Patients with chronic pain experience tremendous variability in pain intensity and how their pain impacts daily function. This variability is likely due to a complex interaction of brain and behavioural factors. The association between negative psychological factors, pain, and brain dysfunction is well studied, but the brain–behavior relationship between positive factors, such as resilience, and pain is not known. Therefore, we examined the interplay between functional connectivity (FC) of intrinsic brain networks within the dynamic pain connectome,26 pain intensity, and behavioural resilience, a key determinant of the pain experience.62
Intrinsic brain networks, including the default mode network (DMN) consist of brain regions with synchronous fluctuations in activity in a resting state, termed FC. Normally, activity in the DMN is anticorrelated with activity of the salience network (SN), and not correlated with the sensorimotor network (SMN).14,15 However, pain alters DMN2,8,21,28,29 and these abnormalities are linked to pain attention, rumination, and pain-related depression.2,13,28,29 Previous studies in chronic pain demonstrate that both reduced DMN anticorrelation with the SN and attention networks, and hyperconnectivity between nodes of the DMN.2,3,8,20–22,40 But there is no consensus on the precise relationship between the DMN, its relationship with other brain networks, and pain. We have proposed a tri-network model of chronic pain, based on our finding that pain is associated with hyperconnectivity among networks of the dynamic pain connectome; the DMN, SMN, and SN.20
Resilience is an individual's ability to “bounce back” in response to adverse life events, including pain.45 We found that acute pain affect and resilience are linked in healthy individuals.19 In chronic pain, resilience is closely associated with pain-related outcomes, including pain acceptance, catastrophizing, and adjustment.42,46,47,50–52,62 The neural correlates of resilience in chronic pain are unknown, but the relationship between resilience and the DMN has been studied in other conditions in which resilience affects health outcomes.5,12,23,54 For example, heightened within-DMN FC is found in posttraumatic stress disorder.5,12,54 Considering the role of resilience in pain, it is crucial to understand the relationship between chronic pain, resilience, and the DMN, including within-network FC and cross-network FC with the dynamic pain connectome (SN and SMN).
Here, we examined whether DMN connectivity tracks pain specifically, or whether this brain–behaviour relationship in chronic pain is more generally related to resilience. The study aims were to determine the relationships between (1) resilience and chronic pain severity, (2) chronic pain and disease severity and within-/cross-network FC with the DMN, and (3) resilience and DMN within-/cross-network FC, in healthy individuals and patients with chronic pain. We hypothesized that within-DMN FC is related to resilience, and that patients with chronic pain would exhibit within- and cross-network connectivity that tracks their clinical pain. Towards these goals, we examined resting-state FC in back pain patients with ankylosing spondylitis (AS), a form of spondyloarthritis, and measured trait resilience, clinical pain report, and arthritis disease activity.
We recruited 102 men (51 patients with a diagnosis of AS and 51 healthy controls) between the ages of 18 and 61 years through advertisements posted at University Health Network hospitals in Toronto, ON, Canada. The study was restricted to men because AS presents predominately in young or middle-aged men, who are otherwise healthy.6
All study procedures were approved by the University Health Network Research Ethics Board, and all participants provided informed written consent. Diagnosis of active AS was made or confirmed by the AS clinic at Toronto Western Hospital, Toronto, ON, Canada using the modified New York criteria,33 which include the presence of lower back pain, limited spinal mobility, and radiographic sacroiliitis. All patients with AS were receiving either a nonsteroidal anti-inflammatory drug for pain or alternatively, a stable dose of a TNF-alpha inhibitor biologic drug for a minimum of 16 weeks. Exclusion criteria for both patients and healthy controls were: (1) currently experiencing ongoing pain such as headaches or toothaches (with the exception of AS-related pain for patients), (2) any history of chronic pain (with the exception of AS for patients), (3) a diagnosis of a psychiatric or neurological disorder such as depression or attention deficit hyperactive disorder, (4) taking any ongoing medications (with the exception of the aforementioned medications for AS for patients), and (5) diagnosis of any major health condition, such as heart disease or diabetes. All participants were instructed to refrain from caffeine consumption for a minimum of 1 hour before the study and from alcohol for a minimum of 8 hours before the study.
2.2. Clinical assessment and questionnaires
All participants completed the full 25-item Resilience Scale (range 25-175, with 140 considered average resilience)55 on the same day as the magnetic resonance imaging (MRI) scan. The Resilience Scale is well validated with high internal consistency and test–retest reliability.55,57 Patients were asked to rate the intensity of their back pain by making a mark on a scale to indicate their pain “now” on a printed 0 to 10 scale where 0 indicated “none” and 10 indicated “maximum,” on the same day as the MRI scan. In addition, the patient's most recent clinic report for the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) score was obtained (0-10 scale, 10 is high disease activity). The BASDAI scores were reported on average within 52 days of the MRI scan (median 35 days and maximum 11 months). The scale is a composite measure of back pain and nocturnal back pain, fatigue, and morning stiffness, and the severity of each symptom (on a 0-10 scale) is averaged together to create a composite BASDAI score.
2.3. Image acquisition
A 3T GE Signa HDx MRI scanner was used to acquire brain images (GE Medical Systems, Milwaukee, WI). The scanner was fitted with an 8-channel phased-array head coil. A T1-weighted anatomical scan was first acquired for each participant, with the scan parameters: 3D IR-FSPGR sequence, 1 × 1 × 1 mm3 voxels, 256 × 256 matrix, 180 axial slices, FOV 25.6 cm, time to echo = 3 ms, T1 = 450 ms, TR = 7.8 ms, and flip angle = 15°. Next, a 10-minute resting-state scan was obtained. Participants were instructed to keep their eyes closed, not to think of anything in particular, and not to fall asleep. The following parameters were used: T2*-weighted echo-planar imaging sequence, 64 × 64 matrix, 40 axial slices, time to recovery = 2000 ms, time to echo = 30 ms, FOV = 20 cm, 3.125 × 3.125 × 4 mm3 voxels, and flip angle = 85°.
2.4. Data preprocessing
Functional MRI (fMRI) data were preprocessed using MATLAB version 188.8.131.52 (Mathworks), fMRISTAT, and FSLv5.0, using methods previously described in detail.10,20,29 First, the first 4 volumes of each 4D scan volume were deleted. Motion correction was applied using FSL's MCFLIRT. After performing brain extracting using optiBET36 on the T1-weighted anatomical scan, FSL's FLIRT was used to perform linear registration between the participant's fMRI scan and their anatomical scan. Next, nonlinear registration to standard MNI152 2 mm space (10 mm warp) was performed using FNIRT. To remove artifacts related to scanner and physiological noise, aCompCor procedures were implemented.9 Briefly, the T1-weighted anatomical images were first segmented into white matter (WM), gray matter, and cerebrospinal fluid (CSF) using FSL's FAST, and then transformed into fMRI space. White matter and CSF maps were thresholded to retain the top 20 and 198 cm3, respectively. Principal component analysis was performed on the WM and CSF signals, and the top 5 WM and CSF components, along with 6 motion parameters obtained using MCFLIRT, were regressed out of the data. A bandpass temporal filter was applied (0.01-0.1 Hz), and the data were spatially smoothed with a 6-mm full-width half-maximum kernel.
2.5. Definition of resting-state networks and seed regions
We defined 7 key regions of the SN: the right temporoparietal junction (rTPJ) and left temporoparietal junction, right and left dorsolateral prefrontal cortex, mid-cingulate cortex, and right and left anterior insula. We also defined 6 key regions of the DMN, including the DMN core (medial prefrontal cortex [mPFC]) and posterior cingulate cortex (PCC), as well as the right and left lateral temporal cortex and right and left inferior parietal cortex (Fig. 1). These seed regions were defined from seed-to-voxel connectivity maps created using data from 50 healthy individuals and have been previously reported.20,29 The rTPJ time series was used as the seed region for the SN, and the PCC was used for the DMN in these seed-to-voxel analyses to identify these connectivity maps and key brain regions of each network, as has been previously reported.20 The PCC and rTPJ were chosen because they are major hubs of the DMN7,31,32,44 and SN,11,27 respectively. The resulting z score maps were thresholded at z equal to 1 unit below the z score of the peak voxel in each key brain region, resulting in the identification of 7 and 6 key brain regions to represent the SN and DMN, respectively. Key regions of the SMN (the left primary sensorimotor cortex, the right primary sensorimotor cortex, and the supplementary motor cortex) were defined using the Greicius functional atlas (Fig. 1).49
2.6. Statistical analysis
2.6.1. Behavioural measures
A Student t test was used to test for differences between the resilience scores of patients with chronic pain and healthy control participants, and Spearman correlations were performed to assess the relationship between resilience and clinical pain score, and resilience and BASDAI score in patients, with alpha set at 0.05. To account for 2 comparisons, the P values were adjusted using the false discovery rate (FDR).
2.6.2. Within- and cross-network seed-based functional connectivity analysis
To assess the strength of FC within (ie, intranetwork) the DMN in patients with chronic pain and healthy control participants, we calculated the Fisher-transformed correlation between the 2 core DMN nodes; the mPFC and PCC. These 2 regions of the DMN were chosen because they represent the common core of the network, which can be dissociated into several subsystems.1,17,25,44 We represented within-DMN strength by the correlation between these 2 core brain regions because the connectivity between other subsystems of the DMN and the common core varies over time and with different modes of cognition or tasks.1,25 Therefore, the common core may best represent the strength of the network.
To assess the strength of FC between the DMN and SN or DMN and SMN (ie, cross-network connectivity), we calculated the Fisher-transformed correlation between each node in the DMN and each node in the SN or with each node in the SMN, as we have performed previously.20 We then averaged the Fisher-transformed Pearson correlations for all connections to obtain a single value that represents cross-network connectivity of the DMN with the SN and the DMN with the SMN.
2.6.3. Neural correlates of clinical pain and disease activity
We performed Spearman correlations to assess the relationship between within- and cross-network DMN connectivity (within-DMN, DMN–SN, and DMN–SMN) and aspects of chronic pain (clinical pain report on the day of the study and overall disease activity score). We adjusted P values using FDR to account for 6 multiple comparisons (2 clinical measures across 3 networks or network pairs). When statistically significant correlations were found, we checked for any differences in FC between the patient and healthy control groups using a Student t test. To do this, we split the patients into a high pain subgroup (reporting pain ≥3/10, n = 22) and a low/no pain subgroup (reporting pain <3/10, n = 28), as described previously.20 It is important to note that all patients had received a diagnosis of AS for which the diagnostic criteria include the presence of chronic back pain.33 Therefore, all patients can be considered to be patients with chronic pain. However, we wished to further differentiate between individuals who experience higher clinical pain intensity on the day of the scan and those who did not because chronic AS-related pain—such as many forms of chronic pain—varies day-to-day. A clinical pain-intensity report on the day of the study was not available for 1 patient.
Although we made every effort to precisely age-match between our patient and control groups, the patients with chronic pain (n = 51) were slightly older than the healthy control participants (n = 51) (Table 1; healthy mean age ± SD = 31.1 ± 9.1, patient mean age ± SD = 36.7 ± 10.8, t = 2.84, P = 0.005). Therefore, to ensure that each patient subgroup (low/no pain subgroups and high pain subgroups described above) was appropriately age-matched, we selected 2 subgroups of healthy controls in which participants were individually age-matched to correspond to the 2 patient subgroups. Statistical tests in which patients with chronic pain were directly compared with healthy control subjects were performed using these age-matched subgroups.
2.6.4. Neural correlates of resilience
To assess the relationship between resilience and default mode within- and cross-network connectivity, we calculated both (1) the relationship between resilience and within- and cross-network connectivity in healthy control participants, and (2) the relationship between resilience and within- and cross-network connectivity in patients with chronic pain. We used Pearson correlations when the data were normally distributed, but when the data were non-normally distributed, we determined Spearman correlations. To correct for multiple comparisons, FDR correction was performed for each network or pair or networks within each group, totaling 3 multiple comparisons per group. To assess whether significant differences existed between the relationships between within- or cross-network connectivity and resilience across groups, we performed Spearman correlations and then used a Fisher r-to-z transformation to transform the rho values to z values.38
2.6.5. Supplementary analysis
We carefully ensured proper motion correction procedures in preprocessing of the fMRI data to minimize the potential effects of head motion on the results. However, it is possible that head motion could still be a confounder. Therefore, we also checked for group differences in relative frame-wise displacement using a Student t test and repeated all main analyses including relative frame-wise displacement as a covariate.
We also confirmed whether our significant findings related to cross-network connectivity (calculated as the connectivity between all nodes in the DMN and the SN or SMN) remained consistent when the core nodes of the DMN (the PCC and mPFC) were used instead, as these key nodes were used in our analysis of within-network strength of the DMN.
3.1. Descriptive statistics and behaviour
Clinical and demographic information for both patients and healthy control participants are presented in Table 1.
There was no age difference between the low/no pain subgroup and their individually matched healthy controls, (n = 28, t = 0.04, P = 0.96), nor the high pain subgroup and their individually matched healthy controls (n = 22, t = 1.20, P = 0.24).
There was a large range (0-8) in the clinical pain intensity experienced by patients with chronic pain on the day of the scan, confirming the necessity to assess differences between patients and healthy controls in 2 separate subgroups (high pain and low/no pain). The average BASDAI score reported by patients was 3.7/10. As expected, individuals with higher pain also reported higher BASDAI scores (5.1/10) vs 2.1/10 in patients with low pain.
There was no statistically significant difference in resilience between patients and controls; however, there was a trend towards a lower level of resilience in the patients with chronic pain (t = 1.77, P = 0.08), driven by a small number of patients with very low resilience scores (Fig. 2). We found that there was a significant negative relationship between resilience and both clinical pain scores (rho = −0.35, P = 0.03, FDR corrected) and BASDAI disease activity scores in patients (rho = −0.35, P = 0.02, FDR corrected) (Fig. 2).
3.2. Default mode network within- and cross-network functional connectivity and chronic pain
Cross-network connectivity of the DMN was abnormal for patients with high levels of clinical pain but not for patients with low levels of clinical pain. Specifically, in patients (n = 22) reporting high levels of clinical pain, there was significantly stronger (positive) cross-network connectivity between the DMN and the SMN compared with 22 age-matched healthy control participants (t = 2.92, P = 0.006, healthy mean z = −0.007, patient mean z = 0.129) (Fig. 3). Conversely, in patients reporting lower levels of clinical pain, there was no group difference between the patients and their matched healthy control group (t = 0.98, P = 0.33, healthy mean r = −0.003, patient mean r = −0.38). Notably, while the DMN and SMN were predominantly decoupled (correlation around zero) or anticorrelated in age-matched healthy control participants, 18 patients exhibited significant positive connectivity between the DMN and SMN (r > 0.12 P < 0.05, uncorrected, dotted gray line in Fig. 3). Furthermore, across all patients (n = 51), clinical pain scores were positively correlated with cross-network connectivity between the DMN and SMN (rho = 0.47, P = 0.004, FDR corrected) (Fig. 4A). The relationship was significant even when a partial correlation controlling for BASDAI score was performed, indicating some specificity of the relationship to pain, as opposed to disease activity more generally (r = −0.35, P = 0.01). No significant relationship was found between pain and within-DMN (rho = 0.13, P = 0.64, FDR corrected) or DMN–SN cross-network connectivity (rho = −0.09, P = 0.64, FDR corrected) (Fig. 4A). Furthermore, no significant relationship was found between BASDAI score and within- or cross-network connectivity (DMN; rho = 0.03, P = 0.82, DMN–SN; rho = −0.18, P = 0.60, DMN–SMN; rho = 0.10, P = 0.64, FDR corrected) (Fig. 4B).
3.3. Within- and cross-network functional connectivity in the default mode network and resilience in healthy control participants and patients with chronic pain
The relationship between resilience and within- and cross-network connectivity of the DMN was assessed in both healthy control participants and patients with chronic pain. In healthy control participants, higher within-DMN connectivity (r = −0.43, P = 0.001, FDR corrected), but not cross-network connectivity (DMN–SN; r = −0.07, P = 0.61, DMN–SMN; r = 0.11, P = 0.61, FDR corrected), was associated with lower resilience scores (Fig. 5). In patients, there was no relationship between resilience and within- or cross-network connectivity (DMN; rho = 0.06, P = 0.88, DMN–SN; rho = 0.02, P = 0.88, DMN–SMN; rho = −0.2, P = 0.48, FDR corrected) (Fig. 5). The negative relationship between within-DMN connectivity and resilience in healthy individuals was significantly different from the (insignificant) relationship in patients (z = 2.61, P = 0.005). Although there was no significant relationship between resilience and cross-network DMN–SMN connectivity in either group, there was a significant group difference in DMN–SMN cross-network connectivity (the relationship was more negative in patients; z = 1.69, P = 0.046). There was no group difference in the relationship between cross-network DMN–SN connectivity in patients with chronic pain vs healthy controls (z = 0.35, P = 0.363).
As stated above, chronic pain patients with higher clinical pain exhibited abnormal DMN cross-network connectivity (Fig. 3), and resilience and clinical pain scores are moderately negatively correlated (Fig. 2). The within-DMN–resilience relationship might be different in patients with high pain vs patients with low pain. Therefore, we further probed the relationship between within-DMN FC and resilience in patients with chronic pain while including clinical pain intensity and an interaction term in a multiple linear regression model [overall model: F(3,46) = 11.69, P < 0.001]. There were significant main effects of within-DMN connectivity (t = −2.652, P = 0.01), clinical pain score (t = −4.88, P < 0.01), and their interaction (t = 3.93, P < 0.01). As seen in Figure 5D, the relationship between resilience and within-DMN connectivity in patients is most similar to that of healthy control participants when little or no pain (<3/10) is experienced. By contrast, the opposite relationship is seen in patients reporting high clinical pain intensity on the day of the MRI scan. Thus, in patients experiencing greater pain, who tend to exhibit more positive DMN–SMN cross-network FC, the typically negative (in healthy control participants) relationship between resilience and within-DMN connectivity tends to be disrupted.
3.4. Additional analysis: no significant effects of head motion or choice of default mode network seed regions on main findings
We found no significant differences between patients and healthy controls in relative frame-wise displacement (patients: mean = 0.09 mm, healthy controls: mean = 0.09 mm, t = 0.69, P = 0.49). One main finding of this study was a positive relationship between DMN–SMN cross-network connectivity and clinical pain score. We performed a partial Spearman correlation between DMN–SMN cross-network connectivity and clinical pain score correcting for relative head motion, and found that the original relationship was still statistically significant (rho = 0.5, P < 0.001). Another prominent finding was a negative relationship between resilience and within-DMN connectivity in the healthy control population. Therefore, we also performed a partial Pearson correlation between within-DMN connectivity and resilience in healthy individuals, controlling for relative head motion. The relationship remained statistically significant (r = −0.43, P = 0.001). Finally, when relative head motion was included as a covariate in our linear model explaining resilience, each main effect and interaction remained significant (clinical pain report, t = −4.86, P < 0.001; within-DMN FC, t = −2.63, P = 0.01; interaction t = 3.93, P < 0.001), and there was no significant effect of motion (t = 0.49, P = 0.63).
In the previous section, we found that cross-network connectivity between the DMN and SMN was significantly positively correlated with clinical pain and was significantly higher in the high pain patient group than a matched healthy control subgroup. When we analysed the cross-network correlations again using only the core nodes of the DMN, the findings remained significant (DMN–SMN FC correlated with pain: r = 0.43, P = 0.001; DMN–SMN FC in patients with high pain vs controls: t = 2.52, P = 0.016).
In this study, we reveal a triad relationship between resilience, within- and cross-network connectivity of the DMN, and clinical pain in individuals with chronic pain due to AS, a form of spondyloarthritis. A summary of the key findings is depicted in Figure 6 and highlights our findings that (1) in healthy individuals, low resilience scores were observed for those with greater within-DMN FC, (2) in patients with chronic pain, resilience was negatively associated with clinical pain intensity and disease activity, and (3) in patients with high levels of clinical pain, the DMN and SMN are abnormally functionally connected, and DMN-SMN functional connectivity is related to the level of clinical pain. Furthermore, we highlight an interaction between the within-DMN–resilience brain–behaviour relationship and clinical pain intensity; in patients experiencing high pain, who also exhibit abnormally high DMN–SMN cross-network connectivity, the negative relationship between within-DMN connectivity and resilience observed in healthy individuals disappears, possibly because they are not experiencing a true “resting state” due to their chronic pain.
The importance of resilience in chronic pain is gaining traction. More research study on positive psychological factors is critical because resilience predicts functioning in patients with chronic pain independent from the commonly studied vulnerability factors, such as anxiety and negative affect.62 Here, we report that more resilient individuals report lower clinical pain and arthritis disease activity, in line with previous studies on resilience and chronic back pain.47,51,62 For example, resilience has been found to predict adjustment to pain, pain acceptance, and pain catastrophizing in patients with chronic pain.42,46,47 Furthermore, resilience can predict pain-related outcomes (eg, depression and quality of life).4,41,42 Finally, in patients with osteoarthritis, resilience is a better predictor of pain and physical functioning than disease severity, which emphasizes the importance of resilience in chronic pain.58 However, given that AS is caused by genetic and environmental rather than psychological factors,53 it was not surprising that the range of resilience scores in our patients with AS was similar to the controls. Consistent with the literature, we propose that resilience in patients with AS is a relatively stable trait that may promote positive adaption to chronic pain.18 However, resilience is also partially malleable and could therefore be a target for intervention aimed at improving one's ability to adapt to life with chronic pain. For example, Acceptance and Commitment Therapy promotes flexible thinking (a mechanism of resilience), and Positive Activity Interventions work to promote positive behaviour (as opposed to reduce negative or problematic behaviour).18 These interventions could be a target of future study.
Resilience is a positive psychological factor; however, most studies of the DMN and psychological or cognitive factors focus on “risk factors” that have a negative impact on mental health. The DMN is most active when an individual is not engaged in a task requiring external attention, but instead focused on internal tasks or cognition.7,43 In individuals with depression, hyperconnectivity is found within the DMN and is linked to the degree to which an individual ruminates.48 Furthermore, lower levels of happiness in healthy individuals have also been associated with high within-DMN FC.35 Therefore, although positive and negative psychological traits are not precisely inverse constructs, we might expect that resilience, a positive psychological trait, would be negatively associated with within-DMN FC. Indeed, our findings confirm this negative relationship in healthy individuals. In individuals with chronic pain, both rumination about pain (a form of pain catastrophizing) and pain intensity have been found to be positively correlated with FC within the DMN.28,34 The DMN is generally deactivated during a task involving an external stimulus, including pain.29 We have shown, in a task in which healthy individuals were exposed to a painful stimulus, that the degree of DMN activation depends on whether the individual reports mind-wandering away from pain.29 In the current study, in patients with chronic pain, an interaction between clinical pain intensity and within-DMN FC explained resilience in a multiple regression model: In individuals with less clinical pain, a negative resilience–DMN relationship was found similar to that of healthy controls, but the opposite relationship was found in individuals experiencing higher clinical pain. We speculate that individuals with higher levels of pain might experience altered pain-related cognition and within-DMN connectivity during the study, and that altered DMN–SMN cross-network connectivity in this group may also contribute to altered within-DMN connectivity.
Two novel findings in the current study were a positive relationship between DMN–SMN connectivity and clinical pain report, and hyperconnectivity between these networks in higher pain patients. Previously, we and others have found a link between SN- or insula-DMN FC and chronic pain; however, a DMN–SMN abnormality in chronic pain has not been directly explored.2,8,20,21,29 Notably, the relationship between pain and cross-network connectivity was driven predominantly by individuals with high pain intensity and concurrent positive DMN and SMN FC. This abnormality could be due to neuroplasticity following a lengthy period of sensory stimulation (pain). In general, the DMN and SMN are decoupled (uncorrelated),37 but in amputees, increased DMN–SMN FC is associated with changes within the SMN after injury.37 An alternative explanation for this abnormality considers the pain–attention relationship. Building on previous work in which we demonstrated DMN deactivation while attending to pain, when the SMN is activated due to constant pain, an individual may no longer simultaneously deactivate the DMN.29 Thus, these 2 typically separate networks may become hyperconnected. In support of the latter explanation, we found that clinical pain report was still associated with DMN–SMN cross-network connectivity after controlling for overall disease activity score. Finally, our findings corroborate our previous work on a separate cohort of patients with AS: In a series of studies, we demonstrated structural and functional changes in the dynamic pain connectome, and proposed a triple-network model of chronic pain (involving the DMN, SN, and SMN) in AS including a link between DMN–SMN connectivity and daily functional ability.20,59–61 Although our findings confirm our work and that of others on cross-network connectivity in chronic pain, the relationship between cross-network connectivity, attention to pain, and changes in pain intensity is not yet known and should be a target of future study.
The distinction we have made between the separate neural correlates of 2 tightly linked chronic pain–related factors—resilience and pain intensity—is important because it highlights integral aspects of the brain–behaviour relationship in chronic pain, which are not understood. Commonly, when DMN function is reported to be different in individuals with chronic pain, it has been related to a measure of pain intensity or severity, with minimal exploration into the potentially diverse roles of different behavioural factors.39,40,56,63 Chronic pain is multifaceted, and different aspects of emotion, affect, and perception that may be relevant to patients with chronic pain may be associated with different underlying mechanisms. Recently, more emphasis has been placed on teasing apart and identifying commonalities between the roles of the DMN in pain, cognitive control, and emotion.24 Our work builds on this shift by differentiating further between individuals with chronic pain.
We note some limitations of our study. First, the relationship between chronic pain and DMN–SMN cross-network connectivity may be specific to chronic pain intensity reported on the day of the study but not to overall arthritis disease activity. In the future, it would be important to perform a longitudinal study on the natural fluctuations in chronic pain intensity that occur on a timescale of days to weeks and to determine whether brain networks differ because of these fluctuations. Future longitudinal studies in patients with chronic pain are also needed to delineate the relationship between pain perception while the patient is undergoing a resting-state MRI scan, and the neural correlates of more stable factors such as resilience. Second, here, we have only assessed chronic pain due to AS, a form of arthritis that affects the spine. Different forms of chronic pain can exhibit different brain–behaviour relationships and therefore, the relationship between the DMN and resilience may differ in other chronic pain conditions.30 Third, we acknowledge that this study was restricted to men because they account for the vast majority of patients with AS.6,16 Therefore, whether the observed brain–behaviour relationships and chronic pain–related functional brain abnormalities generalize to women is not known.
In conclusion, we have characterized the multifaceted relationship between chronic pain, resilience, and the DMN. Although resilience, pain, and disease activity were tightly linked in individuals with chronic pain, they were each linked to different features of the dynamic pain connectome: We provide evidence of abnormal cross-network DMN–SMN connectivity in patients with higher clinical pain, and demonstrate a link between DMN–SMN cross-network connectivity and pain. Regarding resilience, we uncover a negative relationship between resilience and within-DMN connectivity in healthy individuals, and suggest that this relationship is disrupted in individuals experiencing higher clinical pain. Our findings point to the DMN as a functional brain network that plays a key role in multiple dimensions in chronic pain.
Conflict of interest statement
The authors have no conflicts of interest to declare.
This work was supported by the Canadian Institutes of Health Research and the Mayday Fund. K.S. Hemington is a recipient of a Queen Elizabeth II/Purdue Pharma Scholarship in Science and Technology, an Arthritis Society of Canada Scholarship, and a CIHR Doctoral Research Award.
The authors thank Dr Aaron Kucyi for the development of the fMRI data preprocessing pipeline and the participants for volunteering their time. They also thank Renise Ayearst for clinical data retrieval, Daeria Lawson and Ammepa Anton for patient recruitment, and Eugene Hlasny and Keith Ta for expert technical assistance with MRI acquisition.
Supplemental video content
Video content associated with this article can be found online at http://links.lww.com/PAIN/A573.
. Andrews-Hanna JR, Reidler JS, Sepulcre J, Poulin R, Buckner RL. Functional-anatomic fractionation of the brain's default network. Neuron 2010;65:550–62.
. Baliki MN, Geha PY, Apkarian AV, Chialvo DR. Beyond feeling: chronic pain
hurts the brain, disrupting the default-mode network dynamics. J Neurosci 2008;28:1398–403.
. Baliki MN, Mansour AR, Baria AT, Apkarian AV. Functional reorganization of the default mode network
across chronic pain
conditions. PLoS One 2014;9:e106133.
. Bauer H, Emeny RT, Baumert J, Ladwig KH. Resilience
moderates the association between chronic pain
and depressive symptoms in the elderly. Eur J Pain 2016;20:1253–65.
. Bluhm RL, Williamson PC, Osuch EA, Frewen PA, Stevens TK, Boksman K, Neufeld RW, Théberge J, Lanius RA. Alterations in default network connectivity in posttraumatic stress disorder related to early-life trauma. J Psychiatry Neurosci 2009;34:187.
. Braun J, Sieper J. Ankylosing spondylitis
. Lancet 2007;369:1379–90.
. Buckner RL, Andrews-Hanna JR, Schacter DL. The brain's default network. Ann N Y Acad Sci 2008;1124:1–38.
. Čeko M, Shir Y, Ouellet JA, Ware MA, Stone LS, Seminowicz DA. Partial recovery of abnormal insula and dorsolateral prefrontal connectivity to cognitive networks in chronic low back pain after treatment. Hum Brain Mapp 2015;36:2075–1092.
. Chai XJ, Castañón AN, Ongür D, Whitfield-Gabrieli S. Anticorrelations in resting state networks without global signal regression. Neuroimage 2012;59:1420.
. 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.
. Downar J, Crawley AP, Mikulis DJ, Davis KD. A multimodal cortical network for the detection of changes in the sensory environment. Nat Neurosci 2000;3:277–83.
. Dunkley BT, Doesburg SM, Sedge PA, Grodecki RJ, Shek PN, Pang EW, Taylor MJ. Resting-state hippocampal connectivity correlates with symptom severity in post-traumatic stress disorder. Neuroimage 2014;5:377–84.
. Erpelding N, Davis KD. Neural underpinnings of behavioural strategies that prioritize either cognitive task performance or pain. PAIN 2013;154:2060–71.
. Fox MD, Snyder AZ, Vincent JL, Corbetta M, Van Essen DC, Raichle ME. The human brain is intrinsically organized into dynamic, anticorrelated functional networks. Proc Natl Acad Sci U S A 2005;102:9673–8.
. Fox MD, Zhang D, Snyder AZ, Raichle ME. The global signal and observed anticorrelated resting state brain networks. J Neurophysiol 2009;101:3270–83.
. Golder V, Schachna L. Ankylosing spondylitis
: an update. Aust Fam Physician 2013;42:780.
. Greicius MD, Krasnow B, Reiss AL, Menon V. Functional connectivity
in the resting brain: a network analysis of the default mode hypothesis. Proc Natl Acad Sci U S A 2003;100:253.
. Hassett AL, Finan PH. The role of resilience
in the clinical management of chronic pain
. Curr Pain Headache Rep 2016;20:39.
. 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.
. 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.
. Hubbard CS, Khan SA, Keaser ML, Mathur VA, Goyal M, Seminowicz DA. Altered brain structure and function correlate with disease severity and pain catastrophizing in migraine patients. eNeuro 2014;1:e20.14.
. Ichesco E, Schmidt-Wilcke T, Bhavsar R, Clauw DJ, Peltier SJ, Kim J, Napadow V, Hampson JP, Kairys AE, Williams DA, Harris RE. Altered resting state connectivity of the insular cortex in individuals with fibromyalgia. J Pain 2014;15:815–26.e811.
. Kong F, Wang X, Hu S, Liu J. Neural correlates of psychological resilience
and their relation to life satisfaction in a sample of healthy young adults. NeuroImage 2015;123:165–72.
. Kragel PA, Kano M, Oudenhove LV, Ly HG, Dupont P, Rubio A, Delon-Martin C, Bonaz BL, Manuck SB, Gianaros PJ, Ceko M, Losin EAR, Woo CW, Nichols TE, Wager TD. Generalizable representations of pain, cognitive control, and negative emotion in medial frontal cortex. Nat Neurosci 2018;21:283–9.
. Kucyi A, Davis KD. Dynamic functional connectivity
of the default mode network
tracks daydreaming. Neuroimage 2014;100:471–80.
. Kucyi A, Davis KD. The dynamic pain connectome
. Trends Neurosci 2015;38:86–95.
. Kucyi A, Hodaie M, Davis KD. Lateralization in intrinsic functional connectivity
of the temporoparietal junction with salience- and attention-related brain networks. J Neurophysiol 2012;108:3382–92.
. Kucyi A, Moayedi M, Weissman-Fogel I, Goldberg MB, Freeman BV, Tenenbaum HC, Davis KD. Enhanced medial prefrontal-default mode network functional connectivity
in chronic pain
and its association with pain rumination. J Neurosci 2014;34:3969–75.
. Kucyi A, Salomons TV, Davis KD. Mind wandering away from pain dynamically engages antinociceptive and default mode brain networks. Proc Natl Acad Sci 2013;110:18692–7.
. Lee YC, Nassikas NJ, Clauw DJ. The role of the central nervous system in the generation and maintenance of chronic pain
in rheumatoid arthritis, osteoarthritis and fibromyalgia. Arthritis Res Ther 2011;13:211.
. Leech R, Kamourieh S, Beckmann CF, Sharp DJ. Fractionating the default mode network
: distinct contributions of the ventral and dorsal posterior cingulate cortex to cognitive control. J Neurosci 2011;31:3217–24.
. Leech R, Sharp DJ. The role of the posterior cingulate cortex in cognition and disease. Brain 2014;137:12–32.
. Linden SVD, Valkenburg HA, Cats A. Evaluation of diagnostic criteria for ankylosing spondylitis
. Arthritis Rheum 1984;27:361–8.
. Loggia ML, Kim J, Gollub RL, Vangel MG, Kirsch I, Kong J, Wasan AD, Napadow V. Default mode network
connectivity encodes clinical pain: an arterial spin labeling study. PAIN 2013;154:24–33.
. Luo Y, Kong F, Qi S, You X, Huang X. Resting-state functional connectivity
of the default mode network
associated with happiness. Soc Cogn Affect Neurosci 2016;11:516–24.
. Lutkenhoff ES, Rosenberg M, Chiang J, Zhang K, Pickard JD, Owen AM, Monti MM. Optimized brain extraction for pathological brains (optibet). PLoS One 2014;9:e115551.
. Makin TR, Filippini N, Duff EP, Henderson Slater D, Tracey I, Johansen-Berg H. Network-level reorganisation of functional connectivity
following arm amputation. Neuroimage 2015;114:217–25.
. Myers L, Sirois MJ. Spearman correlation coefficients, differences between. Encyclopedia Stat Sci 2006;12.
. Napadow V, Kim J, Clauw DJ, Harris RE. Decreased intrinsic brain connectivity is associated with reduced clinical pain in fibromyalgia. Arthritis Rheum 2012;64:2398–403.
. Napadow V, LaCount L, Park K, As-Sanie S, Clauw DJ, Harris RE. Intrinsic brain connectivity in fibromyalgia is associated with chronic pain
intensity. Arthritis Rheum 2010;62:2545–55.
. Ong AD, Bergeman CS, Bisconti TL, Wallace KA. Psychological resilience
, positive emotions, and successful adaptation to stress in later life. J Pers Soc Psychol 2006;91:730–49.
. Ong AD, Zautra AJ, Reid MC. Psychological resilience
predicts decreases in pain catastrophizing through positive emotions. Psychol Aging 2010;25:516–23.
. Raichle ME. The brain's default mode network
. Annu Rev Neurosci 2015;38:433–47.
. Raichle ME, MacLeod AM, Snyder AZ, Powers WJ, Gusnard DA, Shulman GL. A default mode of brain function. Proc Natl Acad Sci 2001;98:676–82.
. Reich JW, Zautra AJ, Hall JS. Handbook of Adult Resilience
. New York, NY: Guilford Press, 2010.
. Ruiz-Párraga G, López-Martínez A. The role of experiential avoidance, resilience
and pain acceptance in the adjustment of chronic back pain patients who have experienced a traumatic event: a path analysis. Ann Behav Med 2015;49:247–57.
. Ruiz-Párraga GT, López-Martínez AE, Esteve R, Ramírez-Maestre C, Wagnild G. A confirmatory factor analysis of the Resilience
Scale adapted to chronic pain
(RS-18): new empirical evidence of the protective role of resilience
on pain adjustment. Qual Life Res 2015;24:1245–53.
. Sheline YI, Barch DM, Price JL, Rundle MM, Vaishnavi SN, Snyder AZ, Mintun MA, Wang S, Coalson RS, Raichle ME. The default mode network
and self-referential processes in depression. Proc Natl Acad Sci 2009;106:1942–7.
. Shirer WR, Ryali S, Rykhlevskaia E, Menon V. Decoding subject-driven cognitive states with whole-brain connectivity patterns. Cereb Cortex 2012;22:158–65.
. Smith BW, Zautra AJ. Vulnerability and resilience
in women with arthritis: test of a two-factor model. J Consult Clin Psychol 2008;76:799–810.
. Sturgeon JA, Zautra AJ. Resilience
: a new paradigm for adaptation to chronic pain
. Curr Pain Headache Rep 2010;14:105–12.
. Suffeda A, Meissner W, Rosendahl J, Guntinas-Lichius O. Influence of depression, catastrophizing, anxiety, and resilience
on postoperative pain at the first day after otolaryngological surgery: a prospective single center cohort observational study. Medicine 2016;95:e4256.
. Taurog JD, Chhabra A, Colbert RA. Ankylosing spondylitis
and axial spondyloarthritis. New Engl J Med 2016;374:2563–74.
. van der Werff SJA, Pannekoek JN, Veer IM, van Tol MJ, Aleman A, Veltman DJ, Zitman FG, Rombouts SARB, Elzinga BM, van der Wee NJA. Resilience
to childhood maltreatment is associated with increased resting-state functional connectivity
of the salience network with the lingual gyrus. Child Abuse Negl 2013;37:1021–9.
. Wagnild GM, Young HM. Development and psychometric evaluation of the resilience
scale. J Nurs Meas 1993;1:165–78.
. Wasan AD, Loggia ML, Chen LQ, Napadow V. Neural correlates of chronic low back pain measured by arterial spin labeling. Anesthesiology 2011;115:364–74.
. Windle G, Bennett KM, Noyes J. A methodological review of resilience
measurement scales. Health Qual Life Outcomes 2011;9:8.
. Wright LJ, Zautra AJ, Going S. Adaptation to early knee osteoarthritis: the role of risk, resilience
, and disease severity on pain and physical functioning. Ann Behav Med 2008;36:70–80.
. Wu Q, Inman RD, Davis KD. Neuropathic pain in ankylosing spondylitis
: a psychophysics and brain imaging study: neuropathic pain in as. Arthritis Rheum 2013;65:1494–503.
. Wu Q, Inman RD, Davis KD. Fatigue in ankylosing spondylitis
is associated with the brain networks of sensory salience and attention: neuroimaging of fatigue-associated brain networks in as. Arthritis Rheumatol 2014;66:295–303.
. Wu Q, Inman RD, Davis KD. Tumor necrosis factor inhibitor therapy in ankylosing spondylitis
: differential effects on pain and fatigue and brain correlates. PAIN 2015;156:297–304.
. Yeung EW, Arewasikporn A, Zautra AJ. Resilience
and chronic pain
. J Social Clin Psychol 2012;31:593–617.
. Yu R, Gollub RL, Spaeth R, Napadow V, Wasan A, Kong J. Disrupted functional connectivity
of the periaqueductal gray in chronic low back pain. Neuroimage Clin 2014;6:100–8.
Chronic pain; Resilience; Default mode network; Functional connectivity; Resting-state fMRI; Ankylosing spondylitis; Sensorimotor network; Dynamic pain connectome
Supplemental Digital Content
© 2018 International Association for the Study of Pain