Longitudinal neurofunctional changes in medication overuse headache patients after mindfulness practice in a randomized controlled trial (the MIND-CM study)

Abstract

Background: Mindfulness practice has gained interest in the management of Chronic Migraine associated with Medication Overuse Headache (CM-MOH). Mindfulness is characterized by present-moment self-awareness and relies on attention control and emotion regulation, improving headache-related pain management. Mindfulness modulates the Default Mode Network (DMN), Salience Network (SN), and Fronto-Parietal Network (FPN) functional connectivity. However, the neural mechanisms underlying headache-related pain management with mindfulness are still unclear. In this study, we tested neurofunctional changes after mindfulness practice added to pharmacological treatment as usual in CM-MOH patients.

Methods: The present study is a longitudinal phase-III single-blind Randomized Controlled Trial (MIND-CM study; NCT03671681). Patients had a diagnosis of CM-MOH, no history of neurological and severe psychiatric comorbidities, and were attending our specialty headache centre. Patients were divided in Treatment as Usual (TaU) and mindfulness added to TaU (TaU + MIND) groups. Patients underwent a neuroimaging and clinical assessment before the treatment and after one year. Longitudinal comparisons of DMN, SN, and FPN connectivity were performed between groups and correlated with clinical changes. Vertex-wise analysis was performed to assess cortical thickness changes.

Results: 177 CM-MOH patients were randomized to either TaU group or TaU + MIND group. Thirty-four patients, divided in 17 TaU and 17 TaU + MIND, completed the neuroimaging follow-up. At the follow-up, both groups showed an improvement in most clinical variables, whereas only TaU + MIND patients showed a significant headache frequency reduction (p = 0.028). After one year, TaU + MIND patients showed greater SN functional connectivity with the left posterior insula (p-FWE = 0.007) and sensorimotor cortex (p-FWE = 0.026). In TaU + MIND patients only, greater SN-insular connectivity was associated with improved depression scores (r = -0.51, p = 0.038). A longitudinal increase in cortical thickness was observed in the insular cluster in these patients (p = 0.015). Increased anterior cingulate cortex thickness was also reported in TaU + MIND group (p-FWE = 0.02).

Conclusions: Increased SN-insular connectivity might modulate chronic pain perception and the management of negative emotions. Enhanced SN-sensorimotor connectivity could reflect improved body-awareness of painful sensations. Expanded cingulate cortex thickness might sustain improved cognitive processing of nociceptive information. Our findings unveil the therapeutic potential of mindfulness and the underlying neural mechanisms in CM-MOH patients.

Trial registration: Name of Registry; MIND-CM study; Registration Number ClinicalTrials.gov identifier: NCT0367168; Registration Date: 14/09/2018.

Keywords: Chronic pain; Cortical thickness; Functional connectivity; Headache; Medication overuse headache; Migraine; Mindfulness; Pain management; Resting state fMRI; Salience network.

Fig. 1

CONSORT flow diagram. Trial profile of 191 patients screened for eligibility. 91 patients were included in the MRI subsample. A total of 35 participants completed the study, of whom 34 analysed

Fig. 2

Resting-state fMRI results. Panel (a) shows longitudinal differences (Δ = T1 – T0) in SN functional connectivity between TaU + MIND and TaU groups (Δ TaU + MIND vs. Δ TaU). The left sensorimotor cluster is located at MNI −40, -20, +38; the left insular cluster is located at MNI −38, -14, +12. Results are corrected at p-FWE < 0.05 at the cluster level and p < 0.001 uncorrected at the voxel level. Three brain regions used as nodes of the SN are reported with dashed lines (rPFC = bilateral rostral prefrontal cortex, aINS = bilateral anterior insula, SMG = bilateral supramarginal gyrus; ACC = anterior cingulate cortex, not shown in the present figure). Panel (b) shows the unthresholded longitudinal SN functional connectivity maps (Δ TaU + MIND vs. Δ TaU). Panel (c) shows a significant negative correlation (error bar = 95% Confidence Interval) between longitudinal functional connectivity changes in SN-left insular cluster and changes in Beck Depression Inventory II (BDI) scores in TaU + MIND group. Panel (d) shows longitudinal difference in cortical thickness of the left insular functional cluster in TaU + MIND group compared to the TaU group. Abbreviations: L = left hemisphere; R = right hemisphere; T = t-values(degrees-of-freedom); Z = z-values(degrees-of-freedom)

Fig. 3

Structural results. Panel (a) shows longitudinal between-group vertex-wise differences (Δ TaU + MIND vs. Δ TaU) in right caudal anterior cingulate cortex (ACC) cortical thickness (MNI: x = + 6, y = + 9, z = + 36; p-FWE cluster-wise = 0.02). The ACC brain region used as node of the SN is reported with dashed lines. Panel (b) shows the unthresholded longitudinal between-group vertex-wise contrast map (Δ TaU + MIND vs. Δ TaU). Abbreviations: L = left hemisphere; R = right hemisphere