Secondary Logo

Imaging vs quantitative sensory testing to predict chronic pain treatment outcomes

Davis, Karen D.a,b

doi: 10.1097/j.pain.0000000000001479
NeuPSIG Closing Debate
Free

In this article, I review the concept of personalized pain management and consider how brain imaging and quantitative sensory testing can be used to derive biomarkers of chronic pain treatment outcome. I review how different modalities of brain imaging can be used to acquire information about brain structure and function and how this information can be linked to individual measures of pain.

Brain imaging is providing insight into developing brain- and behaviour-based biomarkers of chronic pain treatment outcomes.

aDepartment of Surgery and Institute of Medical Science, University of Toronto, Toronto, ON, Canada

bDivision of Brain, Imaging, and Behaviour - Systems Neuroscience, Krembil Brain Institute, Krembil Research Institute, Toronto Western Hospital, University Health Network, Toronto, ON, Canada

Corresponding author. Address: Toronto Western Hospital, Room: MP12-306, 399 Bathurst St, Toronto, ON M5T 2S8, Canada. Tel.: (416) 603-5662. E-mail address: karen.davis@uhnresearch.ca (K.D. Davis).

Sponsorships or competing interests that may be relevant to content are disclosed at the end of this article.

Received December 02, 2018

Accepted December 20, 2018

Back to Top | Article Outline

1. Introduction

Despite the availability of many therapeutics that can effectively manage chronic pain in some individuals, there is a tremendous unmet need to provide adequate pain relief for all who suffer from chronic pain. Thus, a one-size-fits-all approach is time-consuming, expensive, and in many cases is ineffective to treat individuals with chronic pain. Developing a treatment for everyone based on group-based findings is one fallout of the “group-to-individual” (G2i) problem.23 However, over the last decade, there has been a transition from the traditional approach to treat physical and mental health conditions similarly for all, to a more personalized and precision model.34 The cornerstone of this approach is to select a specific therapeutic option based on an individual's particular characteristics known to be sensitive to that therapeutic. The framework introduced to develop precision medicine for mental health disorders is an instructive example of how we might integrate information across multiple domains towards a personalized approach to management pain (Fig. 1). Although it is important to have a fundamental understanding of how different types of treatments work, this knowledge is not always necessary to predict treatment outcomes. Rather, the key information needed to prognosticate treatment outcome is to have biomarkers of the pain relief achieved by the treatment. Such biomarkers might comprise measures of brain structure and/or function, as well as how someone feels, thinks, perceives, or behaves (Fig. 2).

Figure 1

Figure 1

Figure 2

Figure 2

In this review, I will focus on the brain and behaviour. A modern view of pain recognizes that each person experiences pain differently based on their pain sensitivity, capacity to modulate, tolerate, and cope with pain. My view is that pain mechanisms are dictated by a system of anatomical scaffolding, activity, function, and dynamics of brain circuits that mediate pain sensing, pain modulation, and salience/attention 6,28 (Figs. 3 and 4). We have called this system, the dynamic pain connectome.39

Figure 3

Figure 3

Figure 4

Figure 4

Brain biomarkers of chronic pain likely will involve the dynamic pain connectome because it is a system for pain sensing, attention, and modulation.39 In addition, biomarkers based on an individual's inherent characteristics (sex, gender, age, personality, traits, etc.) should be considered for their modulating impact on the brain and behaviour.

Back to Top | Article Outline

2. Individual variability and behavioural biomarkers

Sex, gender, genetics, race, environmental and cultural factors, behavioural states, and traits could potentially contribute to the constellation of biomarkers used to predict chronic pain treatment outcomes. Individuals can vary tremendously in terms of their individual make-up of these factors, which may impact how they sense and respond to noxious stimuli, and their ability to cope with pain. For example, my laboratory has identified intersubject variability (including some sex differences) in many aspects of pain sensitivity (such as threshold, habituation, and adaptation) and resilience.10,28,29 We have also introduced 2 types of behavioural phenotypes related to pain-attention interactions: The A/P phenotype spectrum is based on the disruptive effect of pain on concurrent cognitive task performance; P-type individuals lie at one end of the spectrum (pain dominates and task performance diminished during concurrent pain) and A-type individuals (attention dominates and task performance maintained) lie at the other end.21,52 Another phenotype is based on an individual's intrinsic attention to pain (IAP) score, which is calculated from the degree to which an individual attends to pain vs mind wanders away from pain.41 Both the A/P and IAP phenotypes potentially could be used as biomarkers of treatment outcome, but this has yet to be tested.

Additional behavioural attributes that are potential biomarkers of pain and pain treatment outcomes include individual characteristics measured using standardized questionnaires (eg, for pain catastrophizing, resilience, anxiety, etc.), and threshold and suprathreshold psychophysical measures of pain sensitivity and modulation (eg, quantitative sensory testing). In particular, 2 well-known behavioural measures of pain sensitivity and modulationtemporal summation of pain (TSP) and conditioned pain modulation (CPM) hold promise for predicting chronic pain treatment outcome. Conditioned pain modulation (a type of counterirritation paradigm) is used as a proxy measure for antinociceptive brain attributes such as descending antinociceptive activity within the dynamic pain connectome and pain modulation.57 By contrast, TSP can serve as a proxy measure for pronociceptive brain attributes such as activity in the dynamic pain connectome (in particular, the ascending sensing pathway), pain facilitation, and central sensitization.

Back to Top | Article Outline

3. Imaging biomarkers and brain-behaviour relationships

There are many types of brain attributes that can potentially be used to develop biomarkers of pain treatment outcomes. These indicators can be acquired using structural and/or functional brain imaging with magnetic resonance imaging (MRI), electroencephalograph (EEG), magnetoencephalography (MEG), or positron emission tomography. For a general description of these approaches, see Refs. 6, 14, 18, 39, 40, and 46. Here, I will focus on MRI and MEG.

Structural MRI scans and diffusion tensor imaging can be used to measure structural metrics that provide insight into the cellular and scaffolding aspects of the brain. Thus, one can measure aspects of gray matter (volume, density, cortical thickness, etc.) and of white matter. There are 2 types of white matter assessments that can be used to measure integrity and connectivity. Measures such as fractional anisotropy, as well as mean diffusivity, radial diffusivity, and axial diffusivity can provide information pertaining to the health of the white matter, and abnormalities of these metrics have been associated with a variety of conditions such as edema, inflammation, and demyelination.18 Tractography can be used to assess the strength of structural connectivity between brain areas, typically mapped using fractional anisotropy values.

Functional imaging measures include evoked responses, functional connectivity (FC), and regional activity. Most relevant to chronic pain conditions is that FC (also known as functional coupling in MEG studies) and regional activity can be assessed during “resting state” conditions using functional MRI (fMRI), arterial spin labeling, positron emission tomography, EEG, and MEG. These are useful to examine the “chronic pain” brain associated with ongoing pain because no additional stimulus needs to be applied. There are 2 main types of FC assessments—those that measure how brain areas show general synchrony over time (static FC [sFC]) and those that measure how a pair or network of brain areas exhibit dynamic (flexible) synchrony over time (dynamic FC [dFC]). Dynamics in the activity within a region of the brain can also be examined by measuring regional BOLD variability or fractional amplitude of low-frequency fluctuations during a resting state fMRI scan, or through spectral analyses of MEG or EEG data.

My laboratory and others has linked many attributes of individual variability in acute pain sensitivity measures, pain-attention phenotypes (A/P and IAP), and personality described above with brain measures of structure, connectivity, and function of the dynamic pain connectome.9,10,20–22,30,41,48,50–52,55 For example, we found that there is a strong link between TSP and FC. Individuals who exhibited high TSP had strong FC in the ascending nociceptive pathway and weak FC in the descending antinociceptive pathway, whereas individuals with low TSP had the opposite connectivity pattern10 (Fig. 5). These findings indicate that individuals not only lie on a behavioural spectrum of pronociception to antinociception as described by Yarnitsky's group25,57 but also a brain connectivity spectrum of pronociception to antinociception.

Figure 5

Figure 5

Back to Top | Article Outline

4. The dynamic pain connectome dictates how we feel pain

Unlike static models of pain, the dynamic pain connectome concept emphasizes temporal dynamics across brain networks of attention, default mode, salience, ascending nociceptive, and descending modulation systems. This concept provides new thinking about brain mechanisms underlying pain and fluctuations across time and conditions. The fine temporal features within regions of the dynamic pain connectome are largely unknown because previous human studies of pain used fMRI whose signals arise from slow hemodynamics. But a deep understanding of pain requires a technique that can measure signals on a timescale that directly relates to neuronal activity. One such technique is MEG that measures fast brain signals with millisecond precision. Magnetoencephalography also has several advantages over EEG, although it is less sensitive to radial dipoles and debates continue regarding deep source localization.4 Nonetheless, MEG provides more accurate source localization than EEG, deep sources are possible,4 and it is more tolerable for patients with chronic pain because the scalp electrode placement is less onerous.4 Furthermore, unlike fMRI, the presence of ferromagnetic implants such as spinal cord stimulator systems is not a contraindication for using MEG.

Back to Top | Article Outline

5. What could go wrong? Coding and communication in the brain, and how to model it all

The brain uses temporal codes to represent perception. Neurons in the brain exhibit spontaneous (ongoing) activity in the absence of a particular task or stimulus. This activity provides a baseline (ie, a type of prior) from which a response to a given stimulus or situation can arise and shape the subsequent stimulus response and behavior. Thus, this prior contributes to behavioral responsivity and sensitivity.

At the population level, synchronous activity of large numbers of neurons produces oscillatory patterns of activity. It has been proposed that an oscillatory form of activity at rest is an efficient way for neuronal assemblies to be kept close to their firing thresholds.8 Brain oscillations are categorized into different frequency bands (delta, theta, alpha, beta, and gamma),8 and different brain areas have characteristic power spectra.36 Each frequency band is associated with several purported functions including roles in excitatory and inhibitory processes within localized regions or long range communications.35 The assessment of firing patterns from populations of neurons within these frequency bands requires a technique that can measure signals on millisecond timescales such as MEG or EEG.

Brain areas that are connected structurally might not be functionally communicating with each other effectively. Magnetoencephalography can be used to evaluate the communication functionality between brain areas. As noted above, FC is a measure of the synchrony of activity between 2 brain areas. Magnetoencephalography can characterize this type of brain communication by measuring different metrics of functional coupling based on the phase and amplitude of oscillations.19 There are several ways in which ineffectual communication between brain areas or networks can arise (Fig. 6); abnormal communication can arise if (1) regional activity of 2 normally functionally connected brain areas exhibit abnormal coupling or (2) if there is abnormal oscillatory activity within one or both of 2 connected brain areas.

Figure 6

Figure 6

Over the last decade, functional and structural abnormalities within areas, networks, and connections of the dynamic pain connectome have been identified in patients with a variety of chronic pain conditions (for reviews see Refs. 6, 14, 15, 18, and 39), mostly using MRI-based brain imaging technologies. Many imaging studies have also linked these abnormalities to levels of pain and the impact of pain on activities of daily living.7,30,31,37,49 In addition to brain abnormalities, we and others have used diffusion tensor imaging–MRI studies to identify structural abnormalities in the trigeminal nerve of patients with orofacial pains such as trigeminal neuralgia and temporomandibular disorders.14,17,18,46 These brain and peripheral nerve studies provide insight into neural dysfunction that have been reversed in some cases after effective treatment14,16,24 (see next section).

It is not clear yet how best to construct a model of chronic pain. Most imaging studies of chronic pain abnormalities have used conventional, univariate statistical approaches. However, given the complexities of brain mechanisms underlying pain and the limitations of fMRI, many groups have developed more sophisticated, multivariate, and machine learning approaches to understand and predict acute pain,54 the transition to chronic pain,1,27 and chronic pain conditions.43–45,53 Despite some promising results, previous, there remain many challenges to establish models of chronic pain that will be applicable to predict treatment outcome in an individual. One area for future development is to explore what type of brain imaging metric (function, structure, sFC/dFC, oscillations, spectral patterns, etc.), should be used from what type of imaging modality (MRI, MEG, etc.) and what individual attributes (for a discussion see Ref. 13) to include in the model such as (1) pain attributes (intensity and quality), (2) pain state (current pain) vs pain trait (average pain over time), (3) pain etiology (neuropathic, inflammatory, etc.), and (4) impact on activities of daily living, mood, behaviour, etc. For example, my laboratory recently used a machine learning approach to construct models of chronic pain based on FC data for patients with a form of arthritis, ankylosing spondylitis.11 We found that the models for state and trait pains were similar but had distinct features. These models were driven by pain severity, patients with neuropathic pain (rather than inflammatory pain), and had features of predominantly dFC (rather than sFC).

Back to Top | Article Outline

6. Predicting chronic pain treatment outcomes

Much of the effort to predict chronic pain treatment outcome is being directed at assessing the brain (see below). However, we have reported that structural abnormalities in the trigeminal nerve are reversed in patients with significant pain relief after microvascular decompression or gamma knife surgery (but not in patients with no clinical benefit of surgery).16 Furthermore, our group has shown that there are a constellation of pretreatment abnormalities in the trigeminal nerve and brain white matter that are predictors of treatment outcome.33,59

In the central nervous system, the analgesic effect of a treatment may depend not only on the ability to resolve a dysfunction (eg, an overactive ascending nociceptive pathway) but also on the presence of a healthy or modifiable pain modulation system. That is, an effective treatment may work through 2 mechanisms: (1) enhancement of the descending pain modulation system and/or (2) removal of an abnormality that is suppressing or overriding a potentially strong endogenous pain modulation system.

Evidence for the first scenario (ie, boosting inefficient descending modulation) comes from studies by Yarnitsky's group56,58 that found duloxetine to be more efficacious to treat diabetic neuropathy pain in people exhibiting deficient pain modulation and/or enhanced pain facilitation. The interpretation was that the drug was more effective in those with less-efficient CPM because the drug is boosting the descending modulation system, whereas patients whose CPM is operating efficiently do not benefit from this drug because the system cannot be further boosted. Furthermore, because CPM is opiate-mediated, it can be potentiated by opioids2,38 but remains unchanged by drugs with a nonopioid mechanism of action.42,47 The TSP response reflects the magnitude of central neuronal sensitization and can be attenuated by N-methyl-D-aspartate antagonists,3 and calcium and sodium channel blockers but not by serotonin–norepinephrine reuptake inhibitors.26 Furthermore, enhanced TSP but not less-efficient CPM was a successful predictor for response to pregabalin in pain patients.47 Thus, accumulating evidence indicates that simple tests of pain modulation (CPM) and pain magnification (TSP) can indicate whether a certain family of drugs is likely or unlikely to benefit an individual patient. Also, our work showing individual and sex differences in the dynamic pain connectome9,10,12,21,28,48,55 suggests that some patients may benefit from boosting or blocking activity in specific brain areas or networks with techniques such as deep brain stimulation, transcranial magnetic stimulation, or spinal cord stimulation based on identifying brain areas in these patients that are operating suboptimally or with weak functional connections to other brain areas.

We recently provided evidence for the second scenario (ie, reducing an overactive ascending nociceptive system that is suppressing a strong descending modulation system)5 (Fig. 7): We showed that the patients who achieved good pain relief from a treatment of intravenous ketamine exhibited 2 pretreatment features that distinguished them from patients who did not benefit from treatment and healthy controls: unusually high TSP and a strong and flexible connectivity within the descending modulation pathway. Although both the brain and TSP were predictive of pain relief, the functional connectivity was a stronger predictor of pain relief than TSP. Ketamine is an N-methyl-D-aspartate inhibitor that acts to reduce spinal cord windup3,32 and thus can attenuate abnormally high activity in the ascending sensing pathway associated with TSP.10 Thus, the ketamine treatment response may have occurred because the drug blocked an overactive ascending sensing pathway, which released and allowed a strong and flexible descending modulation system to work effectively (Fig. 7).

Figure 7

Figure 7

Back to Top | Article Outline

7. Conclusion

There are promising developments in the use of both imaging-based metrics of brain function and structure, as well as measures of pain sensitivity and related individual attributes to predict pain treatment outcome. In the future, the choice of which or how many predictive biomarkers to use most effectively may depend on the type of pain, treatment modality, and practical and cost issues.

Back to Top | Article Outline

Conflict of interest statement

The author has no conflict of interest to declare.

Back to Top | Article Outline

Acknowledgments

Funding for studies discussed in this review include grants to K.D. Davis from the Canadian Institutes of Health Research (including funding from the CHIR SPOR program to the Chronic Pain Network), the Mayday Fund, and the MS Society of Canada. The author thanks all her trainees and collaborators for their contributions to the findings and ideas included in this article.

Back to Top | Article Outline

References

[1]. Apkarian AV, Baliki MN, Farmer MA. Predicting transition to chronic pain. Curr Opin Neurol 2013;26:360–7.
[2]. Arendt-Nielsen L, Andresen T, Malver LP, Oksche A, Mansikka H, Drewes AM. A double-blind, placebo-controlled study on the effect of buprenorphine and fentanyl on descending pain modulation: a human experimental study. Clin J Pain 2012;28:623–7.
[3]. Arendt-Nielsen L, Mansikka H, Staahl C, Rees H, Tan K, Smart TS, Monhemius R, Suzuki R, Drewes AM. A translational study of the effects of ketamine and pregabalin on temporal summation of experimental pain. Reg Anesth Pain Med 2011;36:585–91.
[4]. Baillet S. Magnetoencephalography for brain electrophysiology and imaging. Nat Neurosci 2017;20:327–39.
[5]. Bosma RL, Cheng JC, Rogachov A, Kim JA, Hemington KS, Osborne NR, Venkat Raghavan L, Bhatia A, Davis KD. Brain dynamics and temporal summation of pain predicts neuropathic pain relief from ketamine infusion. Anesthesiology 2018;129:1015–24.
[6]. Bosma RL, Hemington KS, Davis KD. Using magnetic resonance imaging to visualize the brain in chronic pain. PAIN 2017;158:1192–3.
[7]. Bosma RL, Kim JA, Cheng JC, Rogachov A, Hemington KS, Osborne NR, Oh J, Davis KD. Dynamic pain connectome functional connectivity and oscillations reflect multiple sclerosis pain. PAIN 2018;159:2267–76.
[8]. Buzsáki G, Draguhn A. Neuronal oscillations in cortical networks. Science 2004;304:1926–9.
[9]. Cheng JC, Bosma RL, Hemington KS, Kucyi A, Lindquist MA, Davis KD. Slow-5 dynamic functional connectivity reflects the capacity to sustain cognitive performance during pain. Neuroimage 2017;157:61–8.
[10]. Cheng JC, Erpelding N, Kucyi A, DeSouza DD, Davis KD. Individual differences in temporal summation of pain reflect pronociceptive and antinociceptive brain structure and function. J Neurosci 2015;35:9689–700.
[11]. Cheng JC, Rogachov A, Hemington KS, Kucyi A, Bosma RL, Lindquist MA, Inman RD, Davis KD. Multivariate machine learning distinguishes cross-network dynamic functional connectivity patterns in state and trait neuropathic pain. PAIN 2018;159:1764–76.
[12]. Coulombe MA, Erpelding N, Kucyi A, Davis KD. Intrinsic functional connectivity of periaqueductal gray subregions in humans. Hum Brain Mapp 2016;37:1514–30.
[13]. Davis KD, Kucyi A, Moayedi M. The pain switch: an “ouch” detector. PAIN 2015;156:2164–6.
[14]. Davis KD, Moayedi M. Central mechanisms of pain revealed through functional and structural MRI. J Neuroimmune Pharmacol 2013;8:518–34.
[15]. Davis KD, Taylor KS, Anastakis DJ. Nerve injury triggers changes in the brain. Neuroscientist 2011;17:407–22.
[16]. DeSouza DD, Davis KD, Hodaie M. Reversal of insular and microstructural nerve abnormalities following effective surgical treatment for trigeminal neuralgia. PAIN 2015;156:1112–23.
[17]. DeSouza DD, Hodaie M, Davis KD. Abnormal trigeminal nerve microstructure and brain white matter in idiopathic trigeminal neuralgia. PAIN 2014;155:37–44.
[18]. DeSouza DD, Hodaie M, Davis KD. Structural magnetic resonance imaging can identify trigeminal system abnormalities in classical trigeminal neuralgia. Front Neuroanat 2016;10:95.
[19]. Dunkley BT. Differential intrinsic coupling modes in psychological and physical trauma. Front Psychiatry 2015;6:140.
[20]. Emerson NM, Zeidan F, Lobanov OV, Hadsel MS, Martucci KT, Quevedo AS, Starr CJ, Nahman-Averbuch H, Weissman-Fogel I, Granovsky Y, Yarnitsky D, Coghill RC. Pain sensitivity is inversely related to regional grey matter density in the brain. PAIN 2014;155:566–73.
[21]. Erpelding N, Davis KD. Neural underpinnings of behavioural strategies that prioritize either cognitive task performance or pain. PAIN 2013;154:2060–71.
[22]. Erpelding N, Moayedi M, Davis KD. Cortical thickness correlates of pain and temperature sensitivity. PAIN 2012;153:1602–9.
[23]. Fisher AJ, Medaglia JD, Jeronimus BF. Lack of group-to-individual generalizability is a threat to human subjects research. Proc Natl Acad Sci 2018;115:E6106–E6115.
[24]. Goswami R, Anastakis DJ, Katz J, Davis KD. A longitudinal study of pain, personality, and brain plasticity following peripheral nerve injury. PAIN 2016;157:729–39.
[25]. Granovsky Y, Yarnitsky D. Personalized pain medicine: the clinical value of psychophysical assessment of pain modulation profile. Rambam Maimonides Med J 2013;4:e0024.
[26]. Harding LM, Kristensen JD, Baranowski AP. Differential effects of neuropathic analgesics on wind-up-like pain and somatosensory function in healthy volunteers. Clin J Pain 2005;21:127–32.
[27]. Hashmi JA, Baliki MN, Huang L, Baria AT, Torbey S, Hermann KM, Schnitzer TJ, Apkarian AV. Shape shifting pain: chronification of back pain shifts brain representation from nociceptive to emotional circuits. Brain 2013;136:2751–68.
[28]. Hashmi JA, Davis KD. Deconstructing sex differences in pain sensitivity. PAIN 2014;155:10–3.
[29]. Hemington KS, Cheng JC, Bosma RL, Rogachov A, Kim JA, Davis KD. Beyond negative pain-related psychological factors: resilience is related to lower pain affect in healthy adults. J Pain 2017;18:1117–28.
[30]. Hemington KS, Rogachov A, Cheng JC, Bosma RL, Kim JA, Osborne NR, Inman RD, Davis KD. Patients with chronic pain exhibit a complex relationship triad between pain, resilience, and within- and cross-network functional connectivity of the default mode network. PAIN 2018;159:1621–30.
[31]. Hemington KS, Wu Q, Kucyi A, Inman RD, Davis KD. Abnormal cross-network functional connectivity in chronic pain and its association with clinical symptoms. Brain Struct Funct 2016;221:4203–19.
[32]. Herrero JF, Laird JMA, Lopez-Garcia JA. Wind-up of spinal cord neurones and pain sensation: much ado about something? Prog Neurobiol 2000;61:169–203.
[33]. Hung PS, Chen DQ, Davis KD, Zhong J, Hodaie M. Predicting pain relief: use of pre-surgical trigeminal nerve diffusion metrics in trigeminal neuralgia. Neuroimage Clin 2017;15:710–18.
[34]. Insel TR, Cuthbert BN. Medicine. Brain disorders? Precisely. Science 2015;348:499–500.
[35]. Jensen O, Spaak E, Zumer JM. Human brain oscillations: from physiological mechanisms to analysis and cognition. In: Supek SA, Aine CJ, editors. Magnetoencephalography: From signals to dynamic cortical networks. Berlin: Springer-Verlag, 2014. p. 359–403.
[36]. Keitel A, Gross J. Individual human brain areas can be identified from their characteristic spectral activation fingerprints. PLoS Biol 2016;14:e1002498.
[37]. Kim JA, Bosma RL, Hemington KS, Rogachov A, Osborne NR, Cheng JC, Oh J, Crawley AP, Dunkley BT, Davis KD. Neuropathic pain and pain interference are linked to alpha-band slowing and reduced beta-band magnetoencephalography activity within the dynamic pain connectome in patients with multiple sclerosis. PAIN 2019;160:187–97.
[38]. Kim JH, Kim DW, Chang WH, Kim YH, Kim K, Im CH. Inconsistent outcomes of transcranial direct current stimulation may originate from anatomical differences among individuals: electric field simulation using individual MRI data. Neurosci Lett 2014;564:6–10.
[39]. Kucyi A, Davis KD. The dynamic pain connectome. Trends Neurosci 2015;38:86–95.
[40]. Kucyi A, Davis KD. The neural code for pain: from single-cell electrophysiology to the dynamic pain connectome. Neuroscientist 2017;23:397–414.
[41]. Kucyi A, Salomons TV, Davis KD. Mind wandering away from pain dynamically engages antinociceptive and default mode brain networks. Proc Natl Acad Sci USA 2013;110:18692–7.
[42]. Kunz M, Scholl KE, Schu U, Lautenbacher S. GABAergic modulation of diffuse noxious inhibitory controls (DNIC): a test by use of lorazepam. Exp Brain Res 2006;175:363–71.
[43]. Labus JS, Van Horn JD, Gupta A, Alaverdyan M, Torgerson C, Ashe-McNalley C, Irimia A, Hong JY, Naliboff B, Tillisch K, Mayer EA. Multivariate morphological brain signatures predict patients with chronic abdominal pain from healthy control subjects. PAIN 2015;156:1545–54.
[44]. Lopez-Sola M, Woo CW, Pujol J, Deus J, Harrison BJ, Monfort J, Wager TD. Towards a neurophysiological signature for fibromyalgia. PAIN 2017;158:34–47.
[45]. Mansour A, Baria AT, Tetreault P, Vachon-Presseau E, Chang PC, Huang L, Apkarian AV, Baliki MN. Global disruption of degree rank order: a hallmark of chronic pain. Sci Rep 2016;6:34853.
[46]. Moayedi M, Weissman-Fogel I, Salomons TV, Crawley AP, Goldberg MB, Freeman BV, Tenenbaum HC, Davis KD. White matter brain and trigeminal nerve abnormalities in temporomandibular disorder. PAIN 2012;153:1467–77.
[47]. Olesen SS, Graversen C, Bouwense SA, van GH, Wilder-Smith OH, Drewes AM. Quantitative sensory testing predicts pregabalin efficacy in painful chronic pancreatitis. PLoS One 2013;8:e57963.
[48]. Rogachov A, Cheng JC, Erpelding N, Hemington KS, Crawley AP, Davis KD. Regional brain signal variability: a novel indicator of pain sensitivity and coping. PAIN 2016;157:2483–92.
[49]. Rogachov A, Cheng JC, Hemington KS, Bosma RL, Kim JA, Osborne NR, Inman RD, Davis KD. Abnormal low-frequency oscillations reflect trait-like pain ratings in chronic pain patients revealed through a machine learning approach. J Neurosci 2018;38:7293–302.
[50]. Salomons TV, Moayedi M, Weissman-Fogel I, Goldberg MB, Freeman BV, Tenenbaum HC, Davis KD. Perceived helplessness is associated with individual differences in the central motor output system. Eur J Neurosci 2012;35:1481–7.
[51]. Seminowicz DA, Davis KD. Cortical responses to pain in healthy individuals depends on pain catastrophizing. PAIN 2006;120:297–306.
[52]. Seminowicz DA, Mikulis DJ, Davis KD. Cognitive modulation of pain-related brain responses depends on behavioral strategy. PAIN 2004;112:48–58.
[53]. Ung H, Brown JE, Johnson KA, Younger J, Hush J, Mackey S. Multivariate classification of structural MRI data detects chronic low back pain. Cereb Cortex 2014;24:1037–44.
[54]. Wager TD, Atlas LY, Lindquist MA, Roy M, Woo CW, Kross E. An fMRI-based neurologic signature of physical pain. N Engl J Med 2013;368:1388–97.
[55]. Wang G, Erpelding N, Davis KD. Sex differences in connectivity of the subgenual ACC. PAIN 2014;155:755–63.
[56]. Yarnitsky D. Role of endogenous pain modulation in chronic pain mechanisms and treatment. PAIN 2015;156(suppl 1):S24–S31.
[57]. Yarnitsky D, Granot M, Granovsky Y. Pain modulation profile and pain therapy: between pro- and antinociception. PAIN 2014;155:663–5.
[58]. Yarnitsky D, Granot M, Nahman-Averbuch H, Khamaisi M, Granovsky Y. Conditioned pain modulation predicts duloxetine efficacy in painful diabetic neuropathy. PAIN 2012;153:1193–8.
[59]. Zhong J, Chen DQ, Hung PSP, Hayes DJ, Liang KE, Davis KD, Hodaie M. Multivariate pattern classification of brain white matter connectivity predicts classic trigeminal neuralgia. PAIN 2018;159:2076–87.
Keywords:

Biomarker; Imaging, fMRI; Functional connectivity; Temporal summation of pain; Dynamic pain connectome; Modulation; Default mode network; Ketamine; Prediction

© 2019 International Association for the Study of Pain