000;11(pt 1):805–821.
82. Mechelli A, Price C, Friston K, et al. Voxel- based morphometry of the human brain: methods and applications. Curr Med Imaging Rev. 2005;1:1–9.
83. Buckalew N, Haut MW, Aizenstein H, et al. Differences in brain structure and function in older adults with self-reported disabling and nondisabling chronic low back pain. Pain Med. 2010;11:1183–1197.
84. Petersen ET, Zimine I, Ho YC, et al. Non-invasive measurement of perfusion: a critical review of arterial spin labeling techniques. Br J Radiol. 2006;79:688–701.
85. Letzen JE, Sevel LS, Gay CW, et al. Test-retest reliability of pain-related brain activity in healthy controls undergoing experimental thermal pain. J Pain. 2014;15:1008–1014.
86. Nakamura K, Brown RA, Araujo D, et al. Correlation between brain volume changes and T2 relaxation time induced by dehydration and rehydration: implications for monitoring atrophy in clinical studies. NeuroImage Clin. 2014;6:166–170.
87. Woo C-W, Wager TD. Neuroimaging-based biomarker discovery and validation. Pain. 2015;156:1379–1381.