Sage Journals: Discover world-class research

Abstract

Background

Tension-type headache (TTH) is the most common primary headache, yet its neural basis remains unclear. Magnetoencephalography (MEG) combined with graph-theoretical analysis enables precise mapping of functional brain networks. This study aimed to identify network-level connectivity alterations in TTH using resting-state MEG and graph-based metrics.

Methods

We analyzed resting-state MEG data from 27 TTH patients during the interictal period and 37 age- and gender-matched healthy controls, all with eyes closed. Functional connectivity (FC) across 1–45 Hz was mapped and analyzed using graph theory. Network topology metrics were computed, and their associations with clinical symptoms were assessed.

Results

TTH patients showed increased FC across 1–45 Hz, notably between the right somatomotor and frontal operculum and insula (FrOperIns), and between the temporo-occipital-parietal (TempOccPar) and visual regions, with the latter positively correlated with Headache Impact Test-6 scores. Frequency-specific increases were observed between the left prefrontal and right orbitofrontal cortices (delta, theta), somatomotor and FrOperIns (theta), and TempOccPar and visual areas (beta). Graph theory analysis revealed nodal abnormalities, particularly in the left precuneus and posterior cingulate and prefrontal cortices, along with elevated local efficiency and clustering coefficient.

Conclusions

These findings indicate that TTH is associated with frequency-specific alterations in functional connectivity and disrupted network topology, particularly involving regions implicated in pain processing and cognitive control. Graph-theoretical MEG analysis may offer valuable insights into the neural mechanisms of TTH and support the development of network-based biomarkers.

Trial Registration: ClinicalTrials.gov Identifier: NCT05334927

Graphical abstract

This is a visual representation of the abstract.

Keywords

functional connectivity graph theory magnetoencephalography tension-type headache

Introduction

Tension-type headache (TTH) is the most prevalent primary headache disorder, affecting more than two billion people worldwide (1,2). Despite its high prevalence and substantial burden on quality of life, the neurobiological mechanisms underlying TTH remain poorly defined. Unlike migraine, which has been extensively studied using neuroimaging techniques, TTH has historically been regarded as lacking specific central nervous system abnormalities. However, recent findings suggest that subtle changes in brain network organization may contribute to its pathophysiology (3,4).

Functional connectivity (FC) analysis has emerged as a powerful tool to investigate large-scale brain network organization in both healthy and clinical populations (5,6). Altered FC has been consistently reported in chronic pain conditions, including migraine, fibromyalgia and back pain, and may reflect disrupted sensory integration, emotional regulation or descending pain control (7). Magnetoencephalography (MEG) offers a unique advantage for investigating resting-state FC due to its millisecond temporal resolution and ability to capture neural oscillations in source space. Unlike functional magnetic resonance imaging (fMRI), which measures slow hemodynamic responses, MEG enables direct analysis of neural dynamics across distinct frequency bands, each of which is associated with specific cognitive and sensory processes (6,8,9). Previous MEG studies have demonstrated altered FC in various primary headache, including migraine (10,11) and new daily persistent headache (6), and suggest that headache chronification may be associated with disrupted neural integration across cortical regions. Frequency-specific FC analysis is particularly relevant in the context of pain research (12,13). Prior studies have shown that alpha, beta and gamma oscillations are differentially involved in sensory discrimination, attentional control and affective evaluation of pain (14,15). However, MEG-based studies of FC in TTH remain scarce, and few have explored the frequency-specific functional architecture of TTH-related brain networks.

Graph theory provides a powerful framework to quantify the topological organization of brain networks, treating brain regions as nodes and functional connections as edges. Key metrics (such as local and global efficiency, clustering coefficient and small worldness) capture the balance between network integration and segregation. These properties are particularly relevant for understanding how chronic pain may alter large-scale neural communication (16). Altered network topology has been reported in several psychiatric and neurological conditions (17), including migraine (18), although their characterization in TTH is limited. Most previous studies have either focused on structural changes or lacked fine-grained analyses of functional network connectivity, leaving a substantial gap in our understanding of how brain networks are altered in TTH.

In the present study, we aimed to address these gaps by investigating resting-state FC and network topology in TTH using MEG and graph-theoretical analysis. By leveraging the high temporal resolution of MEG and analyzing multiple frequency bands, we sought to capture fine-grained alterations in network connectivity. We hypothesized that patients with TTH would show frequency-specific disruptions in functional brain networks, particularly in regions involved in pain perception and cognitive control, and that these alterations would correlate with clinical symptom severity. This study represents one of the first attempts to combine MEG with graph theory to systematically characterize network-level dysfunction in TTH, providing novel insights into its neurophysiological basis.

Methods

The analysis pipeline from MEG signal acquisition to functional brain network construction is presented in Figure 1. Resting-state MEG signals were acquired and preprocessed, then filtered into five frequency bands: delta (1–4 Hz), theta (4–8 Hz), alpha (8–13 Hz), beta (13–30 Hz) and gamma (30–45 Hz). Structural MRI images were co-registered with MEG data to enable source localization. Source-level functional connectivity networks were then constructed. Graph-theoretical metrics, including node attributes and global parameters, were computed across a range of sparsity thresholds. Group differences in connectivity and network topology were statistically assessed and correlations with clinical measures were examined.

Figure 1.

The analysis pipeline from magnetoencephalography (MEG) signal acquisition to functional brain network construction.

Participants

This cross-sectional observational study included 64 participants, comprising 27 patients with TTH and 37 age- and gender-matched healthy controls (HCs), who were consecutively recruited from the headache outpatient clinic at Beijing Tiantan Hospital, Capital Medical University, between October 2020 and August 2023. Inclusion criteria for TTH patients were: (i) diagnosis consistent with the International Classification of Headache Disorders, 3rd edition (ICHD-3) (2); (ii) no contraindications for MRI examination; and (iii) no preventive medication use for at least three months, with no use of acute pain medication on the day prior to scanning. General exclusion criteria for all participants included: (i) comorbid primary headache or other pain disorders; (ii) pregnancy or breastfeeding; (iii) presence of neurological, cardiovascular, cerebrovascular or endocrine disorders; (iv) history of drug or alcohol abuse; and (v) poor MRI data quality (e.g. severe susceptibility artifacts or incomplete data). All TTH patients underwent MRI and MEG scanning during the interictal period, defined as at least 24 hours after the last headache episode, and confirmed by both neurologists and radiologists. Clinical evaluations included the Headache Impact Test-6 (HIT-6; a validated questionnaire assessing the impact of headaches on daily functioning, with higher scores indicating greater disability) (19), Patient Health Questionnaire-9 (PHQ-9)(20), Pittsburgh Sleep Quality Index (PSQI)(21), Visual Analogue Scale (VAS; the maximum pain level experienced in the past month without taking analgesics) and Montreal Cognitive Assessment (MoCA)(22). These scales were used to assess headache-related disability, depressive symptoms, sleep quality, pain intensity and cognitive function, respectively.

This study was approved by the local ethics committee of Beijing Tiantan Hospital (Approval No. KY2022-044), as a sub-study of the ongoing China HeadAche DIsorders RegiStry Study (CHAIRS; ClinicalTrials.gov Identifier: NCT05334927). All participants provided written informed consent in accordance with the Declaration of Helsinki.

MRI data acquisition

Structural MRI scans were acquired using a 3.0 Tesla scanner (GE Healthcare, Milwaukee, WI, USA) at the Nuclear Medicine Department of Beijing Tiantan Hospital. All scans were performed by an experienced neuroradiologist who was blinded to participant group assignments. During the scan, participants were instructed to remain awake, keep their eyes closed, and minimize head and neck movement. Noise-reduction and motion-minimization measures were employed. High-resolution T1-weighted anatomical images were obtained using a 3D BRAVO sequence (coronal acquisition; field of view = 256 mm; matrix = 256 × 256; 192 slices; flip angle = 15°; repetition time = 850 ms; echo time = 320 ms; voxel size = 1 × 1 × 1.5 mm³).

MEG data acquisition and preprocessing

Resting-state neuromagnetic activity was recorded using a 306-channel Elekta Neuromag system (102 magnetometers and 204 planar gradiometers) (Elekta, Stockholm, Sweden). Head position was tracked continuously using four head-position indicator (HPI) coils, and recordings were repeated if excessive movement was detected. If sleep occurs, recording will restart. Head shape was digitized using the Fastrak® system (Polhemus, Colchedster, VT, USA), with approximately 300 scalp points marked at the nasion and preauricular points, and across the scalp for co-registration with MRI. Data were sampled at 2000 Hz with a low-pass filter at 660 Hz. Electrooculogram (EOG) and electrocardiogram (ECG) were simultaneously recorded. Participants were instructed to remain awake, relaxed, and motionless with eyes closed during the five-minutes resting-state MEG recording.

Preprocessing began with spatiotemporal signal space separation using Elekta's MaxFilter to suppress environmental noise and remove bad channels (23). Head motion correction was applied based on continuous HPI tracking. Subsequent preprocessing was conducted using the OHBA Software Library (OSL) (https://github.com/OHBA-analysis). Data were downsampled to 500 Hz, and notch filters were applied at 50 Hz and its harmonics to remove power line artifacts. A bandpass filter from 0.5 to 250 Hz was applied to reduce computational load. Noisy segments and channels were automatically identified and removed. Independent component analysis was used to estimate 60 components, and components associated with ocular and cardiac artifacts were identified via correlation with EOG and ECG signals and manually confirmed for removal. Following artifact removal, data were filtered to 1–45 Hz. Individual T1-weighted MRI images were co-registered to MEG data via an iterative closest-point algorithm aligning digitized scalp points to surfaces extracted using FSL's BET tool (24). Source reconstruction was performed using a linearly constrained minimum variance beamformer with a single-shell head model on an 8-mm isotropic grid (25,26). The reconstructed source space was parcellated into 100 anatomical regions of interest (ROI), and regional time series were extracted using a spatial basis projection method (27), which constrains dipole sources to the cortical surface and represents them as linear combinations of spatial basis functions aligned with the local cortical geometry (see supplementary material, Table S1). This approach reduces the dimensionality of the inverse problem and improves spatial accuracy in source estimation. To reduce spurious correlations caused by source leakage, symmetric multivariate leakage correction was applied (28). This method uses a symmetric orthogonalization procedure to remove zero-lag correlations across all ROI time series, producing mutually orthogonal signals at the same time as preserving their original structure. This correction minimizes spurious functional connectivity caused by volume conduction and is particularly suitable for multivariate analyses in source-space MEG data. Given the arbitrary sign of source-reconstructed time series, a sign-flipping algorithm was employed to ensure consistency across participants by maximizing sign agreement of pairwise covariance matrices.

Functional connectivity construction and network analysis

FC was estimated by computing power envelope correlations between orthogonalized source-level signals. Cortical sources were parcellated into 100 regions based on the Schaefer atlas (29), with each region treated as a network node for subsequent functional connectivity and graph-theoretical analyses. FC between each pair of regions defined the edges of the network. This approach has demonstrated high reproducibility for resting-state MEG connectivity estimation (30). To identify group-level differences in FC, a network-based statistic (NBS) was performed. To investigate frequency-specific network dynamics, band-pass filtering was applied to the source signals in five canonical frequency bands: delta (1–4 Hz), theta (4–8 Hz), alpha (8–13 Hz), beta (13–30 Hz) and gamma (30–45 Hz). FC matrices were computed separately for each band and subject. Graph-theoretical metrics were calculated to assess both nodal and global topological properties of the functional network. Network analyses were conducted across a sparsity threshold range of 0.05 to 0.40 (step size = 0.01), which was selected based on previous studies to ensure that the brain networks remained fully connected and biologically plausible across subjects (31 –33). An example is shown in the supplementary material (Figure S1). To reduce the influence of any single threshold, the area under the curve (AUC) for each graph metric was calculated over the entire sparsity range, enabling threshold-independent comparisons between groups. Group-level differences in AUC values were statistically assessed using two-sample t-tests (34). All network construction and analysis were performed using the GRETNA toolbox (http://www.nitrc.org/projects/gretna) and visualization was completed using BrainNet Viewer (http://www.nitrc.org/projects/bnv).

Statistical analysis

The sample size (n = 64; 27 TTH patients and 37 HCs) was determined based on data availability and prior literature (12), providing 90% power to detect group differences at a two-sided alpha of 0.05, assuming no negative correlation between endpoints. Statistical analyses were performed using SPSS, version 27.0 (IBM Corp., Armonk, NY, USA). Normality of continuous variables was tested using the Shapiro–-Wilk test. Data following a normal distribution were reported as the mean ± SD; otherwise, as median and interquartile range. Categorical data were expressed as counts (percentages). Continuous variables with normal distribution were compared using independent-sample t-tests, while non-normally distributed variables were analyzed using the Mann–Whitney U-test. Categorical variables were compared using the chi-square test. Correlations between network parameters and clinical or demographic variables were evaluated using Pearson correlation, with false discovery rate (FDR) correction applied for multiple comparisons. In both NBS and graph-theoretical analyses, age and gender were included as covariates. All tests were two-tailed, with significance defined as p < 0.05.

Results

Clinical characteristics and demographics

In total, 32 patients with TTH and 43 HCs were initially enrolled. Three patients and three HCs were excluded due to poor data quality or segmentation errors, and an additional two patients and three HCs were excluded due to incomplete data (Figure 2). The final sample included 27 TTH patients (mean ± SD age: 41.96 ± 12.27 years; 14 females; headache frequency: 30 days/month (interquartile range = 20–30)) and 37 HCs (mean ± SD age: 39.15 ± 15.30 years; 23 females). There were no significant differences in age or gender between the two groups (p = 0.127 and 0.409, respectively). Detailed clinical characteristics of the TTH group are presented in Table 1.

Figure 2.

Illustration of the recruitment and exclusion process. HC = healthy control; TTH = tension-type headache.

Table 1.

Demographic and clinical data.

	Controls (n = 37)	TTH (n = 27)	p value
Age (years)	39.15 ± 15.30	41.96 ± 12.27	0.127
Sex (male/female)	14/23 (37.8/62.2%)	13/14 (48.1/51.9%)	0.409
Nausea, vomiting, n (%)	NA	0 (0.0%)	NA
Photophobia, n (%)	NA	8 (29.6%)	NA
Phonophobia, n (%)	NA	12 (44.4%)	NA
Disease duration (years)	NA	5.00 (1.00–20.00)	NA
Headache frequency (days/month)	NA	30 (20–30)	NA
VAS score (0–10)	NA	6.42 ± 2.02	NA
MoCA score (0–30)	NA	26.17 ± 3.92	NA
HIT-6 score (36–78)	NA	63.00 ± 2.53	NA
PHQ-9 score (0–27)	NA	8.00 ± 7.13	NA
GAD-7 score (0–21)	NA	6.50 ± 5.96	NA
PSQI score (0–21)	NA	9.00 ± 4.82	NA
MIDAS score	NA	101.17 ± 83.71	NA

Continuous variables are presented as the mean ± SD or as median (interquartile range (IQR)). Categorical data are expressed as counts (percentages).

Abbreviations: TTH = tension-type headache; NA = not applicable; VAS = visual analogue scale (the maximum headache intensity during the past month without medication); MoCA = Montreal Cognitive Assessment; HIT-6 = Headache Impact Test-6; PHQ-9 = Patient Health Questionnaire-9; GAD-7 = Generalized Anxiety Disorder-7; PSQI = Pittsburgh Sleep Quality Index; MIDAS = Migraine Disability Assessment Questionnaire

NBS analysis and clinical correlations

NBS analysis revealed significantly increased FC in TTH patients compared to HCs. Specifically, increased FC was observed between the right somatomotor and right frontal operculum and insula (FrOperIns) across the whole bands (t = 4.849, p = 9.17 × 10⁻⁶) and in the theta band (t = 5.421, p = 1.11 × 10⁻⁶). Increased FC was also found between the right temporo-occipital-parietal (TempOccPar) and right visual regions in the whole bands (t = 4.982, p = 5.65 × 10⁻⁶) and beta band (t = 5.042, p = 4.54 × 10⁻⁶). In the delta (t = 6.009, p = 1.19 × 10⁻⁷) and theta bands (t = 5.388, p = 1.26 × 10⁻⁶), increased FC was observed between the left prefrontal cortex and right orbitofrontal cortex (OFC) (Figure 3 and Table 2).

Figure 3.

Network-based statistic analysis and clinical correlations. (A) In patients with tension-type headache (TTH), increased functional connectivity (FC) was observed between the right somatomotor cortex and the right frontal operculum/insula (FrOperIns), as well as between the right temporo-occipital-parietal (TempOccPar) and right visual regions across the whole band. (B) In the delta band, increased FC was observed between the left prefrontal cortex and right orbitofrontal cortex (OFC) in the TTH patient. (C) In the theta band, increased FC was observed between the right somatomotor cortex and the right FrOperIns, as well as between the left prefrontal cortex and right OFC in the TTH patient. (D) In the beta band, increased FC was observed between the right TempOccPar and right visual regions in the TTH patient. (E) In the whole band, FC between the right TempOccPar and right visual regions was positively associated with Headache Impact Test-6 (HIT-6) score. (F) In the beta band, FC between the right TempOccPar and right visual regions was positively associated with HIT-6 score.

Table 2.

The differences of functional connection between the control group and TTH group.

Frequency	Brain region of functional connection	t (TTH-HC)	p
Whole (1–45 Hz)	R_somatomotor - R_frontal operculum insula	4.849	9.17 × 10⁻⁶
	R_temporal occipital parietal - R_visual	4.982	5.65 × 10⁻⁶
Delta (1–4 Hz)	L_prefrontal cortex - R_orbital frontal cortex	6.009	1.19E × 10⁻⁷
Theta (4–8 Hz)	L_prefrontal cortex - R_orbital frontal cortex	5.388	1.26 × 10⁻⁶
	R_somatomotor - R_frontal operculum insula	5.421	1.11 × 10⁻⁶
Beta (13–30 Hz)	R_temporal occipital parietal - R_visual	5.042	4.54 × 10⁻⁶

Abbreviations: TTH = tension-type headache; HC = healthy controls; R = right. L = left.

Correlation analysis further demonstrated that FC between the right TempOccPar and right visual regions was positively associated with headache impact (HIT-6 score) across the whole bands (Pearson, r = 0.853; FDR, q = 0.018) and in the beta band (Pearson, r = 0.818; FDR, q = 0.024) (Figure 3). No significant correlations were found between other FC metrics and clinical features in the TTH group.

Graph theory analysis

Graph-theoretical analysis revealed significant differences in network topology between TTH patients and HCs.

As shown in Table 3, multiple nodal attributes exhibited altered across groups. Compared to controls, TTH patients showed significantly lower betweenness centrality variance in the left precuneus and posterior cingulate cortex (pCunPCC) and right medial cortex (p = 2.54 × 10⁻⁴ and 2.22 × 10⁻⁴, respectively). In contrast, degree centrality variance in the left pCunPCC was significantly higher in the TTH group (p = 1.44 × 10⁻⁴). The TTH group also exhibited increased nodal clustering coefficient variance in the right visual and right temporal cortices (p = 1.55 × 10⁻⁵ and 1.77 × 10⁻⁴, respectively), as well as higher nodal efficiency variance in the left prefrontal cortex and left pCunPCC (p = 4.87 × 10⁻⁴ and 9.12 × 10⁻⁵, respectively). Nodal local efficiency variance was elevated in the right visual cortex (p = 4.17 × 10⁻⁴). In addition, nodal shortest path variance was reduced in the left prefrontal cortex and left pCunPCC in TTH patients compared to controls (p = 3.96 × 10⁻⁴ and 1.49 × 10⁻⁵, respectively).

Table 3.

The differences of nodal attributes between the control group and TTH group.

Brain region	Controls group	TTH group	p _var	p _var
Betweenness centrality
Left precuneus posterior cingulate cortex	35.093 ± 13.556	22.699 ± 7.673	−3.889	2.54 × 10⁻⁴
Right medial	45.300 ± 17.451	28.337 ± 7.164	−3.931	2.22 × 10⁻⁴
Degree centrality
Left precuneus posterior cingulate cortex	9.973 ± 2.293	12.136 ± 1.521	−4.060	1.44 × 10⁻⁴
Nodal clustering coefficient
Right visual	0.132 ± 0.022	0.144 ± 0.018	−4.703	1.55 × 10⁻⁵
Right temporal	0.154 ± 0.024	0.155 ± 0.016	−3.999	1.77 × 10⁻⁴
Nodal efficiency
Left prefrontal cortex	0.173 ± 0.012	0.200 ± 0.011	3.689	4.87 × 10⁻⁴
Left precuneus posterior cingulate cortex	0.165 ± 0.011	0.208 ± 0.009	−4.196	9.12 × 10⁻⁵
Nodal local efficiency
Right visual	0.198 ± 0.026	0.217 ± 0.023	−3.737	4.17 × 10⁻⁴
Nodal shortest path
Left prefrontal cortex	0.533 ± 0.039	0.458 ± 0.026	−3.753	3.96 × 10⁻⁴
Left precuneus posterior cingulate cortex	0.560 ± 0.037	0.441 ± 0.020	4.051	1.49 × 10⁻⁴

Note: Values are the mean ± SD. Abbreviations: TTH = tension-type headache. “Left precuneus–posterior cingulate cortex” refers to a single Schaefer atlas parcel encompassing both the medial precuneus and posterior cingulate regions.

In the analysis of global network topology, the TTH group exhibited significantly higher local efficiency compared to the HC group (p < 0.01), while global efficiency showed no significant difference between the two groups (p > 0.05). Regarding small-world attributes, gamma and sigma values were significantly increased in the TTH group relative to HCs (both p < 0.05), whereas lambda did not differ significantly between groups (p > 0.05). Additionally, the clustering coefficient was significantly higher in the TTH group (p < 0.05), while the shortest path length remained comparable between groups (p > 0.05) (Figure 4). No significant correlations were found between graph-theoretical metrics and clinical variables in the TTH group.

Figure 4.

Results of global network topology. (A) The tension-type headache (TTH) group exhibited significantly higher local efficiency compared to the healthy (HC) group, while global efficiency showed no significant difference between the two groups. (B) Regarding small-world attributes, gamma and sigma values were significantly increased in the TTH group relative to HCs, whereas lambda did not differ significantly between groups. Additionally, the clustering coefficient was significantly higher in the TTH group, while shortest path length remained comparable between groups. *p < 0.05, **p < 0.01. ns, no significant.

Discussion

This study revealed frequency-specific alterations in functional connectivity and network topology in TTH patients using resting-state MEG and graph theory. Increased FC was observed between somatomotor and frontal operculum/insula regions, as well as between temporo-occipital-parietal and visual areas, with the latter positively correlated with HIT-6 scores. Additional FC increases were found between the left prefrontal and right OFC. Graph analysis showed disrupted nodal properties, particularly in the left prefrontal and precuneus/posterior cingulate cortices, characterized by increased degree, efficiency and clustering, along with reduced path length. At the global level, TTH patients exhibited higher local efficiency, clustering and small-worldness, suggesting enhanced local integration and altered network organization.

Increased FC between the right somatomotor cortex and right FrOperIns across multiple frequency bands in TTH patients may reflect heightened interoceptive and sensorimotor integration. The insula, particularly its anterior portion, plays a critical role in pain perception, emotional salience and interoception (35) and has been repeatedly implicated in migraine and chronic pain conditions (36,37). Heightened connectivity in this region may indicate altered pain processing or increased vigilance to somatic sensations in TTH. Although only 29.6% of TTH patients reported photophobia, we observed significantly increased FC between temporo-occipital-parietal and visual regions, which correlated with HIT-6 scores. This suggests that visual network alterations may contribute to headache burden even without overt photophobia. Previous studies reported normal or subtly altered brain excitability in TTH, particularly in the somatosensory cortex, and emphasized differences from migraine (38 –40). Our findings differ in that we used MEG-based network analysis and identified frequency-specific visual system changes. Importantly, we excluded patients with migraine or migrainous features based on ICHD-3 criteria, minimizing diagnostic overlap. Therefore, the observed visual hyperconnectivity likely reflects intrinsic network reorganization in TTH rather than comorbid migraine. Although TTH and migraine may share some sensory processing abnormalities, our results support distinct, modality-specific alterations in TTH. Furthermore, increased FC between the left prefrontal cortex and right OFC in the delta and theta bands suggests aberrant fronto-orbitofrontal communication, which may relate to cognitive-affective dysregulation in chronic headache. The OFC is a key node in the emotional and reward network (41), involved in evaluating the affective component of pain. Prior studies have shown altered OFC activity in both depression and pain (42,43), implicating it in the chronicity of discomfort and reduced hedonic tone.

Although classical pain-processing hubs (e.g. anterior cingulate cortex, thalamus) were not prominently identified, this may be due to the resting-state nature of our MEG paradigm, which reflects spontaneous network dynamics rather than evoked pain responses. Altered connectivity in sensory and integrative regions such as the insula and temporoparietal cortex may still reflect dysfunctional pain modulation in TTH. Thus, although the absence of strong activation in core pain areas may limit anatomical specificity, our findings provide meaningful insight into large-scale network dysfunction in chronic headache. Notably, TTH patients exhibited right-dominant functional connectivity, particularly involving the somatomotor, insular, and temporo-occipital-parietal regions. This lateralization may reflect the preferential role of the right hemisphere in pain processing, interoceptive awareness and attentional modulation of sensory input (44 –46). Prior studies have highlighted the involvement of the right insula in salience detection and affective appraisal of pain (35). Thus, the observed asymmetry may indicate right-lateralized maladaptive plasticity in TTH, potentially linked to altered sensory integration or compensatory pain modulation mechanisms. The observed frequency-specific alterations in TTH may reflect distinct aspects of pain pathophysiology. The theta- and delta-band increases in FC between the left prefrontal cortex and right OFC may reflect disrupted affective or cognitive pain modulation. Beta-band hyperconnectivity between the right TempOccPar and visual regions may indicate abnormal sensory gain or cortical excitability, potentially contributing to visual discomfort or heightened sensory load. Theta-band increases in the somatomotor-FrOperIns circuit may reflect enhanced interoceptive or somatosensory integration. These oscillatory disruptions align with known neurophysiological mechanisms of chronic pain and may have clinical relevance in characterizing symptom profiles or guiding targeted neuromodulatory interventions.

Graph-theoretical analysis revealed widespread alterations in brain network topology in TTH patients, highlighting disrupted integration and segregation in pain-related regions. Notably, nodal metrics within the left pCunPCC showed significant abnormalities, with reduced betweenness centrality but increased degree and efficiency. As core hubs of the default mode network, the precuneus and PCC are critical for internally directed attention and self-referential processing (47). Alterations in these regions have been repeatedly linked to chronic pain and attentional dysregulation (48,49). Increased degree and efficiency in TTH may reflect compensatory hyperconnectivity, while reduced betweenness centrality suggests diminished control over information flow across distributed systems (50). Furthermore, increased clustering coefficient and local efficiency in the right visual and temporal cortices suggest enhanced regional segregation and potential hypersynchrony in sensory areas. Similar changes have been observed in migraine (51), and are thought to reflect altered sensory gating and heightened visual responsiveness.

At the global level, the TTH group exhibited significantly higher local efficiency and small-worldness (gamma and sigma), indicating a shift toward enhanced local information processing at the expense of global integration (52,53). Such a shift is commonly observed in chronic pain conditions and may underlie difficulties in switching between cognitive states or integrating multisensory information (54,55). The absence of group differences in global efficiency and path length further suggests that the core large-scale communication infrastructure is preserved, but its functional dynamics are altered. Taken together, these network-level changes imply that TTH is associated with reorganization of key topological properties, particularly in default mode, sensory, and prefrontal areas, that may reflect maladaptive neuroplasticity in response to sustained nociceptive input.

Several limitations should be considered when interpreting the results of this study. First, the relatively modest sample size and cross-sectional design represent key limitations of this study. The former may restrict the generalizability of our findings, while the latter precludes causal inference regarding the relationship between functional connectivity alterations and headache chronicity. Future longitudinal and multicenter studies are needed to validate and extend these observations. Second, although we carefully controlled for confounding factors such as age and gender, other unmeasured variables (e.g. medication use, comorbid psychiatric symptoms) may have influenced brain connectivity patterns. Third, the use of a fixed parcellation scheme (29) and MEG source reconstruction algorithms may introduce methodological bias; future studies employing individualized parcellation or multimodal imaging could improve spatial specificity and robustness. Fourth, we used AUC across the sparsity range 0.05–0.40 to reduce threshold-related bias. However, we acknowledge that the choice of threshold range may still affect network topology. In addition, anxiety and depression scores were not assessed in the control group, limiting our ability to directly evaluate the impact of emotional factors on functional connectivity. Lastly, while MEG provides excellent temporal resolution, its spatial resolution is limited compared to other modalities such as fMRI, and interpretation of deep structures (e.g. thalamus or brainstem) should be made with caution.

Conclusions

This study provides novel evidence of altered resting-state functional connectivity and network topology in patients with TTH using MEG and graph-theoretical analysis. TTH patients exhibited frequency-specific hyperconnectivity within pain- and sensory-related networks, alongside disruptions in both nodal and global network properties. Collectively, these findings indicate that TTH involves maladaptive reorganization of large-scale brain networks subserving pain processing and regulation. This network-level insight may guide the development of more targeted neuromodulatory or behavioral interventions.

Article highlights

Frequency-specific hyperconnectivity in TTH is identified using resting-state MEG.

Increased connectivity is revealed in pain- and sensory-related networks.

Graph theory shows disrupted nodal/global topologies (e.g. elevated local efficiency, clustering).

Altered networks correlate with clinical symptom severity (HIT-6).

The study findings suggest maladaptive reorganization in TTH-related brain networks.

Supplemental Material

sj-docx-1-cep-10.1177_03331024251386425 - Supplemental material for Disrupted functional network topology in tension-type headache: A cross-sectional magnetoencephalography study

Supplemental material, sj-docx-1-cep-10.1177_03331024251386425 for Disrupted functional network topology in tension-type headache: A cross-sectional magnetoencephalography study by Zhonghua Xiong, Dong Qiu, Jie Liang, Xiaoshuang Li, Zhi Guo, Mantian Zhang, Geyu Liu, Tianshuang Gao and Yonggang Wang in Cephalalgia

Supplemental Material

sj-jpg-2-cep-10.1177_03331024251386425 - Supplemental material for Disrupted functional network topology in tension-type headache: A cross-sectional magnetoencephalography study

Supplemental material, sj-jpg-2-cep-10.1177_03331024251386425 for Disrupted functional network topology in tension-type headache: A cross-sectional magnetoencephalography study by Zhonghua Xiong, Dong Qiu, Jie Liang, Xiaoshuang Li, Zhi Guo, Mantian Zhang, Geyu Liu, Tianshuang Gao and Yonggang Wang in Cephalalgia

Supplemental Material

sj-docx-3-cep-10.1177_03331024251386425 - Supplemental material for Disrupted functional network topology in tension-type headache: A cross-sectional magnetoencephalography study

Supplemental material, sj-docx-3-cep-10.1177_03331024251386425 for Disrupted functional network topology in tension-type headache: A cross-sectional magnetoencephalography study by Zhonghua Xiong, Dong Qiu, Jie Liang, Xiaoshuang Li, Zhi Guo, Mantian Zhang, Geyu Liu, Tianshuang Gao and Yonggang Wang in Cephalalgia

Footnotes

Acknowledgments

We extend our sincere gratitude to the National Neurological Imaging Centre of Beijing Tiantan Hospital, Capital Medical University, for their invaluable technical and equipment support throughout this study. Additionally, we express our heartfelt appreciation to the headache specialists whose expertise was instrumental in ensuring accurate diagnoses.

Author contributions

All authors contributed to the study conception and design. ZHX and DQ wrote the original draft. ZHX, DQ, XSL, ZG, MTZ, GYL and TSG analyzed the data. ZHX prepared figures and tables. YGW, DQ, JL, XSL, ZG, MTZ, GYL and TSG reviewed and edited the final draft. All authors contributed to the article and approved the final version submitted for publication.

Data availability

All data generated or analyzed during this study are included in this published article and its supplementary information files.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Ethical statement

This study was approved by the Ethics Committee of Beijing Tiantan Hospital, Capital Medical University (approval number: KY2022–044), as a sub-study of the ongoing China HeadAche DIsorders RegiStry Study (CHAIRS; NCT05334927). Written informed consent was obtained from all participants in accordance with the Declaration of Helsinki.

Funding

The authors disclosed receipt of the following financial support for the research, authorship and/or publication of this article: This work was supported by the Beijing Natural Science Foundation, National Key R&D Program of China, National Natural Science Foundation of China, Joint Funds of the National Natural Science Foundation of China (grant number F252058, 2024YFC2510100, 32170752, U24A20683).

ORCID iDs

Zhonghua Xiong

Xiaoshuang Li

Geyu Liu

Yonggang Wang

Supplemental material

Supplemental material for this article is available online.

References

GBD 2021 Diseases and Injuries Collaborators . Global incidence, prevalence, years lived with disability (YLDs), disability-adjusted life-years (DALYs), and healthy life expectancy (HALE) for 371 diseases and injuries in 204 countries and territories and 811 subnational locations, 1990–2021: a systematic analysis for the global burden of disease study 2021. Lancet 2024; 403: 2133–2161.

Headache Classification Committee of the International Headache Society (IHS) . The international classification of headache disorders, 3rd edition. Cephalalgia 2018; 38: 1–211.

Wang

Chen

, et al. Regional homogeneity abnormalities in patients with tension-type headache: a resting-state fMRI study. Neurosci Bull 2014; 30: 949–955.

Zhao

Chen

Cheng

, et al. Altered brain functional connectivity in patients with tension-type headache. Headache 2025; 65: 216–229.

Wang

Qiu

Mei

, et al. Altered functional connectivity of brainstem nuclei in new daily persistent headache: evidence from resting-state functional magnetic resonance imaging. CNS Neurosci Ther 2024; 30: e14686.

Qiu

Mei

, et al. Mapping brain functional networks topological characteristics in new daily persistent headache: a magnetoencephalography study. J Headache Pain 2023; 24: 161.

Yuan

Wang

Zhang

, et al. Altered functional connectivity of the right caudate nucleus in chronic migraine: a resting-state fMRI study. J Headache Pain 2022; 23: 154.

Baillet

. Magnetoencephalography for brain electrophysiology and imaging. Nat Neurosci 2017; 20: 327–339.

Grent-'t-Jong

Gajwani

Gross

, et al. Association of magnetoencephalographically measured high-frequency oscillations in visual Cortex with circuit dysfunctions in local and large-scale networks during emerging psychosis. JAMA Psychiatry 2020; 77: 852–862.

10.

Ren

Yao

Tian

, et al. Altered effective connectivity in migraine patients during emotional stimuli: a multi-frequency magnetoencephalography study. J Headache Pain 2022; 23: 6.

11.

Liu

Xiang

, et al. Resting state brain activity in patients with migraine: a magnetoencephalography study. J Headache Pain 2015; 16: 525.

12.

Wang

Shen

, et al. Functional connectivity across multi-frequency bands in patients with tension-type headache: a resting-state fMRI retrospective study. BMC Med Imaging 2025; 25: 145.

13.

Shen

, et al. Frequency-Specific alterations in brain function in patients with primary dysmenorrhea. Pain Med 2022; 23: 902–911.

14.

Ploner

Sorg

Gross

. Brain rhythms of pain. Trends Cogn Sci 2017; 21: 100–110.

15.

Tiemann

Schulz

Gross

, et al. Gamma oscillations as a neuronal correlate of the attentional effects of pain. Pain 2010; 150: 302–308.

16.

Seguin

Sporns

Zalesky

. Brain network communication: concepts, models and applications. Nat Rev Neurosci 2023; 24: 557–574.

17.

Avena-Koenigsberger

Misic

Sporns

. Communication dynamics in complex brain networks. Nat Rev Neurosci 2017; 19: 17–33.

18.

Mei

Qiu

Xiong

, et al. Disrupted topologic efficiency of white matter structural connectome in migraine: a graph-based connectomics study. J Headache Pain 2024; 25: 204.

19.

Martin

Blaisdell

Kwong

, et al. The Short-Form Headache Impact Test (HIT-6) was psychometrically equivalent in nine languages. J Clin Epidemiol 2004; 57: 1271–1278.

20.

Spitzer

Kroenke

Williams

. Validation and utility of a self-report version of PRIME-MD: the PHQ primary care study. Primary Care Evaluation of Mental Disorders. Patient Health Questionnaire. JAMA 1999; 282: 1737–1744.

21.

Buysse

Reynolds

Monk

, et al. The Pittsburgh Sleep Quality Index: a new instrument for psychiatric practice and research. Psychiatry Res 1989; 28: 193–213.

22.

Nasreddine

Phillips

Bédirian

, et al. The Montreal Cognitive Assessment, MoCA: a brief screening tool for mild cognitive impairment. J Am Geriatr Soc 2005; 53: 695–699.

23.

Taulu

Simola

. Spatiotemporal signal space separation method for rejecting nearby interference in MEG measurements. Phys Med Biol 2006; 51: 1759–1768.

24.

Smith

. Fast robust automated brain extraction. Hum Brain Mapp 2002; 17: 143–155.

25.

Nolte

. The magnetic lead field theorem in the quasi-static approximation and its use for magnetoencephalography forward calculation in realistic volume conductors. Phys Med Biol 2003; 48: 3637–3652.

26.

Veen

BDV

Buckley

. Beamforming: a versatile approach to spatial filtering. IEEE ASSP Mag 1988; 5: 4–24.

27.

Dale

Liu

Fischl

, et al. Dynamic statistical parametric mapping: combining fMRI and MEG for high-resolution imaging of cortical activity. Neuron 2000; 26: 55–67.

28.

Colclough

Brookes

Smith

, et al. A symmetric multivariate leakage correction for MEG connectomes. Neuroimage 2015; 117: 439–448.

29.

Schaefer

Kong

Gordon

, et al. Local-Global parcellation of the human cerebral Cortex from intrinsic functional connectivity MRI. Cereb Cortex 2018; 28: 3095–3114.

30.

Colclough

Woolrich

Tewarie

, et al.

How reliable are MEG resting-state connectivity metrics?

Neuroimage 2016; 138: 284–293.

31.

Chen

Zheng

, et al. Disrupted brain connectivity networks in aphasia revealed by resting-state fMRI. Front Aging Neurosci 2021; 13: 666301.

32.

Yan

Han

, et al. Diminished functional segregation and resilience are associated with symptomatic severity and cognitive impairments in schizophrenia: a large-scale study. Gen Psychiatr 2024; 37: e101613.

33.

Yao

Zhou

Zhang

, et al. Inflammation affects dynamic functional network connectivity pattern changes via plasma NFL in cognitive impairment patients. CNS Neurosci Ther 2024; 30: e14391.

34.

Bullmore

Sporns

. Complex brain networks: graph theoretical analysis of structural and functional systems. Nat Rev Neurosci 2009; 10: 186–198.

35.

Craig

. How do you feel? Interoception: the sense of the physiological condition of the body. Nat Rev Neurosci 2002; 3: 655–666.

36.

Ashar

Gordon

Schubiner

, et al. Effect of pain reprocessing therapy vs placebo and usual care for patients with chronic back pain: a randomized clinical trial. JAMA Psychiatry 2022; 79: 13–23.

37.

Skorobogatykh

van Hoogstraten

Degan

, et al.

Functional connectivity studies in migraine: what have we learned?

J Headache Pain 2019; 20: 108.

38.

Chen

Hsiao

, et al. Comparison of somatosensory cortex excitability between migraine and “strict-criteria” tension-type headache: a magnetoencephalographic study. Pain 2018; 159: 793–803.

39.

Chen

Chou

Lee

, et al. Comparison of gray matter volume between migraine and “strict-criteria” tension-type headache. J Headache Pain 2018; 19: 4.

40.

Chen

Hsiao

Wang

. Brain excitability in tension-type headache: a separate entity from migraine? Curr Pain Headache Rep 2021; 24: 82.

41.

Rudebeck

Rich

. Orbitofrontal cortex. Curr Biol 2018; 28: R1083–r1088.

42.

Rolls

. Emotion, motivation, decision-making, the orbitofrontal cortex, anterior cingulate cortex, and the amygdala. Brain Struct Funct 2023; 228: 1201–1257.

43.

Becker

Gandhi

Pomares

, et al. Orbitofrontal cortex mediates pain inhibition by monetary reward. Soc Cogn Affect Neurosci 2017; 12: 651–661.

44.

Kheirkhah

Baumbach

Leistritz

, et al. The right hemisphere is responsible for the greatest differences in human brain response to high-arousing emotional versus neutral stimuli: A MEG study. Brain Sci 2021; 11: 960.

45.

Allen

Bobnar

Kolber

. Left and right hemispheric lateralization of the amygdala in pain. Prog Neurobiol 2021; 196: 101891.

46.

Kong

Loggia

Zyloney

, et al. Exploring the brain in pain: activations, deactivations and their relation. Pain 2010; 148: 257–267.

47.

Buckner

Andrews-Hanna

Schacter

. The brain's default network: anatomy, function, and relevance to disease. Ann N Y Acad Sci 2008; 1124: 1–38.

48.

Napadow

LaCount

Park

, et al. Intrinsic brain connectivity in fibromyalgia is associated with chronic pain intensity. Arthritis Rheum 2010; 62: 2545–2555.

49.

Loggia

Kim

Gollub

, et al. Default mode network connectivity encodes clinical pain: an arterial spin labeling study. Pain 2013; 154: 24–33.

50.

Zaoli

Mazzarisi

Lillo

. Betweenness centrality for temporal multiplexes. Sci Rep 2021; 11: 4919.

51.

Zeng

, et al. Abnormal thalamocortical network dynamics in migraine. Neurology 2019; 92: e2706–e2716.

52.

Bassett

Bullmore

. Small-World brain networks revisited. Neuroscientist 2017; 23: 499–516.

53.

Bassett

Sporns

. Network neuroscience. Nat Neurosci 2017; 20: 353–364.

54.

Yang

Chang

. Chronic pain: structural and functional changes in brain structures and associated negative affective states. Int J Mol Sci 2019; 20: 3130.

55.

DeSouza

Woldeamanuel

Sanjanwala

, et al. Altered structural brain network topology in chronic migraine. Brain Struct Funct 2020; 225: 161–172.

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

0.03 MB

0.01 MB

1.37 MB