Disentangling self from pain: mindfulness meditation–induced pain relief is driven by thalamic–default mode network decoupling

Abstract

For millenniums, mindfulness was believed to diminish pain by reducing the influence of self-appraisals of noxious sensations. Today, mindfulness meditation is a highly popular and effective pain therapy that is believed to engage multiple, nonplacebo-related mechanisms to attenuate pain. Recent evidence suggests that mindfulness meditation–induced pain relief is associated with the engagement of unique cortico-thalamo-cortical nociceptive filtering mechanisms. However, the functional neural connections supporting mindfulness meditation–based analgesia remain unknown. This mechanistically focused clinical trial combined functional magnetic resonance imaging with psychophysical pain testing (49°C stimulation and pain visual analogue scales) to identify the neural connectivity supporting the direct modulation of pain-related behavioral and neural responses by mindfulness meditation. We hypothesized that mindfulness meditation-based pain relief would be reflected by greater decoupling between brain mechanisms supporting appraisal (prefrontal) and nociceptive processing (thalamus). After baseline pain testing, 40 participants were randomized to a well-validated, 4-session mindfulness meditation or book-listening regimen. Functional magnetic resonance imaging and noxious heat (49°C; right calf) were combined during meditation to test study hypotheses. Mindfulness meditation significantly reduced behavioral and neural pain responses when compared to the controls. Preregistered (NCT03414138) whole-brain analyses revealed that mindfulness meditation–induced analgesia was moderated by greater thalamus–precuneus decoupling and ventromedial prefrontal deactivation, respectively, signifying a pain modulatory role across functionally distinct neural mechanisms supporting self-referential processing. Two separate preregistered seed-to-seed analyses found that mindfulness meditation–based pain relief was also associated with weaker contralateral thalamic connectivity with the prefrontal and primary somatosensory cortex, respectively. Thus, we propose that mindfulness meditation is associated with a novel self-referential nociceptive gating mechanism to reduce pain.

In Brief

These findings indicate that mindfulness meditation engages a unique self-referential nociceptive filtering mechanism to reduce the elaboration of pain.

1. Introduction

Pain is constructed and modulated by a dynamic interplay between real-time appraisals of corresponding sensory, cognitive, and affective processes. Mindfulness meditation, a self-regulatory technique, is highly effective at reducing self-reported experimental and chronic pain.^7,9,15,38 It is practiced by promoting a “detached observation” to reduce the self-referential value of all arising sensations. Mindfulness meditation is more effective and engages distinct neural,⁵² endogenous,^26,46 and parasympathetic¹ mechanisms from placebo to reduce pain. Using perfusion-based functional magnetic resonance imaging (fMRI) acquisition, we found that meditation-induced pain reductions were directly associated with higher cerebral blood flow (CBF) in the ventrolateral prefrontal/orbitofrontal (OFC) cortex and lower CBF in the contralateral thalamus.^52,53 Based on these reproducible effects, we proposed that mindfulness meditation-induced pain relief would be driven by unique cortico-thalamo-cortical nociceptive filtering mechanisms indicated by (1) stronger connectivity between the right OFC and contralateral thalamus to reflect excitatory innervations on the inhibitory thalamic reticular nuclei (TRN) and (2) weaker contralateral thalamus–primary somatosensory cortex (S1) connectivity. However, previous studies using perfusion fMRI^52,53 were not optimized to dissect event-related functional connectivity because of low temporal resolution (repetition times > 9 seconds). Thus, there are no known studies that have identified the functional connections supporting the direct attenuation of pain by mindfulness meditation.

Growing evidence^10,54 also indicate a relationship between mindfulness meditation-related health benefits and attenuated default mode network processing. The default mode network is characterized by oscillating activity between the medial prefrontal cortex (mPFC), posterior cingulate cortex (PCC)/precuneus, and inferior and lateral temporal cortices. The default mode network is critically involved in facilitating self-referential processes^36,39 and pain-related ruminations.^3,24 Higher self-reported trait mindfulness was associated with lower pain sensitivity and weaker processing in the PCC/precuneus during noxious heat stimulation in meditation-naive individuals.^21,55 However, the brain mechanisms supporting trait mindfulness are different from those engaged by the active practice of mindfulness meditation in response to pain-evoking stimulation. Thus, the proposed work extends on previous investigations by examining the neural connectivity explicitly engaged by mindfulness meditation during noxious stimulation to reduce pain.

The proposed preregistered mechanistically focused clinical trial (NCT03414138) combined blood oxygen level–dependent signaled fMRI acquisition with psychophysical pain testing (49°C stimulation and pain-focused visual analogue scales [VASs]) to identify the neural connectivity supporting the direct modulation of pain-related behavioral and neural responses by mindfulness meditation. We postulated that mindfulness meditation reduces pain by eliciting a prefrontal-mediated nociceptive gating mechanism at the level of the thalamus. Separate 2 (mindfulness vs control group) × 2 (prerest) vs (postmindfulness/rest) general linear model (GLM) analysis of variance (ANOVA) examined whether mindfulness meditation produced significant reductions in behavioral and neural pain responses when compared to the controls. All neuroimaging analyses were conducted during pain-evoking 49°C heat to identify brain mechanisms supporting the direct attenuation of pain by mindfulness meditation. Based on updated working hypotheses and to test the primary aim of the study, whole-brain and seed-to-seed psychophysiological interaction (PPI) analyses³³ were performed to determine whether mindfulness meditation-induced pain relief would be moderated by (1) stronger lateral OFC connectivity with the contralateral thalamus and (2) weaker thalamic connectivity with the somatosensory cortex (SI) representation of the noxious stimulation site (right calf). The present findings provide novel evidence that the direct modulation of pain by mindfulness meditation is moderated by brain mechanisms supporting the attenuation of self-referential and nociceptive processing.

2. Methods

2.1. Sample size determination

Based on previous work,^51,52 we expected a large behavioral and neural effect size difference between mindfulness meditation and nonmanipulated controls. Our previous fMRI-based contrast of parameter estimates (COPEs) (left thalamus; right OFC)⁵³ and fMRI-based statistical power calculating software (fmripower.org) were used for sample size determination. Forty participants (n = 20/group) provided over 85% statistical power to detect the hypothesized large-effect sizes (r = 0.50) exhibited in predefined⁵³ brain mechanisms supporting meditation-induced pain reductions.

2.2. Participants

Wake Forest School of Medicine's Institutional Review Board approved (IRB#182082) all study procedures. One hundred thirty-seven healthy, pain-free, and meditation-naive participants (aged 18-65 years) were screened from the local community. Exclusion criteria included individuals who were claustrophobic, pregnant, and/or had MRI contraindications. All participants provided written informed consent at the initial study visit with all methods clearly explained, acknowledging that (1) they would experience painful heat stimuli and (2) they were free to withdraw from the study at any time.

At screening, 79 participants were excluded for not meeting the inclusion and exclusion criteria (Fig. 1). Fifty-eight participants were enrolled in the study, but 18 individuals did not complete study procedures for several reasons (Fig. 1). Forty participants (all right-handed; mean age = 30 years ± 10 years; 20 males; 20 females) were included in the final analysis (22 = White, 13 = Black, 3 = Hispanic, 1 = Asian, and 1 = Native American; Fig. 1).

Figure 1.: Consolidated Standards of Reporting Trials (CONSORT) flow diagram: at screening, 79 participants were excluded for not meeting the inclusion and exclusion criteria. Fifty-eight participants were enrolled into the study. After session 1, participants were dropped for low pain sensitivity (n = 6), equipment not fitting (n = 1), and chronic pain (n = 1). After randomization, a subject from each group was dismissed for “no-shows.” In the MRI session, a total of 4 participants from the mindfulness group and 2 from the control group were dismissed for several reasons. Two participants completed their respective fMRI scans but were removed from the final analysis (and replaced at random) due to a procedural error and an exhibited fMRI-related artifact, respectively. The targeted sample size of 40 participants were included in the study. MRI, magnetic resonance imaging; fMRI, functional magnetic resonance imaging.

2.3. Randomization procedure

Group randomization was stratified by sex (10/group) after successfully completing session 1 (Fig. 2). Men and women were randomized without replacement using a random number generator by a research technician not involved in data collection. Participants were informed of their respective group assignments after session 1.

Figure 2.: Study procedures: session 1: after attaining informed consent, participants were trained to use pain intensity and unpleasantness visual analogue scale (VAS) in response to noxious heat stimulation (psychophysical training). The experimental heat series included ten, 10-second plateaus of 49°C interleaved with 14 seconds of 35°C stimulation applied to the back of the right calf. Participants were instructed to keep their eyes closed and rest in response to 2 heat series. Visual analogue scale ratings were collected after each respective heat series. After 2 heat series, participants were instructed to “sit quietly for 10 minutes” to control for the time spent initiating meditation in session 6. After this 10-minute period, participants were administered 2 more heat series, and VAS ratings were collected afterward. Participants were then randomized to a 4-session (20 minutes/session) mindfulness meditation or a control group that listened to the Natural History of Selborne across four 20-minute sessions. After successful completion of their respective regimens (sessions 2-5), participants reported to the MRI center for postintervention functional MRI sessions. Session 6: after being positioned in the scanner, we administered 2 heat series while participants rested with their eyes closed during fMRI acquisition. Visual analogue scale ratings were collected after each respective heat series. Before anatomical acquisition, we instructed the members of the mindfulness group to “begin meditating and to continue until the end of the experiment.” The control group was instructed to “keep eyes closed.” After 10 minutes elapsed, we administered 2 more heat series during functional acquisition and collected pain ratings after each respective series. MRI, magnetic resonance imaging; fMRI, functional magnetic resonance imaging.

2.4. Stimuli

An fMRI-compatible thermal sensory analyzer (TSA-II, Medoc, Inc, Raleigh, NC) fitted with a 16-mm² surface area probe delivered all thermal stimuli. Heat series during the experiment (300-second duration) consisted of 10 alternating 10-second plateaus of 49°C interleaved with 14 seconds of 35°C stimulation delivered to the back of the right calf. To minimize habituation, the thermal probe was moved to a new stimulation site on the right calf after each experimental series. Participants were free to remove the stimuli at any time by lifting their limb away from the probe holder.

2.5. Psychophysical assessment of pain

Pain intensity and unpleasantness ratings were assessed with a 15-cm sliding VAS. The minimum rating (0) was designated as “no pain sensation” and “not at all unpleasant,” whereas the maximum (10) was labeled as “most intense pain sensation imaginable” or “most unpleasant sensation imaginable”. Participants were instructed to use the VAS scale only if they felt the stimulus to be painful. Thus, if there was no pain to report, participants verbalized with “0” or “no pain.”

2.6. Anatomical magnetic resonance imaging acquisition

Magnetic resonance imaging data were acquired on a 3T Siemens MAGNETOM Skyra scanner with a 32-channel head coil (Siemens Healthineers AG, Munich, Germany). High-resolution structural scans were acquired using an MP-RAGE sequence (TI = 900 milliseconds [ms]; TR = 2300 ms; flip angle = 9°; TE = 2.98 ms; 1 mm isotropic spatial resolution; 192 slices, GRAPPA factor = 2; scan time = 312 seconds [s]).

2.7. Functional magnetic resonance imaging acquisition

For functional imaging, blood oxygen level–dependent images were acquired using echo planar imaging (EPI; TE = 25 ms; TR = 2000 ms; 35 × 4-mm-thick slices with no gap; 4.00 × 4.00 mm in-plane resolution; flip angle = 75°; 180 repetitions; scan time = 300 seconds).

2.8. Experimental design

2.8.1. Experimental session 1 (psychophysical training + baseline pain testing)

After providing written consent, participants were (1) familiarized with thirty-two 5-second duration thermal stimuli (35-49°C) on the ventral aspect of the forearm and (2) trained to use the pain VAS. Afterward, participants placed their right calf on a custom-made thermal probe holder. Baseline psychophysical pain responses were assessed in response to noxious heat by administering a total of 4 heat series. Visual analogue scale pain intensity and unpleasantness ratings were collected after each series, and the thermal probe was moved to a different region on the back of the right calf. Throughout session 1 (Fig. 2), participants were instructed to remain still and sit quietly. After the first 2 heat series, participants were told to continue to “sit quietly” for 10 minutes to control for the time elapsed in the postintervention MRI session. Afterward, 2 heat series were administered, and pain ratings were collected, respectively. After successful completion of sensory testing, participants were randomized to one of the groups and informed of their respective group assignments.

2.8.2. Experimental session 2 to 5 (intervention sessions)

2.8.2.1. Mindfulness meditation training regimen

As adapted in our previous work,^1,46,53 participants in the mindfulness meditation group participated in 4 separate sessions (20 minutes each and on separate days) of mindfulness-based mental training. These sessions were facilitated by experienced mindfulness instructors. Across all the meditation training sessions, participants were instructed to focus on the changing sensations of the breath and to reduce self-referential judgments by acknowledging arising thoughts, feelings, and emotions without judgment or emotional reaction. When attention to breathing sensations disengaged, participants were encouraged to “simply return their attention back to the breath sensations.” Participants were instructed not to explicitly change their breathing rate. Contrary to traditional meditation training interventions, participants were explicitly instructed to not practice meditation outside of study training to reduce interindividual variability in practice time effects. In meditation training session 3, the same basic principles of the previous sessions were reiterated. An audio recording of fMRI scanner sounds was introduced during the last 10 minutes of meditation training to familiarize participants with meditating in an MRI environment. On the final training session (session 4), participants received minimal meditation instruction and were required to lie in the supine position and meditate during an audio recording of the fMRI sounds to simulate the scanner environment.

2.8.2.2. Book-listening control regimen

The control group listened to an audio recording of The Natural History and Antiquities of Selborne across 4 sessions (20 minutes each and on separate days). This book has been used as a neutral comparison regimen in our studies because it does not improve mood but controls for the time elapsed in the meditation intervention, facilitator attention, and a group session.⁵¹ Audio recordings continued serially from where they ended in the previous book-listening session. Therefore, participants who successfully completed the book-listening regimen listened, in total, to 80 minutes of the Natural History and Antiquities of Selborne. Study volunteers were not allowed to sleep, use their phones, or talk to the experimenter during book listening. In control session 3, we introduced the sounds of the scanner in the last 10 minutes of book listening. To better match the procedures to the meditation group, participants were instructed to lie in the supine position and listen to the audiobook and sounds of the scanner during book-listening session 4.

2.8.3. Experimental session 7 (functional magnetic resonance imaging)

Participants first reported to the Wake Forest Center for Biomolecular Imaging and were positioned in the MRI scanner with a respiratory transducer around the chest and a pulse oximeter on the left index finger (respiration and heart rate data not presented here). Participants positioned their right calf on a custom-made thermal probe holder fitted with a force transducer to continuously confirm contact of the thermal probe with the calf. All physiological and thermal heat stimuli were logged with a digital chart recorder (MP160; Biopac Systems Inc., Goleta, CA).

2.8.4. Premanipulation

During the premanipulation condition (characterized as “rest”), all participants were instructed to “not move” and keep their “eyes closed.” There were 2 types of thermal stimulation series: (1) heat series (300-second duration) = 10 alternating 10-second plateaus of 49°C and 14 seconds of 35°C stimulation and (2) neutral series (300-second duration; neutral series not presented here) = continual 35°C stimulation. Participants were administered 2 neutral (not presented here) and 2 heat series, to the back of the right calf, in an alternating fashion, and pain ratings were collected after each series.

An anatomical scan was then collected after participants in the mindfulness meditation group were instructed to “begin meditating and to continue until the end of the experiment.” As in our previous meditation fMRI studies,^52,53 we provided meditation group participants supplementary time during the anatomical acquisition to establish a meditative state before functional acquisition. Members of the control group were instructed to continue to “lie still with your eyes closed” and to not move. After the structural scan, participants were provided 3 supplemental minutes in silence before continuing fMRI acquisition and heat stimulation.

2.8.5. Postmanipulation

Participants were administered alternating series of 2 neutral (not presented here) and 2 heat series during the postmanipulation scan. Postmanipulation is characterized as “control-rest” for the control group and “meditation” for the mindfulness meditation group. Pain ratings were collected after each heat series.

2.9. Analysis of behavioral data

A 2 manipulation (pre vs post) × pain type (intensity vs unpleasantness) × group (mindfulness vs control) × session (1 and 6) ANOVA tested (SPSS v27, IBM, Armonk, NY) the primary hypothesis that mindfulness meditation would produce greater pain reductions than rest and the book-listening control group. Significant main effects and interactions (P < 0.05) were investigated with planned simple effects tests.

2.10. Analysis of neuroimaging data

All neuroimaging data were analyzed using FSL version 6.00 using customized analytical methods.

2.10.1. Preprocessing

Six seconds of nonsteady 4D data were removed from the start of each functional acquisition. To correct for B0 field nonuniformity, fieldmap unwarping was applied (FSL FUGUE). Distortion-corrected functional images were coregistered to their corresponding T1w reference using FLIRT boundary-based registration, the standard Montreal Neurological Institute template through 12-parameter linear, and 10 mm warp resolution nonlinear transformations. Head motion was corrected, and the analogous 6 rotation and translation parameters were estimated. From these motion parameters, framewise displacement was calculated and used to flag motion outliers above 0.9 mm displacement. Functional images were then spatially smoothed using a Gaussian kernel (full-width half-maximum of 5 mm) and high-pass filtered with a period of 100 seconds. Autocorrelation correction and model estimation was applied using FILM.

White matter (WM) segmentation was performed for each structural T1 (FSL FAST), thresholded at 99% probability then eroded, before being registered to each respective functional image. A time series (eigenvariate) was extracted from each subject's WM mask to regress out nonneural sources of noise.

2.10.2. General linear model

For the 4 fMRI-based noxious heat series, the general linear model (GLM) design matrix included a stimulus regressor indicating the onsets and durations of the ten 10 heat stimuli (a total of 100 seconds of noxious 49°C) relative to an implicit baseline (a total of 180 seconds of innocuous 35°C). This regressor was scaled to +1, temporally filtered, and convolved with a double gamma model of the hemodynamic response function, and its temporal derivative was added to the model. Each model also included covariates of no interest—a high-variance signal from WM, 6 head motion parameters, and separate regressors for each volume flagged as high motion, to reduce physiological and motion-related noise. The results of this GLM were effect size (β) maps of the stimulus-related regional signal changes.

We performed intrasubject fixed effects (second level) analyses within each subject separately for the stimulus regressor and proceeded to intersubject mixed effects (third level) analysis using a whole-brain between-group independent samples t test on the contrast maps of premanipulation (rest) vs postmanipulation (meditation or rest). Within each group, demeaned and residualized (prerest vs postmanipulation) pain intensity and unpleasantness (orthogonalized to intensity) ratings were entered as covariates to determine whole-brain voxel-wise correlations with changes in pain intensity. Importantly, Z (gaussianized T/F) statistic images were determined using FLAME 1 + 2 and a corrected cluster of height z ±3.10 and size P < 0.05, controlling for the family-wise error rate.¹²

2.10.3. Seed-to-seed and whole-brain psychophysiological interaction analyses

To directly test our working hypotheses, we conducted 2 seed-to-seed and 2 seed to whole-brain PPI analyses to investigate heat stimulus-modulated connectivity³³ comparing premanipulation (rest) and postmanipulation (meditation or rest). Within each group, we tested preregistered connections of interest (1) left thalamic to whole brain, (2) right OFC to whole brain, (3) right OFC to left thalamus, and (4) left thalamus to left SI representation of the stimulation site on the right calf.

Seed regions were created from the higher-order dictionaries of functional modes-128 component (DiFuMo) atlas.¹¹ Seeds were selected from the contralateral left thalamus and right OFC that overlapped with the peak voxel coordinates of brain mechanisms supporting meditation-related analgesia in a previous study.⁵³ A seed corresponding to the left-lateralized SI representation of the right leg was manually created using a 5-mm mask around the right calf representation² that overlapped with the coordinates corresponding to peak right calf activation from our previous study.⁵³ The standard space of these seed regions was linearly transformed to each functional image and then binarized at 75% probability, and the eignenvariate values were extracted.

Psychophysiological interaction models were generated for each heat series using 3 regressors. The first regressor (psychological regressor) modeled the zero-centered convolved heat stimulus (see the General linear model section), the second regressor (physiological regressor) was the mean-centered time series values from the seeds (respectively), and the third regressor was the interaction regression term between the first (psychological) and second (physiological) regressors. Denoising covariates paralleled those of the GLM described above. Within each subject, we then compared premanipulation (rest) vs postmanipulation (meditation) PPI connectivity during 49°C as compared to 35°C. The contrast maps of these second-level analyses were entered into the third-level within-group analyses. Furthermore, demeaned residuals (postmanipulation vs premanipulation) of pain intensity and unpleasantness (orthogonalized to intensity) ratings were entered as covariates. All whole-brain PPI analyses accounted for mixed effects (FLAME 1 + 2) and were thresholded using a corrected cluster of height z ± 3.1 and size P < 0.05, controlling for the family-wise error rate.^12,47 Seed-to-seed tests restricted the analysis space to a priori regions of interest using nonparametric permutation testing (FSL RANDOMISE). Statistical inferences were based on threshold-free cluster enhancement, P < 0.05.⁴⁰

3. Results

3.1. Behavioral findings

3.1.1. Mindfulness meditation significantly reduced pain

There were no significant between-group differences in preintervention pain ratings, F(1, 34) = 0.11, P = 0.75, ${\mathrm{\eta }}_{\text{P}}^2$= 0.03 (Fig. 3). The RM ANOVA revealed a significant group × manipulation × session interaction, F(1, 38) = 18.59, P < 0.001, ${\mathrm{\eta }}_{\text{P}}^2$ = 0.33. To interpret the significant interaction, simple effects tests determined that mindfulness meditation produced significant reductions in pain intensity (−32%) and pain unpleasantness (−33%) ratings from rest to meditation and when compared premanipulation to postmanipulation in the control group, respectively (Ps < 0.001; Fig. 4). In the postintervention MRI session, the control group reported significantly higher pain from premanipulation (first half of scan) to postmanipulation (second half of scan; P = 0.05). As expected, both groups reported lower pain in premanipulation in session 6 when compared to premanipulation in session 1. Yet, the controls exhibited significant pain habituation⁴⁹ from session 1 to 6 (P = 0.003) and when compared to the meditation group (P = 0.004) during premanipulation. The pain habituation effect is perplexing but may be associated with relief that the experiment was almost over.

Figure 3.: Psychophysical visual analogue scale pain intensity (left) and unpleasantness (right) ratings (± 1 SEM) in response to premanipulation (rest) vs postmanipulation (rest) and noxious heat (49°C) during the preintervention psychophysical testing session. There were no significant within-group or between-group effects. VAS, visual analogue scale.

Figure 4.: Psychophysical visual analogue scale pain intensity (left) and unpleasantness (right) ratings (± 1 SEM) in response to premanipulation (rest) vs postmanipulation (rest; meditation) and noxious heat (49°C) during the postintervention MRI session. Mindfulness meditation produced significant reductions in pain intensity (−32%) and pain unpleasantness (−33%) ratings when compared to rest and the change in pain from premanipulation to postmanipulation in the control group (Ps < 0.001). MRI, magnetic resonance imaging; VAS, visual analogue scale.

3.2. Neuroimaging findings

3.2.1. Mindfulness meditation–induced pain relief is associated with greater vmPFC deactivation

Importantly, statistical power calculation analyses (NCT03414138) successfully predicted this study's large behavioral and brain-based effect sizes. All analyses were conducted during noxious heat stimulation (49°C). A whole-brain independent samples t test found that mindfulness meditation produced significant reductions in nociceptive-processing regions including the bilateral posterior insula, secondary somatosensory cortex (SII), parietal operculum, dorsal ACC (dACC), supplementary motor area (SMA), and cerebellum during noxious heat stimulation (Tables 1 and 2; Fig. 5 left panel) when compared to rest and from premanipulation to postmanipulation in the control group, respectively.

Table 1 - Brain coordinates on significant activation and deactivation.

Region	Z score	P	x, y, z
Figure 5: Meditation (rest vs meditation) < controls (prerest vs postrest)
Bilateral occipital pole	4.87	7.77 × 10−11	−52, −84, −8
Bilateral supplementary motor area and bilateral dorsal anterior cingulate cortex	5.08	2.22 × 10−14	−4, −8, 40
Left secondary somatosensory cortex	4.65	1.52 × 10−4	−60, −24, 10
Left posterior insula	4.05	1.68 × 10−4	−50, −46, 24
Right parietal operculum	5.12	5.96 × 10−8	44, −34, 22
Figure 5: mindfulness group: rest > meditation
Bilateral supplementary motor area and bilateral dorsal anterior cingulate cortex	5.34	9.47 × 10−9	−2, −10, 54
left anterior insula, left amygdala, and left hippocampus	4.82	2.17 × 10−11	−24, −12, −14
Right anterior insula, right amygdala, and right hippocampus	4.91	3.16 × 10−6	34, −8, −16
Right central operculum and right parietal operculum	4.83	2.90 × 10−8	48, −44, 16
Left posterior insula and left secondary somatosensory cortex	4.61	4.14 × 10−9	−54, −20, 22
Figure 5: mindfulness group: meditation–rest and brain correlates of pain intensity reductions
Right ventromedial prefrontal cortex	4.42	0.0418	8, 36, −22
Figure 6: control group: postscan rest > prescan rest
Left occipital lobe			−4
Figure 7: mindfulness group: meditation–rest thalamic decoupling correlates of pain intensity reductions
Precuneus	3.94	8.18 × 10−5	8, −72, 30
Occipital pole	4.02	0.0209	0, −78, −2
Figure 8: mindfulness group: meditation–rest seed-to-seed connectivity correlates of pain intensity reductions
Thalamus	*	0.0164	0, −16, 4
Left primary somatosensory cortex	*	0.0313	−8, −44, 62
Figure 9: mindfulness group: meditation–rest vlPFC connectivity correlates of pain intensity reductions
Left parietal operculum, left posterior insula, and left secondary somatosensory cortex	4.22	0.0499	−50, −38, 24

vlPFC, ventrolateral prefrontal cortex.

Table 2 - Mean (SD) of the parameter estimates extracted from significant brain deactivation (Fig. 5 left panel) during noxious heat for premanipulation and postmanipulation for the mindfulness meditation (meditation) and book-listening (control) group, respectively.

	Meditation group		Control group
	Premanipulation	Postmanipulation	Premanipulation	Postmanipulation
Mean (SD)	0.21 (0.19)	0.04 (0.11)	0.19 (0.18)	0.22 (0.23)

Figure 5.: All neuroimaging analyses were performed during noxious heat (49°C > 35°C). Left panel: the whole-brain independent samples t test revealed that mindfulness meditation significantly reduced activation in the bilateral posterior insula, secondary somatosensory cortex (SII), parietal operculum, dorsal ACC (dACC), supplementary motor area (SMA), and primary visual cortex (V1) during noxious heat when compared to rest and the change in pain from premanipulation to postmanipulation in the control group. Middle panel: a paired-samples t test using a change in pain (demeaned postmanipulation − premanipulation) found that meditation produced significant reductions in the dACC, SMA, bilateral anterior/posterior insula, amygdala, hippocampus, central parietal operculum, and SII when compared to rest. Right panel: greater mindfulness-based pain intensity reductions, from rest to meditation, was associated with stronger vmPFC deactivation (r = 0.73, P < 0.001). R, subject right, slice locations correspond to standard stereotaxic space.

During noxious heat, the within-group ANOVA established that mindfulness meditation reduced bilateral amygdala, hippocampus, and mid and anterior insula activation when compared to rest (Fig. 5 middle panel). Importantly, greater mindfulness meditation–based pain reductions were, from rest to meditation, associated with greater vmPFC deactivation (Fig. 5 right panel). The control group exhibited greater occipital lobe activation in the second half of the scan when compared to the first half of the scan (Fig. 6).

Figure 6.: Neuroimaging analyses were performed during noxious heat (49°C > 35°C). A paired-samples t test using a change in pain (demeaned postmanipulation − premanipulation) found that the controls exhibited a significant increase in occipital lobe activation in the second half of the scan when compared to the first half of the scans. Slice locations correspond to standard stereotaxic space.

3.2.2. Mindfulness meditation–induced pain relief is moderated by weaker thalamic–precuneus connectivity

Contralateral thalamic–whole-brain PPI analyses revealed that greater meditation-induced reductions, from rest to meditation, in pain intensity ratings were associated with greater thalamic decoupling from the precuneus and the primary visual cortex (V1) (Fig. 7). There were no other significant meditation-related mean effects, correlations, or control group effects.

Figure 7.: Left thalamic–whole-brain psychophysiological interaction analyses (PPI; z = ± 3.1, P < 0.05) were performed during 49°C plateaus. Psychophysiological interaction analyses were conducted during meditation as compared to rest. Pain intensity and unpleasantness ratings were regressed on the mean effect, respectively. Greater mindfulness meditation–induced pain intensity relief was moderated by weaker thalamic connectivity with the precuneus and primary visual cortex (V1) (r = 0.74, P < 0.001). R, subject right, slice locations correspond to standard stereotaxic space.

3.2.3. Mindfulness meditation reduces pain by attenuating prefrontal cortex–thalamic and thalamic–somatosensory cortex connectivity

Contrary to our hypotheses, seed-to-seed PPI analyses indicated that weaker right OFC connectivity with the contralateral thalamus predicted mindfulness meditation–induced reductions in pain from rest to meditation (Fig. 8 top panel). However, as predicted, weaker thalamic–SI functional connectivity predicted stronger meditation-induced pain intensity reductions (Fig. 8 bottom panel).

Figure 8.: Separate seed-to-seed psychophysiological interaction analyses were conducted (threshold-free cluster enhancement, P < 0.05)⁴⁰ to determine whether (1) right orbitofrontal cortical connectivity with the left thalamus and (2) left thalamus connectivity with the primary somatosensory cortex predicted mindfulness meditation–induced pain relief. Top panel: weaker right orbitofrontal cortical connectivity with the contralateral thalamus predicted mindfulness-induced pain reductions from rest to meditation (r =0.65, P =0.001). Bottom panel: weaker thalamic–somatosensory cortex functional connectivity predicted stronger meditation-induced pain intensity reductions (r =0.48, P = 0.03). There were no other significant effects for either group. R, subject right, slice locations correspond to standard stereotaxic space. SI, somatosensory cortex; vlPFC, ventrolateral prefrontal cortex.

3.2.4. Mindfulness meditation–induced pain relief is moderated by weaker orbitofrontal cortex–posterior insula/secondary somatosensory cortex connectivity

Right OFC to whole-brain PPI analyses revealed that greater meditation-induced pain relief, from rest to meditation, was associated with weaker OFC connectivity with the contralateral posterior insula, SII, and parietal operculum (Fig. 9). There were no other significant meditation-related mean effects, correlations, or control group effects.

Figure 9.: Right orbitofrontal cortex–whole-brain psychophysiological interaction analyses were performed during 49°C plateaus. Psychophysiological interaction analyses (z = ±3.1, P < 0.05) were conducted during meditation as compared to rest. Pain intensity and unpleasantness ratings were regressed on the mean effect, respectively. Greater mindfulness meditation–induced pain intensity reductions from rest to meditation was moderated by weaker right orbitofrontal cortex connectivity with the contralateral posterior insula, SII, and parietal operculum (r = 0.80, P < 0.001). There were no other significant meditation-related mean effects, correlations, or control group effects. R, subject right, slice locations correspond to standard stereotaxic space. OFC, orbitofrontal cortex; SII, secondary somatosensory cortex; vlPFC, ventrolateral prefrontal cortex.

4. Discussion

The dynamic transformation of nociceptive information into a subjectively available pain experience is predicated on intrinsically reflexive interpretations of ascending noxious input. By contrast, mindfulness meditation is premised on sustaining nonjudgmental awareness of all sensory, cognitive, and self-referential events. This study combined fMRI, pain-evoking heat, and meditation to identify whether mindful-based pain modulation is associated with modifying nociceptive and appraisal processes. Study power calculation analyses using a sample of 40 participants successfully predicted the large behavioral and neural effect sizes. Importantly, we demonstrate that mindfulness meditation directly attenuated self-reported pain (Fig. 4) and corresponding brain mechanisms supporting nociception and self-referential processing (Figures 5 and 7–9).

Mindfulness-based pain relief was moderated by weaker functional connectivity between the contralateral thalamus and the precuneus (Fig. 7). The precuneus and thalamus are critically implicated in the quality of awareness of self and sensory environment (ie, self-consciousness) through connections to brainstem arousal systems.^5,44 Precuneal and thalamic and thalamus predict loss and CBF increase signal recovery of consciousness, respectively.^14,44,48 Together, the precuneus and thalamus form a binding site to integrate to facilitate multisensory integration with internal representations of self.^5,31,44 Yet, the precuneus is not directly connected to sensory ventroposterior lateral thalamus, but rather the anterior central intralaminar complex and pulvinar nuclei of the thalamus.⁸ As the central node of the default mode network, the neural network facilitating self-referential processes, the precuneus exhibits the highest metabolic consumption in the brain¹⁹ and is anatomically situated to integrate somatosensory representations with the sense of self.^34,50 Higher precuneal activation is associated with lower pain sensitivity^4,25 and higher dispositional mindfulness.⁵⁵ Thus, the relationship between mindfulness-induced pain relief and weaker thalamic–precuneal connectivity may reflect volitionally mediated disengagement from self-referential appraisals of nociceptive input,^5,44 a potentially novel pain modulatory mechanism.

The present findings also extend on previous work to demonstrate that mindfulness engages multiple mechanisms to reduce pain. When compared to the controls, mindfulness meditation significantly reduced noxious heat-driven activation in the dACC, SMA, bilateral SII/parietal operculum, and posterior insula (Fig. 5). The whole-brain regression analysis found that greater meditation-induced pain reductions were associated with greater vmPFC deactivation, a cortical midline structure critically involved in facilitating the experience of a “multifaceted self.”¹⁸ Specifically, the vmPFC is a regulator of self-narrative processing of moment-to-moment experience.^5,22,31,37 Subcortical vmPFC connections facilitate generation of incoming sensory information to produce value-based self-representations,^30,32,37 the sense of “self and body-ownership”(ie, “mine-ness”), and personal significance.^5,22,32 It is suitable, then, that mindfulness meditation reduced amygdalar and hippocampal activation during noxious heat, with mechanisms that suggest a reduction in affective reactivity to ongoing sensory input.⁴² Mindfulness meditation is premised on “letting go and acceptance of accepting arising sensory and cognitive events” and may lower pain by reducing the subjective embodiment of noxious sensations.¹³ Thus, these data indicate that mindfulness meditation weakens cortical self-referential midline processing of ascending nociceptive inputs to promote a sensory egocentric decoupling mechanism.

Although not implicative of direct connectivity, seed-to-seed PPI analyses found that mindfulness meditation–induced pain reductions were also associated with weaker functional connectivity between the right OFC and the contralateral thalamus (Fig. 8). As predicted, pain relief produced by mindfulness meditation was associated with greater decoupling between the contralateral thalamus and the left SI corresponding to the stimulation site (right calf). Thus, we provide novel supplementary evidence that mindfulness meditation reduces pain by attenuating low-level afferent processing to diminish the elaboration of nociception throughout the cortex. The OFC–whole-brain PPI analysis provided novel insight to how mindfulness meditation may uniquely uncouple interpretive appraisals of sensory, albeit noxious experience. Weaker PFC connectivity with brain mechanisms that are critical for sensory pain discrimination (contralateral posterior insula, SII, and parietal operculum; Fig. 9) was associated with greater mindfulness meditation–based pain relief. Lateral OFC projections to the dorsomedial or ventroanterior thalamus²⁸ drive self-appraisals of low-level sensory input.^20,27,29 Thus, we propose that mindfulness-based attention inhibits the integration of self-referential and nociceptive processes through a PFC nociceptive gating mechanism^17,45 that may be reflective of a nonevaluative cognitive stance.

As previously described in contemporary and traditional texts,^6,23,35,43 mindfulness meditation may reduce pain by uncoupling egocentric appraisals of salient nociceptive inputs. Converging lines of evidence demonstrate that stronger thalamic–PCC/precuneus connectivity drives chronic pain symptomology^4,41,42 and corresponding affective dysregulation.²⁵ Thus, the enduring chronic pain relief reported in response to mindfulness-based pain therapies^16,38 may reverse such aberrant signaling. These neurobiological findings are remarkably consistent with the central tenets of mindfulness (“non/not-self”; Pali: Anatta) to foster a nonreactive sense of self to alleviate suffering.^18,35 We suggest that mindfulness-based pain relief, after brief mental training, can significantly uncouple self-referential from nociceptive processes, an important finding for the millions of individuals seeking a fast-acting and nonpharmacologic pain treatment.

Conflict of interest statement

The authors have no conflicts of interest to declare.

Supplemental video content

A video abstract associated with this article can be found at https://links.lww.com/PAIN/B679.