Sage Journals: Discover world-class research

Abstract

Objective

To investigate the intranetwork resting state fMRI connectivity within the Salience Network of chronic migraine with and without medication overuse headache.

Methods

We compared 351 pairs of intranetwork connectivity in chronic migraine (n = 13) and chronic migraine with medication overuse headache (n = 16) compared to matched controls, and between each chronic migraine subgroup.

Results

Compared to controls, 17 pairs of intranetwork connections in chronic migraine and 27 pairs in chronic migraine with medication overuse headache were decreased. When comparing chronic migraine with medication overuse headache versus chronic migraine, connectivity between bilateral extended amygdala, and between paracingulate to right ventral tegmental area/substantia nigra were decreased in chronic migraine (chronic migraine < chronic migraine with medication overuse headache). Connectivity between left dorsolateral prefrontal cortex to bilateral ventral striatum/pallidum, to bilateral dorsal anterior cingulate cortex; left anterior prefrontal cortex to contralateral orbitofrontal insula; and left ventral striatum/pallidum to ipsilateral supplementary motor area (SMA)/preSMA were decreased in chronic migraine with medication overuse headache (chronic migraine with medication overuse headache < chronic migraine).

Conclusion

Both chronic migraine subgroups had shared intranetwork connectivity abnormality, however, each subgroup had unique pattern of disruption within the salience network. The results suggest that the aberrant assignment of salience to external and internal stimuli plays an important role in chronic migraine and chronic migraine with medication overuse headache interictally, mostly involving mesolimbic pathways (especially bilateral extended amygdala) in chronic migraine, and prefrontal-subcortical limbic pathways in chronic migraine with medication overuse headache.

Keywords

Medication overuse headache amygdala prefrontal cortex limbic subcortical

Introduction

Despite the prevalence and comorbidities of chronic migraine (CM) and CM with medication overuse headaches (MOH), the underlying mechanisms remain poorly understood (1 –5). With recent advances in fMRI, changes in the functional connectivity of brain networks can be investigated. Many of the current neuroimaging studies in CM versus CM with MOH (CM^MOH) were performed using task activated fMRI (6 –8). During a risk decision task, decreased activation in the ventral tegmental area and substantia nigra (VTA/SN) was demonstrated in CM^MOH compared to CM (9). Two recent studies found altered regional functional connectivity in episodic migraine and CM, focused on the amygdala and marginal division of the neostriatum (10,11). However, there is a paucity of research on how the functional connectivity of intrinsic networks is affected in CM and CM^MOH. Given that the brain is a closely integrated system, and chronic pain is a multi-dimensional dynamic process, using established brain networks to evaluate patterns of functional neural disruption may be a unique approach in understanding the pathogenesis of CM and CM^MOH.

Our recent investigation revealed that, within the salience network, there was an overall decrease in network connectivity strength in both CM and CM^MOH; however, we found no difference between each group (16). The salience network is a well-defined, intrinsically correlated baseline brain network important for regulating salience-related information via integrating sensory, emotional, and cognitive inputs (12,13). It is considered to be a filtration and amplification network responsible for evaluating specific external/internal stimuli, and assigning the relevance of the stimuli for continuous processing and modulation of pain, as well as guiding goal-directed behavior (14,15). Given migraine is a complex neurological disorder affecting not only the sensory (painful sensation) but also the emotional, cognitive aspect of pain awareness, evaluation of the salience intranetwork connectivity may shed light on the key brain regions critically dysregulated in CM versus CM^MOH.Currently, there is no report on the salience network intranetwork resting state functional connectivity comparing CM versus CM^MOH.

The purpose of this study is to investigate the intranetwork interictal resting state functional connectivity within the salience network in CM and CM^MOH. We hypothesized that the diminished overall connectivity of this network would be attributed to particular node to node connectivity (NTNC) disruptions within the mesocorticolimbic system, deferentially affected in CM and CM^MOH. Additionally, we hypothesized that there would be a more widespread NTNC disruption in CM^MOH than CM.

Methods

Participants

Women were eligible for the study if they were 18 years old or older, met diagnostic criteria fulfilling International Classification of Headache Disorders III beta for CM as determined by a headache specialist, or were non-pain, non-headache controls. CM subgroups were characterized based on medication overuse status (with or without) as per International Classification of Headache Disorders III beta diagnostic criteria (17). We have carefully selected our CM^MOH patients to ensure that they started with episodic migraine decades ago, then developed CM due to overuse of acute pain medications based on our clinical assessment by a certified headache specialist. Participants were excluded if they had MRI contraindication, neurological or pain disorders other than CM, any chronic illness (i.e. hypertension, diabetes, hepatic, renal, chronic inflammatory, or infectious disease), or inability to follow study protocol while completing assessments. The healthy controls were selected to match the age and gender of the CM participants. Controls were excluded if they had a family history of migraine or used over the counter/prescription pain medication for more than 5 days per month. CM participants were scanned at their baseline level of pain, at least 24 hours outside of their acute pain exacerbation period; any participant who came in within 24 hours of acute pain exacerbation was rescheduled.

All participants underwent vital sign evaluation including BMI, a neurological examination, and completed a standardized questionnaire to ascertain demographics including age, sex, race, and educational level as well as clinical characteristics including: (a) duration of migraine history; (b) duration of CM history; (c) family history of migraine; (d) current medications; (e) number of moderate to severe headache days per month; (f) location of migraine; (g) presence of aura; (h) headache-related disability, as determined by Headache Impact Test (HIT-6); (i) depression, as determined by the Patient Health Questionnaire (PHQ-9); (j) allodynia, as measured by the allodynia symptom checklist (ASC 12) (18).

HIT-6 and PHQ-9 scale

The HIT-6 is a validated questionnaire that consists of six items reflecting quality of life. Higher scores (range of 36–78) indicate an increasing impact of headaches on daily functioning (19). The PHQ-9 is a diagnostic measure for clinical depression. A score of ≥15 on the PHQ-9 is associated with a 68% sensitivity and 95% specificity for diagnosing “major depressive disorder” based on the DSM-IV criteria (20).

Standard protocol approvals, registrations, and patient consents

The study protocol was approved by institutional review board at University of South Carolina. Written informed consent was obtained from all participants.

MR imaging

All participants were scanned on a Siemens 3T scanner located at the McCausland Center for Brain Imaging (Columbia, South Carolina). Participants were instructed to keep their eyes closed, stay awake, relax, and think of “nothing in particular” during the resting state scan. All conditions and lighting were consistent throughout the entire study for all participants. Some participants (five CM and five controls) were scanned prior to a system hardware upgrade; however, any variance due to this upgrade was controlled for in our analysis by adding this as a nuisance regressor variable to the general linear model (GLM) using the Freedman-Lane approach (21).

The imaging parameters for the Trio (12 channel head coil) system consisted of a 6 minute high-resolution T1 weighted magnetization-prepared rapid gradient echo (MP-RAGE) scan (repetition time [TR] = 2250 ms, echo time [TE] = 4.15 ms, 192 slices, 50% slice gap, flip angle = 9°, voxel size = 1.0 mm³, Field of View [FOV] = 256 mm², iPAT factor of 2, and using a sagittal, ascending acquisition) and a 15 minute resting state functional imaging scan using a T2* weighted BOLD contrast-sensitive sequence ([TR] = 1550 ms, [TE] = 34 ms, 42 slices, 20% slice gap, flip angle = 71°, voxel size = 2.5 mm³, FOV = 215 mm², and using a transversal, descending acquisition).

The imaging parameters for the Prisma (20 channel head coil) system included an acquisition of 6 minute high resolution T1 weighted MP-RAGE scan (same parameters as Trio, except that TE = 4.11 ms) and a 15 minute resting state functional imaging scan using a T2* weighted BOLD contrast-sensitive sequence (TR = 1100 ms, TE = 35 ms, 56 slices, 20% slice gap, flip angle = 72°, voxel size = 2.4 × 2.4 × 2.0 mm³, FOV = 216 mm², and using a transversal acquisition).

MRI preprocessing

Rs-fMRI preprocessing was completed using a combination of Statistical Parametric Mapping 12 (SPM12) software and custom MATLAB scripts (https://github.com/neurolabusc/nii_preprocess). The pipeline consisted of standard procedures including motion correction, coregistration, normalization, frequency filtering (0.01 to 0.1 Hz bandpass), and spatial smoothing (8 mm full width at half maximum). For each network, a connectivity atlas was constructed using spherical (15 mm diameter) regions of interest (ROIs) centered on the peak Montreal Neurological Institute (MNI) coordinates for the a priori network of the salience network (12). The salience network (Figure 1) includes the bilateral dorsal anterior cingulate cortex (dACC), bilateral dorsolateral prefrontal cortex (dlPFC), left anterior PFC (aPFC), bilateral hypothalamus, bilateral orbital frontal insula (OFI), paracingulate cortex, bilateral extended amygdala (BA 34), bilateral supramarginal gyrus (SMG), left periaqueductal gray (PAG), bilateral supplementary motor area (SMA)/pre SMA, bilateral ventral tegmental area/substantia nigra (VTA/SN), bilateral superior temporal, bilateral temporal pole, bilateral ventral striatum/pallidum (vSP), and right ventrolateral PFC (vlPFC). The ROIs (nodes) and their MNI coordinates used in the salience network are reported in the supplemental Table 1.

Figure 1.

Axial and sagittal view of the salience network. Locations: (1) Left dlPFC, (2) Right dlPFC, (3) Right vlPFC, (4) Left anterior PFC, (5) Left OFI, (6) Right OFI, (7) Left temporal pole, (8) Right temporal pole, (9) Left SMA/preSMA, (10) Right SMA/preSMA, (11) Paracingulate cortex, (12) Left dACC, (13) Right dACC, (14) Left extended amygdala, (15) Right extended amygdala, (16) PAG, (17) Left hypothalamus, (18) Right hypothalamus, (19) Right dorsomedial thalamus, (20) Left VTA/SN, (21) Right VTA/SN, (22) Left vSP, (23) Right vSP, (24) Left superior temporal, (25) Right superior temporal, (26) Left SMG, (27) Right SMG. Node depth is illustrated by transparency. Images were made with “Surf Ice” (https://www.nitrc.org/projects/surfice/) using the exact MNI coordinate locations.

Table 1.

Demographic and clinical features (CM subgroup only) of subjects based on all subgroups.

Demographics:	CM^MOH	CM	C
n	16	13	19
Age	39 ± 14	39 ± 12	37 ± 11
BMI	26 ± 6	31 ± 4	26 ± 5
Race/ethnicity
white	10 (62.5)	10 (76.9)	12 (63.1)
black	6 (37.5)	3 (23.1)	7 (36.9)
Highest degree completed
Unknown	1 (6.3)	1 (7.7)	1 (5.3)
High school	5 (31.3)	8 (61.5)	4 (21)
Undergraduate	7 (43.8)	2 (15.4)	10 (52.6)
Graduate	3 (18.8)	2 (15.4)	4 (21.0)
Clinical features	CM^MOH	CM	p-value
HA days/month	21 ± 7	15 ± 5	0.005*
Headache location
Right	1 (6.3)	2 (15.4)	n/a
Left	2 (12.5)	3 (23.1)
Bilateral	13 (81.3)	8 (61.5)
Aura	10 (62.5)	5 (38.5)	n/a
Cranial autonomic symptoms	8 (50.0)	2 (15.4)	n/a
Allodynia (ASC 12) score	6 ± 3	5 ± 6	0.22
Family history (1st degree)	13 (81.3)	7 (53.8)	n/a
History of migraine (years)	21 ± 13	20 ± 13	0.371
History of CM (years)	2 ± 1	3 ± 3	0.135
Hit-6 score	67 ± 3	64 ± 4	0.03*
PHQ-9 score	10 ± 6	5 ± 3	0.006*

Values represent mean ± standard deviation or number (% of total). Significance indicated by asterisks. Abbreviations: CM: chronic migraine; CM^MOH: Chronic migraine with medication overuse headache; C: controls; BMI: body mass index; HA: moderate to severe headache days; ASC-12: allodynia symptoms checklist; HIT-6: headache impact test; PHQ-9: patient health questionnaire-9.

Statistical analysis

Using the generated ROIs (“nodes”), functional connectivity matrices were created by extracting the mean BOLD time course from each node in the salience network, and Pearson correlation (r) coefficients were calculated between each of those nodes. Next, Pearson correlation coefficient values were Fischer’s Z transformed (to convert to a normal distribution) to produce the “functional connectivity strengths” that are used in the analysis. Comparison groups were formed to investigate the relationship between the following groups: CM^MOH versus controls, CM versus controls, and CM^MOH versus CM. The intranetwork NTNC (also referred to as “edge”) strengths are defined as the correlation coefficient generated between each pair of ROIs.

When comparing each subgroup to their matched controls, we used a one tailed, two sampled t-test. When comparing between the CM subgroups (CM^MOH vs. CM), a two tailed, two sampled t-test was used. Corrected p values of <0.05 were considered significant for intranetwork NTNC strength differences (CM^MOH vs. CM). Corrected p values of <0.01 were considered significant for intranetwork NTNC strength differences (CM subgroups vs. controls). A larger threshold was used when comparing between the two CM groups, because differences may be subtle compared to the CM versus controls. All statistical tests were conducted using permutation thresholding (10,000 permutations) to control for multiple comparisons (22).

To assess associations between intranetwork functional connectivity strength and clinical parameters, Pearson correlations were conducted between the shared significant intranetwork NTNC strengths and clinical parameters, including years with migraine, years with CM, number of moderate to severe headache per month, PHQ-9, HIT-6, and allodynia symptoms checklist scores (ASC 12). Correlations were corrected for multiple comparison using the Bonferroni correction (n = 351; p = 0.00014).

Results

Participants

A total of 56 women were recruited between January 2015 and August 2016 from the University of South Carolina headache clinic (Columbia, SC). Among the CM participants (n = 33), four were excluded for motion artifact (n = 1), incidental finding on T1 (n = 2), and the presence of non-migraine related pain during the scan (n = 1). The CM group was separated into two subgroups based on the presence of medication overuse headache (MOH) as defined by the ICHD 3. Among the controls (n = 23), four were excluded for: claustrophobia (n = 2), motion artifact (n = 1), and an incidental finding on T1 (n = 1). Individual characteristics for each group are summarized in Table 1. Of the CM participants, 16 had the secondary diagnosis of MOH for the abuse of opioids (n = 8), triptans (n = 9), combined analgesics (n = 14), and combination of triptans with opioids or NSAIDS (n = 10). Numbers of mild headache days were not significantly different between CM (8.4 ± 4.8) and CM^MOH (8.5 ± 3.6). Among 29 CM participants, 15 used daily migraine preventive medications (topamax or propranolol). We did not find any significant differences between those who used preventive medications versus those who did not. Significant intranetwork resting state functional connectivity within the salience network is listed in Table 2 and illustrated in Figure 2. Significant intranetwork connections when comparing between CM subgroups are listed in Table 3 and illustrated in Figure 3. After correcting for multiple comparisons (Bonferroni), we found no correlational relationship between clinical features and intranetwork connectivity using Pearson correlation.

Table 2.

Differences of intranetwork connections in the salience network for each CM subgroup as compared to controls.

Intranetwork connection	CM Subgroup	Control	p-value
Decreased coherence in CM^MOH
Left aPFC × left superior temporal	0.0 ± 0.17	0.20 ± 0.22	0.007
Left aPFC × right superior temporal	−0.01 ± 0.18	0.18 ± 0.19	0.007
Left aPFC × left vSP	0.06 ± 0.18	0.23 ± 0.16	0.008
Right dlPFC × right SMG*	0.29 ± 0.20	0.47 ± 0.15	0.003
Right dlPFC × right vlPFC*	0.48 ± 0.23	0.73 ± 0.19	0.003
Right vlPFC × right SMA/preSMA*	0.28 ± 0.20	0.46 ± 0.15	0.006
Right vlPFC × left SMA/preSMA	0.14 ± 0.21	0.39 ± 0.22	0.004
Right dlPFC × right SMA/preSMA	0.29 ± 0.17	0.48 ± 0.19	0.006
Left dlPFC × left vSP	0.16 ± 0.15	0.34 ± 0.15	0.002
Left dlPFC × right vSP	0.12 ± 0.15	0.33 ± 0.16	<0.001
Left dlPFC × right OFI	0.20 ± 0.19	0.39 ± 0.21	0.008
Left dlPFC × dmThalamus	0.14 ± 0.19	0.30 ± 0.13	0.008
Left dlPFC × left dACC	0.57 ± 0.26	0.79 ± 0.14	0.004
Right vlPFC × right OFI	0.21 ± 0.17	0.42 ± 0.23	0.007
Right vlPFC × left SMG	0.30 ± 0.26	0.58 ± 0.29	0.006
Right vlPFC × right SMG	0.33 ± 0.21	0.62 ± 0.19	<0.001
Right vSP × left SMG	0.17 ± 0.15	0.30 ± 0.13	0.009
Left hypothalamus × right SMG	0.21 ± 0.14	0.33 ± 0.09	0.004
Right vSP × dmThalamus	0.24 ± 0.18	0.41 ± 0.18	0.009
Right vSP × right superior temporal	0.06 ± 0.19	0.26 ± 0.19	0.005
Left VTA/SN × left OFI	0.18 ± 0.19	0.36 ± 0.16	0.006
Right VTA/SN × left hypothalamus	0.11 ± 0.21	0.35 ± 0.22	0.003
Right VTA/SN × right hypothalamus	0.07 ± 0.23	0.32 ± 0.21	0.003
Right VTA/SN × left SMG	−0.01 ± 0.14	0.15 ± 0.15	0.002
Right OFI × left SMA/preSMA	0.47 ± 0.24	0.67 ± 0.14	0.008
Right vSP × left SMA/preSMA	0.23 ± 0.18	0.41 ± 0.17	0.007
Left vSP × left SMA/preSMA	0.26 ± 0.15	0.45 ± 0.17	0.002
Decreased coherence in CM
Left extended amygdala × Left aPFC	−0.03 ± 0.17	0.15 ± 0.15	0.003
Left extended amygdala × Left temporal pole	0.26 ± 0.21	0.45 ± 0.20	0.003
Left extended amygdala × Left VTA/SN	0.17 ± 0.17	0.39 ± 0.21	0.006
Left extended amygdala × right VTA/SN	0.10 ± 0.21	0.36 ± 0.15	<0.001
Left extended amygdala × right extended amygdala	0.55 ± 0.24	0.80 ± 0.23	0.007
Right extended amygdala × right temporal pole	0.29 ± 0.24	0.48 ± 0.23	0.009
Right extended amygdala × right vSP	0.41 ± 0.18	0.57 ± 0.14	0.009
Left OFI × left vSP	0.37 ± 0.13	0.51 ± 0.13	0.005
Left OFI × right vSP	0.33 ± 0.12	0.46 ± 0.12	0.005
Right vlPFC × right hypothalamus	0.20 ± 0.15	0.30 ± 0.13	0.007
Right vlPFC × right SMA/preSMA*	0.29 ± 0.18	0.46 ± 0.16	0.002
Right dACC × right OFI	0.51 ± 0.17	0.67 ± 0.21	0.002
Right dlPFC × right SMG*	0.22 ± 0.20	0.47 ± 0.16	0.001
Right dlPFC × right vlPFC*	0.47 ± 0.15	0.71 ± 0.20	0.001
Left superior temporal × left temporal pole	0.48 ± 0.26	0.71 ± 0.22	0.0099
Right superior temporal × left temporal pole	0.39 ± 0.29	0.57 ± 0.27	0.009
Right OFI × right temporal pole	0.64 ± 0.24	0.89 ± 0.25	0.002

These intranetwork connections are similarly disrupted in both CM subgroups.

Values represent the average Fischer Z transformed correlational efficiency (“Functional connectivity strength”) between the two stated nodes ± standard deviation.

Abbreviations: CM: chronic migraine; CM^MOH: Chronic migraine with medication overuse headache; dm: dorsomedial; dACC: dorsal anterior cingulate cortex; dlPFC: dorsolateral prefrontal cortex; aPFC: anterior PFC; OFI: orbital frontal insula; SMG: supramarginal gyrus; SMA: supplementary motor area; VTA/SN: ventral tegmental area/substantia nigra; vSP: ventral striatum and pallidum; vlPFC: ventrolateral PFC; × = connectivity between the pair of ROIs.

Figure 2.

Significant intranetwork connectivity difference in salience network between CM^MOH versus controls (top) and CM versus controls (bottom). Colored line represents a significant intranetwork connectivity difference (respectively) between the CM groups versus their matched controls. Green represents a unique difference to that comparison group and blue represents a shared difference in both comparison groups. Colored bars represent limbic associations: Blue = corticolimbic, green = limbic, and purple = paralimbic. Images were generated using the Matlab application “Circro” (https://github.com/bonilhamusclab/circro). Abbreviations: CM: chronic migraine; CM^MOH: chronic migraine with medication overuse headache; dm-: dorsomedial; PFC: prefrontal cortex; SMA: supplementary motor area; SN/VTA: substantia nigra/ventral tegmental area.

Table 3.

Differences of intranetwork connections when comparing CM^MOH versus CM.

	Controls (19)	CM^MOH	CM	p-value (CM + MOH vs. CM-MOH)
Decreased coherence in CM^MOH
Left aPFC × right OFI	0.33 ± 0.32	0.14 ± 0.16	0.36 ± 0.21	0.018
Left dlPFC × left vSP*	0.33 ± 0.15	0.16 ± 0.15	0.37 ± 0.16	0.009
Left dlPFC × right vSP*	0.33 ± 0.15	0.12 ± 0.15	0.32 ± 0.16	0.010
Left dlPFC × left dACC*	0.75 ± 0.17	0.57 ± 0.26	0.81 ± 0.12	0.015
Left dlPFC × right dACC	0.61 ± 0.19	0.44 ± 0.27	0.64 ± 0.14	0.036
Left vSP × left SMA/preSMA*	0.42 ± 0.18	0.26 ± 0.15	0.44 ± 0.19	0.026
Decreased coherence in CM
Right extended amygdala × left extended amygdala**	0.85 ± 0.29	0.82 ± 0.31	0.55 ± 0.24	0.044
Right VTA/SN × Paracingulate	0.10 ± 0.18	0.10 ± 0.15	0.04 ± 0.10	0.043

Intranetwork connection was found to be significant in CM^MOH versus control group.

Intranetwork connection was found to be significant in CM versus control group.

Values represent the Fischer Z transformed correlational efficiency (“Functional connectivity strength”) between the two stated nodes ± standard deviation.

Abbreviations: CM: chronic migraine; CM^MOH: chronic migraine with medication overuse headache; dACC: dorsal anterior cingulate cortex; dlPFC: dorsolateral prefrontal cortex; aPFC: anterior PFC; OFI: orbital frontal insula; SMA: supplementary motor area; VTA/SN: ventral tegmental area/substantia nigra; vSP: ventral striatum and pallidum; vlPFC: ventrolateral PFC; ×: connectivity between the pair of ROIs.

Figure 3.

Significant intranetwork connectivity difference in CM^MOH compared to CM in the salience network. Colored line represents an intranetwork connectivity difference between CM^MOH versus CM. Green represents that the intranetwork connection was lower in CM patients (CM^MOH> CM) and blue represents that the intranetwork connection was lower in CM^MOH patients (CM^MOH< CM). Colored bars represent limbic associations: Blue = corticolimbic, green = limbic, and purple = paralimbic. Abbreviations: CM: chronic migraine; CM^MOH: chronic migraine with medication overuse headache; dm-: dorsomedial; PFC: prefrontal cortex; SMA: supplementary motor area; SN/VTA: substantia nigra/ventral tegmental area.

Numerically, we did find increased intranetwork connectivity; however, it did not meet our statistical threshold. We were expecting this, based on our methods, as a “decrease” is best described as dyssynchrony between regions. In a pre-defined network, such as the salience network, if a problem exists within a network, it is more likely to see a “decrease” (or dyssynchrony) of connectivity using this method. In order to see two regions that are significantly “increased”, they would have to be afflicted in the exact same manner and therefore become more synchronized to each other. In other words, to see an “increase”, two regions would have to either be functionally firing together more, or less, at the same time as each other.

Discussion

In this study, we identified the patterns of NTNC disruptions within the salience network in the interictal state of CM and CM^MOH. We observed significant desynchronization between right dlPFC to ipsilateral vlPFC and SMG, as well as between right vlPFC to ipsilateral SMA/preSMA regions in both subgroups. Medication overuse did not have any impact on these changes. These findings suggest that the shared aberrant intranetwork connectivity may represent common efferent pathways in CM and CM^MOH. Furthermore, both CM and CM^MOH presented with disruptions in various regions of the cortico-limbic system. The involvement of the limbic system is not surprising, as it is vital in regulating emotional saliency and pain (23 –25). Conventionally, the limbic system consists of mesocortical and mesolimbic pathways. The structures comprising the limbic system overlap with the majority of regions within the salience network, such as ACC, OFI, regions of the PFC, thalamus, hypothalamus, VTA/SN, extended amygdala, PAG, and vSP (26 –29). Even though both CM subgroups revealed decreased synchronization in regions of the mesocorticolimbic system, it cannot be assumed that similar disruptions are related to the same pathophysiology. In fact, the differences identified when comparing between subgroups provided more insight into the pathogenesis behind CM and CM^MOH.

Salience network intranetwork functional connectivity in CM

One of the most intriguing findings in this study is the prominent desynchronization between the bilateral extended amygdala, and extended amygdala to ipsilateral temporal pole and OFI, in CM compared to controls. Moreover, when comparing between CM subgroups, the disruption between bilateral extended amygdala was an obvious delineator, representing a potential unique type of neuroplasticity in CM that is not seen in CM^MOH. The extended amygdala, otherwise known as the sublenticular extended amygdala or subcallosal region (as illustrated in Figure 4), is defined by its neural connections and immunohistochemical staining. This macrostructure (which includes central and medial amygdaloid nuclei, substantia innominata, the nucleus accumbens shell, and bed nuclei of the striaterminalis) plays an important role in regulating nociception, aversive motivational state, reward, memory, and learning (30 –32). In animal models, excitotoxic lesions in the central amygdaloid nucleus inhibited aversive responses (33) and administration of NMDA into the amygdala elicited cortical spreading depression (34), implicating that the amygdala may serve as an important link between migraine pain and emotional affect. Given the aberrant increase of functional activity in the amygdala during the ictal stage of episodic migraine (35), it is plausible that the increased frequency of migraine may induce excitotoxicity of the amygdala due to over-sensitization after the transition from episodic migraine to CM in susceptible patients.

Figure 4.

Anatomical location (MNI: −28, 4, −18) of left extended amygdala using MRIcron (https://www.nitrc.org/projects/mricron).

Substantial evidence in structural and functional brain imaging research supports the notion that lack of habituation may be important in migraine pathogenesis (36). In contrast to the basolateral amygdala, which has abundant excitatory neurons, the extended amygdala is rich in GABAergic neurons and endogenous opioids, such as enkaphalins, dynorphin, and endorphins (37,38). Moreover, the extended amygdala has strong interconnections with subcortical (i.e. the thalamus, hypothalamus, and brainstem) and cortical regions (i.e. the auditory, olfactory, and visual cortex) (31,32,39 –42). Given the intricate neuroanatomical and neurochemical relationship between the extended amygdala and pain processing/modulation in migraine, we postulate that the decreased functional coactivation of the bilateral extended amygdala contributes to lack of inhibition in pain processing/modulation; that is, the inability to habituate to excessive external or internal stimuli. We speculate that the bilateral extended amygdala serves as the main “hub” for habituation, and deficiencies with this area may result in the misinterpretation of salient input in CM.

Salience network intranetwork functional connectivity in CMMOH

This study revealed widespread disruptions between cortical to limbic structures in CM^MOH compared to healthy controls. This is consistent with the findings that recurrent negative reinforcement of overusing analgesic medication triggers metabolic changes in the cortico-limbic systems, which contributes to the development of addictive and compulsive traits, as well as a decreased pain threshold (43 –47). These decreased connectivity patterns were not as extensive in the CM subgroup when compared to healthy controls. It is plausible that the more prominent disruption in CM^MOH is largely attributed to the consequence of medication overuse. Addiction-related plasticity in the CNS is observed in our CM^MOH subgroup, as evident by dysfunctional connectivity involving dlPFC and aPFC, similar to what has been reported (25,43,46,48 –51). The left dlPFC appears to be a key region preferentially affected in CM^MOH, as decreased functional coactivation was seen predominantly between the left dlPFC to bilateral dACC and vSP. It is conceivable that functional desynchronization of the left dlPFC with other regions of this network contributes to lack of cognitive control in CM^MOH and is associated with increased saliency of pain fueled by the negative reinforcement of analgesic medication overuse. Taken together, we observed a “shift in paradigm” in the functional desynchronization of NTNC within the salience network when comparing CM versus CM^MOH, reflective of the reorganization in neural resource within the salience network. Specifically, the CM cohort revealed a lack of functional coactivation among mesolimbic structures centered around the bilateral amygdala, whereas similar aberrant connectivity among cortical to limbic structures centered around the left dlPFC was observed in CM^MOH.

Strengths and limitations of the study

The strength of this study is our relatively large scale systematic approach in examining intranetwork NTNC (351 pairs) within the salience network. Previous resting state fMRI studies of CM often examined focal regions of interest and their relationship with the whole brain, rather than connectivity changes. The methodology used in this project allowed us to understand how CM may affect this intrinsic network important for the evaluation and interpretation of salient information in chronic migraneurs. In this cohort, we did not include any patients who started out with severe intractable headaches with prolonged use of acute pain medications. This type of patient does exist, but was not included as part of our study. Therefore, the difference we saw in CM vs CM^MOH is suggestive of disruption of the reward/addiction pathway due to overuse of medications, essentially rewiring the brain circuits of the chronic migraineurs. We also acknowledge that there are several limitations to this study. We cannot be certain that the results are consequences or causes of CM (predisposition to CM); however, this is a common limitation in similar studies. We cannot be certain that the results are free from confounders; we did attempt to minimize any potential bias. Some CM participants (n = 15) were on daily migraine prophylaxis medications, but we do not believe this impacted our results based on our previous analysis (16). Participants were not matched exactly according to the anxiety/depression scale, therefore we cannot exclude the possibility that some of the changes in connectivity may be related to anxiety/depression. The prominent desynchronization in the bilateral amygdala observed in CM is unlikely to be attributed to mood disorders, given higher PHQ-9/HIT 6 in CM^MOH. Furthermore, using Pearson’s correlation, we did not find PHQ-9 to be associated with any of the intranetwork NTNC connections in the salience network. Additionally, differences in using acute pain medications (i.e. type, dosing) may impact the functional connectivity. Lastly, even though we did not find any correlation to clinical characteristics, this may be caused by using multiple comparison data in Pearson analysis, this is known to be associated with a higher type II error rate (false negative). A larger sample size is needed to examine the correlational relationship of clinical parameters to NTNC changes using linear regression and multiple comparisons.

In contrast to the whole brain analysis used to investigate the regional reorganization, our methods are based on a pre-defined network with the intention of using this specific network as a “standard” to examine whether the underlying changes are related to CM. We believe a network approach is more appropriate when studying a complex neurological disorder such as CM and CM^MOH. In addition, it is premature to suggest that increased connectivity (that is, rs-fMRI connectivity, not task-based connectivity), or decreased connectivity can be extrapolated to hyperexcitability or hypoexcitability states. Hopefully, new methodology in fMRI in the near future can address this issue. Future studies would benefit by utilizing Granger causality to predict directionality (52).

Footnotes

Clinical implications

We suggest that the aberrant assignment of salience to external and internal stimuli plays an important role in CM and CM^MOH interictally, mostly involving mesolimbic pathways (especially centered on the bilateral extended amygdala) in CM and prefrontal-subcortical limbic pathways in CM^MOH.

The aberrant NTNC has the potential to be used as a potential target for therapeutic intervention, or as a non-invasive marker for evaluating clinical outcomes in management of CM and CM^MOH.

Acknowledgements

The authors would like to acknowledge Taylor Hanayik (University of South Carolina), for his assistance in analysis and interpretation of data, and Charmaine Jenkins (University of South Carolina), for her technical assistance.

References

Natoli

Manack

Dean

et al.

Global prevalence of chronic migraine: A systematic review. Cephalalgia 2009; 30: 599–609.

Dodick

Loder

Manack Adams

et al.

Assessing barriers to chronic migraine consultation, diagnosis, and treatment: Results from the Chronic Migraine Epidemiology and Outcomes (CaMEO) study. Headache 2016; 56: 821–834.

Lipton

Manack Adams

Buse

et al.

A comparison of the Chronic Migraine Epidemiology and Outcomes (CaMEO) study and American Migraine Prevalence and Prevention (AMPP) study: demographics and headache-related disability. Headache 2016; 56: 1280–1289.

Katsarava

Schneeweiss

Kurth

et al.

Incidence and predictors for chronicity of headache in patients with episodic migraine. Neurology 2004; 62: 788–790.

May

Schulte

. Chronic migraine: Risk factors, mechanisms and treatment. Nature Rev Neurol 2016; 12: 455–464.

Zappaterra

Guerzoni

Cainazzo

et al.

Basal cutaneous pain threshold in headache patients. J Headache Pain 2011; 12: 303–310.

Ferraro

Grazzi

Muffatti

et al.

In medication-overuse headache, fMRI shows long-lasting dysfunction in midbrain areas. Headache 2012; 52: 1520–1534.

Lai

Chou

Fuh

et al.

Gray matter changes related to medication overuse in patients with chronic migraine. Cephalalgia 2016; 36: 1324–1333.

Ferraro

Grazzi

Mandelli

et al.

Pain processing in medication overuse headache: A functional magnetic resonance imaging (fMRI) study. Pain Med 2012; 13: 255–262.

10.

Chen

Liu

et al.

Altered functional connectivity of the marginal division in migraine: A resting-state fMRI study. J Headache Pain 2016; 17: 89–89.

11.

Chen

Liu

et al.

Altered functional connectivity of amygdala underlying the neuromechanism of migraine pathogenesis. J Headache Pain 2017; 18: 7–7.

12.

Seeley

Menon

Schatzberg

et al.

Dissociable intrinsic connectivity networks for salience processing and executive control. J Neurosci 2007; 27: 2349–2356.

13.

Borsook

Edwards

Elman

et al.

Pain and analgesia: The value of salience circuits. Progress Neurobiol 2013; 104: 93–105.

14.

Kucyi

Davis

. The dynamic pain connectome. Trends Neurosci 2015; 38: 86–95.

15.

Menon V. Salience Network. In: Toga A (ed.) Brain mapping: An encyclopedic reference, vol 2. Academic Press: Elsevier, 2015, pp.597–611.

16.

Androulakis M, Krebs K, Peterlin B, et al. Modulation of intrinsic resting state fMRI networks in women with chronic migraine. Neurology 2017; 89: 163–169.

17.

Headache Classification Committee of the International Headache Society (IHS). The international classification of headache disorders (beta version). Cephalalgia 2013; 33: 629–808.

18.

Lipton RB, Bigal ME, Ashina S, et al. Cutaneous allodynia in the migraine population. Ann Neurol 2008; 63: 148–158.

19.

Kosinski

Bayliss

Bjorner

et al.

A six-item short-form survey for measuring headache impact: The HIT-6™. Qual Life Res 2003; 12: 963–974.

20.

Kroenke

Spitzer

Williams

. The PHQ-9: Validity of a brief depression severity measure. J Gen Intern Med 2001; 16: 606–613.

21.

Freedman

Lane

. A nonstochastic interpretation of reported significance levels. J Business Econ Stat 1983; 1: 292–298.

22.

Winkler

Ridgway

Webster

et al.

Permutation inference for the general linear model. Neuroimage 2014; 92: 381–397.

23.

Houk

Davis

Beiser

. Models of information processing in the basal ganglia, Cambridge (MA): MIT Press, 1995.

24.

Graybiel

. The basal ganglia and chunking of action repertoires. Neurobiol Learning Memory 1998; 70: 119–136.

25.

Gottfried JA (ed). Neurobiology of sensation and reward. Boca Raton (FL): CRC Press/Taylor & Francis, 2011.

26.

Everitt

Hutcheson

Ersche

et al.

The orbital prefrontal cortex and drug addiction in laboratory animals and humans. Ann N.Y. Acad Sci 2007; 1121: 576–597.

27.

Draganski

Kherif

Klöppel

et al.

Evidence for segregated and integrative connectivity patterns in the human basal ganglia. J Neurosci 2008; 28: 7143–7152.

28.

Wise

. Roles for nigrostriatal – not just mesocorticolimbic – dopamine in reward and addiction. Trends Neurosci 2009; 32: 517–524.

29.

Bowers

Chen

Bonci

. AMPA receptor synaptic plasticity induced by psychostimulants: The past, present, and therapeutic future. Neuron 2010; 67: 11–24.

30.

Heimer

. The human brain and spinal cord 1995; Vol 2, New York: Springer.

31.

Sah

Faber

De Armentia

et al.

The amygdaloid complex: Anatomy and physiology. Physiol Rev 2003; 83: 803–834.

32.

Habel U and Schneider F. Amygdala, functional imaging. In: Gebhart GF and Schmidt RF (eds) Encyclopedia of pain. Berlin Heidelberg: Springer, 2013, pp.118–124.

33.

Tanimoto

Nakagawa

Yamauchi

et al.

Differential contributions of the basolateral and central nuclei of the amygdala in the negative affective component of chemical somatic and visceral pains in rats. Eur J Neurosci 2003; 18: 2343–2350.

34.

Akcali

Sayin

Sara

et al.

Does single cortical spreading depression elicit pain behaviour in freely moving rats?

Cephalalgia 2010; 30: 1195–1206.

35.

Stankewitz

May

. Increased limbic and brainstem activity during migraine attacks following olfactory stimulation. Neurology 2011; 77: 476–482.

36.

Burstein

Noseda

Borsook

. Migraine: Multiple processes, complex pathophysiology. J Neurosci 2015; 35: 6619–6629.

37.

Koepp

Hammers

Lawrence

et al.

Evidence for endogenous opioid release in the amygdala during positive emotion. Neuroimage 2009; 44: 252–256.

38.

Veinante

Yalcin

Barrot

. The amygdala between sensation and affect: A role in pain. J Mol Psych 2013; 1: 10–1186.

39.

Cassell

Freedman

Shi

. The intrinsic organization of the central extended amygdala. Ann N.Y. Acad Sci 1999; 877: 217–241.

40.

Alheid GF. Extended amygdala. In: Binder MD, Hirokawa N and Windhorst U (eds) Encyclopedia of neuroscience. Berlin Heidelberg: Springer, 2009, pp.1501–1506.

41.

Standring S (ed). Gray’s anatomy: The anatomical basis of clinical practice. Philadelphia (PA): Elsevier Health Sciences, 2015.

42.

Kamali

Sair

Blitz

et al.

Revealing the ventral amygdalofugal pathway of the human limbic system using high spatial resolution diffusion tensor tractography. Brain Struct Function 2016; 221: 3561–3569.

43.

Asensio

Romero

et al.

Striatal dopamine D2 receptor availability predicts the thalamic and medial prefrontal responses to reward in cocaine abusers three years later. Synapse 2010; 64: 397–402.

44.

Lundqvist

Aaseth

Grande

et al.

The severity of dependence score correlates with medication overuse in persons with secondary chronic headaches. The Akershus study of chronic headache. Pain 2010; 148: 487–491.

45.

Sances

Galli

Anastasi

et al.

Medication-overuse headache and personality: A controlled study by means of the MMPI-2. Headache 2010; 50: 198–209.

46.

Lüscher

Malenka

. Drug-evoked synaptic plasticity in addiction: From molecular changes to circuit remodeling. Neuron 2011; 69: 650–663.

47.

Navratilova

Xie

Okun

et al.

Pain relief produces negative reinforcement through activation of mesolimbic reward–valuation circuitry. Proc Nat Acad Sci 2012; 109: 20709–20713.

48.

Goldstein

Volkow

. Drug addiction and its underlying neurobiological basis: Neuroimaging evidence for the involvement of the frontal cortex. Am J Psych 2002; 159: 1642–1652.

49.

Goldstein

Leskovjan

Hoff

et al.

Severity of neuropsychological impairment in cocaine and alcohol addiction: Association with metabolism in the prefrontal cortex. Neuropsychologia 2004; 42: 1447–1458.

50.

Goldstein

Alia-Klein

Tomasi

et al.

Is decreased prefrontal cortical sensitivity to monetary reward associated with impaired motivation and self-control in cocaine addiction?

Am J Psych 2007; 164: 43–51.

51.

Yang

Xie

Shao

et al.

Dynamic neural responses to cue-reactivity paradigms in heroin-dependent users: An fMRI study. Hum Brain Map 2009; 30: 766–775.

52.

Bressler

Menon

. Large-scale brain networks in cognition: Emerging methods and principles. Trends Cognitive Sci 2010; 14: 277–290.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.02 MB

Modulation of salience network intranetwork resting state functional connectivity in women with chronic migraine

Abstract

Objective

Methods

Results

Conclusion

Keywords

Introduction

Methods

Participants

HIT-6 and PHQ-9 scale

Standard protocol approvals, registrations, and patient consents

MR imaging

MRI preprocessing

Statistical analysis

Results

Participants

Discussion

Salience network intranetwork functional connectivity in CM

Salience network intranetwork functional connectivity in CMMOH

Strengths and limitations of the study

Footnotes

Clinical implications

Acknowledgements

References

Supplementary Material