Modulatory effects of actual and imagined acupuncture on the functional connectivity of the periaqueductal grey and ventral tegmental area

Abstract

Objective:

Both acupuncture and guided imagery hold promise for treating pain. The mechanisms underlying these alternative interventions remain unclear. The reported study aimed to comparatively investigate the modulation effect of actual and imagined acupuncture on the functional connectivity of descending pain modulation system and reward network.

Methods:

Twenty-four healthy participants (mean [SE], 25.21 [0.77] years of age; 66.67% female) completed a crossover study that included five sessions, a training session and four intervention sessions administered in randomized order. We investigated the modulation effect of real acupuncture, sham acupuncture, video-guided acupuncture imagery treatment (VGAIT) and VGAIT control on the resting state functional connectivity (rsFC) of periaqueductal gray (PAG) and ventral tegmental area (VTA). These are key regions of the descending pain modulatory system (DPMS) and dopaminergic reward system, respectively.

Results:

Compared with sham acupuncture, real acupuncture produced decreased PAG-precuneus (Pcu) rsFC and increased VTA-amygdala/hippocampus rsFC. Heat pain threshold changes applied on the contralateral forearm were significantly associated with the decreased PAG-Pcu (r=0.49, p=0.016) and increased VTA-hippocampus rsFC (r=−0.77, p<0.001).

Compared to VGAIT control, VGAIT produced decreased PAG-paracentral lobule (PCL)/posterior cingulate cortex/Pcu, middle cingulate cortex (MCC), and medial prefrontal cortex rsFC, and decreased VTA-caudate and MCC rsFC.

Direct comparison between real acupuncture and VGAIT showed that VGAIT decreased rsFC in PAG-PCL/MCC, VTA-caudate/anterior cingulate cortex/nucleus accumbens, and VTA-MCC.

Conclusions:

Results suggest that both actual and imagined acupuncture can modulate key regions in DPMS and reward networks, but through different pathways. Identification of different pain-relief mechanisms may facilitate the development of new pain management methods.

Introduction

Pain can affect both physical and psychological well-being and is a major reason for seeking medical care (1,2). Opioid medications have important pain-relief effects, but also have the potential to become addictive and lead to abuse, making the management of pain a major challenge in global health (3). Due to the challenges associated with the pharmacological treatment of pain, particularly chronic pain, integrative health approaches have drawn increased attention from researchers and the public (4).

Acupuncture is a traditional therapeutic approach that stimulates certain points of the body with needles and has proven effective for the treatment of many pain disorders (5). Mental imagery can influence psychologic and physiologic states; its efficacy for pain management is also supported by a large body of research (6–9). Literature suggests that the brain responds to imagined experiences in a similar way to actual experiences (10–17). For example, studies have shown that particular overlapping brain regions, such as the insula, middle cingulate cortex, dorsal anterior cingulate cortex, and periaqueductal grey, are active when one experiences (18–21) or visualizes (22) acupuncture treatment. Therefore, imagined acupuncture needle stimulation may be able to reactivate some overlapping brain pathways that are activated during real acupuncture treatment.

Based on above findings, we have recently developed a complementary intervention, viz., video-guided acupuncture imagery treatment (VGAIT) that incorporates acupuncture and guided imagery to activate / deactivate brain regions that can also be activated / deactivted by real acupuncture. As a support, our previous studies have shown that VGAIT may produce analgesic effects comparable to acupuncture in both healthy controls (23) and chronic low back pain patients (24).

The underlying neural mechanisms of acupuncture and VGAIT for pain relief remain unclear. In a previous study (23), we found that acupuncture needle stimulation and VGAIT activated and deactivated widespread brain areas. Some brain areas, such as insula and anterior cingulate cortex, are associated with analgesia evoked by both acupuncture and imagined acupuncture. In another study, we found that real acupuncture decreased regional homogeneity (ReHo) and functional connectivity (FC) in sensorimotor areas, whereas VGAIT increased ReHo in basal ganglia (BG) (i.e., putamen) and FC between the BG subcortical network and default mode network (25). Nevertheless, the post-treatment effects, particularly that of VGAIT, and how these treatments modulate pain sensitivity remain unclear.

Recently, two brain systems, an opioidergic descending pain modulatory system (DPMS) and dopaminergic reward / motivation system, have drawn the attention of investigators. Findings from both animal and human studies suggest that these two systems play key and interacting roles in relieving pain (26–30). In a recent study (31), we found that both physical and mindful exercises can modulate resting functional connectivity (rsFC) of periaqueductal gray (PAG), which is a key region in DPMS, as well as that of the ventral tegmental area (VTA), which is the origin of dopaminergic projections, to relieve osteoarthritic knee pain. Studies from our group and others have found that acupuncture can modulate the functional connectivity of periaqueductal gray (PAG), as well as of the reward network (32–35).

The work reported here investigated the modulatory effects of acupuncture and VGAIT on the DPMS (using PAG as a seed) and on the reward system (using VTA as a seed). Specifically, we collected resting-state fMRI and behavioral pain threshold data before and after each of four interventions (i.e., acupuncture, sham acupuncture, VGAIT, and VGAIT control) from healthy individuals. We estimated and compared the rsFC of PAG and VTA before and after the interventions, and correlated connectivity patterns with corresponding analgesic effects. We hypothesized that both acupuncture and VGAIT would modulate the functional connectivity of PAG and VTA to produce pain relief and that they would evidence both common and distinct pathways.

Materials and methods

Full details of the experiment and procedure are reported in a previous publication (23), in which we investigated fMRI signal changes evoked by acupuncture needle stimulation and VGAIT administration. The work reported herein investigates the modulation effect of actual acupuncture and VGAIT on PAG and VTA rsFC before and after the interventions, which has not been previously published.

Participants

Twenty-seven healthy, right-handed, and acupuncture-naive individuals (19–33 years of age; 18 females and 9 males) were recruited for the study between January 2017 and August 2017. The study was approved by the Partners Human Research Committee (IRB) of Massachusetts General Hospital (MGH). All participants provided written informed consent prior to participating in the study.

Study overview

Participants went through five experimental sessions. Session 1 was a training session. Intervention sessions (Sessions 2–5) consisted of pain-threshold assessments and fMRI recordings, during which the participants received one of the four interventions in random order: 1) real acupuncture, 2) sham acupuncture, 3) VGAIT, or 4) VGAIT control condition. Each intervention session was separated by a minimum of seven days to avoid residual effects of the previous intervention. During data acquisition, all study personnel, except the acupuncturist, were blinded with respect to the intervention condition.

Session 1: training and familiarity session

Session 1 provided training for the heat and pressure pain threshold assessments using Quantitative Sensory Testing (QST). During Session 1, participants received real acupuncture treatment, which lasted about five minutes. This treatment was videotaped for the VGAIT intervention. In addition, a cotton swab was used to touch non-acupuncture sites, which was videotaped for the VGAIT control condition. Finally, participants were introduced to the MGH Acupuncture Sensation Scale (MASS) (36), which was used to assess sensations experienced during the interventions.

Session 2–5: intervention sessions

Real acupuncture was applied on acupoints Spleen 6 (Sanyinjiao) and Spleen 9 (Yinlingquan) on the right leg. Sham/placebo acupuncture was applied at two sham points using a sham acupuncture needle (i.e. Streitberger needle), which has been validated by studies demonstrating that participants cannot distinguish between real and sham needling (37,38). For each participant, leg position, acupoint location, and needling parameters were kept constant. The total treatment time last about 25 minutes.

At the beginning of the VGAIT session, participants were given the following introduction to read: “You will see a video of acupuncture treatment being applied on your leg. Please focus on the needle manipulation and try to imagine there is an actual needle being placed into your leg at the same spot. You will find that you can actually feel the needle manipulation on your leg at the same spot as in the video, as well as some soreness and an aching, dull pain along with other sensations. It is very important that you stay focused and try to imagine the sensation of receiving acupuncture as vividly as you can. After the treatment, we will provide a scale to measure the sensations you felt during the video.” The VGAIT control condition was the same as VGAIT except that cotton swabs were used to repeatedly touch non-acupoints and gently rotated using the same procedure as for the real and sham acupuncture. The rationale for using a cotton swab applied to non-acupoints (VGAIT control) as the control condition for the real acupuncture needle applied to acupoints (VGAIT) is that VGAIT emphasizes “guided imagery,” during which participants focus on the video and imagine the intervention sensations as vividly as they can. The sensations introduced by real acupuncture needles and cotton swab can be easily imagined and distinguished by participants.

Following these instructions, VGAIT or the VGAIT control was applied. The length of intervention time for VGAIT and VGAIT control was identical to the real and sham acupuncture interventions. Participants were asked to complete the MASS to assess their deqi sensations experienced during each intervention. A full detailed description of the interventions can be found in our previous publication (23). Two resting-state fMRI scans, approximately nine minutes each, were performed before and after the interventions (Figure 1).

Figure 1. Experiment procedures.

A. Overview of experiment procedures. B. Intervention session procedures.

Pain threshold measurements

Two pain modalities (heat and pressure) were assessed using QST before and after the intervention in a quiet room outside the scanner. We chose two pain modalities because heat-evoked pain is predominantly mediated by C-fibers, whereas pressure-evoked pain is predominantly mediated by A-delta fibers (39,40). Pain threshold assessments of two locations (leg and arm/thumbnail) were conducted three times. The thermode (heat) and algometer (pressure) were repositioned between each threshold assessment. In this paper, we focus on distal pain thresholds in order to measure suprasegmental (general) analgesic effects produced by the different interventions (41). Contact heat stimuli were delivered using a PATHWAY CHEPS (Contact Heat-Evoked Potential Stimulator, Medoc Advanced Medical Systems) (42), and pressure pain thresholds were assessed using an algometer (43). A detailed description of the QST can be found in a previous study (23).

fMRI data acquisition

All fMRI data were acquired at the Massachusetts General Hospital Martinos Center for Biomedical Imaging with a 3-axis gradient head coil in a 3 Tesla Siemens MRI System. A high-resolution T1-weighted structural image was acquired by an isotropic multiecho MPRAGE pulse sequence for anatomic localization of significant signal changes. Functional MRI images were acquired using a gradient echo T2*-weighted pulse sequence [time repetition (TR)/time echo (TE)=2000/30ms, flip angle (FA)=90°, field of view (FOV)=192×192 mm, 48 AC-PC aligned slices, slice thickness = 3.0 mm with 0.6 mm inter-slice gap, 90 image volumes per slice, matrix size=96×96] and a 32-channel multiarray coil. During the resting-state fMRI, which lasted approximately nine minutes, participants were asked to keep their eyes open and to blink normally while looking at a darkened screen and to not think about any particular thing.

Data analysis

Analysis of demographic and neuropsychological rating data

The perceived pain threshold changes for each intervention were analyzed using a one-way repeated-measures analysis of variance (ANOVA). A separate ANOVA was conducted for each of the four pain threshold measures, respectively. When the treatment main effect was significant, post hoc (Bonferroni-corrected) pairwise comparisons were performed. Correlation analyses were conducted using the r program in JASP open-source statistical software (Version 0.8.1, http://www.jasp-stats.org) to examine associations amongst demographic, neuropsychological, and behavioral pain threshold data and rsFC for each intervention.

fMRI data preprocessing

Data and calculations of rsFC were preprocessed using the CONN toolbox version 18.b (http://www.nitrc.org/projects/conn). We used the default preprocessing pipeline for seed-to-voxel rsFC analysis. The specific steps were as follows: slice timing correction, head motion correction, skull-stripping using BET, co-registration of the anatomical image to the mean functional image, segmentation of the anatomical gray matter, white matter, and CSF, normalization to Montreal Neurological Institute (MNI) 152 standard template and smoothing with a 6-mm Gaussian kernel. Band-pass filtering was performed with a frequency window of 0.008–0.09 Hz.

Seed-based functional connectivity analyses

To investigate the role of opioid descending pain modulatory pathways and dopamine reward pathways in mediating analgesic effects of acupuncture and VGAIT, we selected PAG and VTA as seeds for the following rsFC analyses. The bilateral PAG with a 2-mm radius sphere (MNI coordinates x=6, y=−28, z=−4; x=−2, y=−28, z=−4) were selected based on the brain activation results during the VGAIT scan (p=0.001) (23) and created using the SPM Wake Forest University (WFU) Pickatlas toolbox (http://fmri.wfubmc.edu/software/pickatlas, version 3.0.5) (44). This seed was chosen because it provided the most robust PAG activation during all four interventions. The bilateral VTA with a 4-mm radius sphere (MNI coordinates x=−4, y=−15, z=−9; x=5, y=−14, z=−8) was defined based on a previous study (45).

In the first-level analysis, we produced a correlation map for each participant by extracting the BOLD time course separately from the bilateral PAG and VTA seeds and computing Pearson correlation coefficients between the time course in the PAG / VTA and every voxel of the whole brain. Correlation coefficients were Fisher transformed into ‘z’ scores to increase normality. Seed-to-voxel between-group connectivity change analyses were performed separately by pairwise t-test using CONN toolbox.

We used pairwise t-tests to compare functional connectivity between different interventions, e.g., real acupuncture and sham acupuncture. The resulting statistical map was thresholded at p<0.005 at whole-brain voxel level. Next, we applied two different strategies to correct for multiple comparisons in the statistical analyses. For the first strategy, we used false discovery rate (FDR) to correct multiple comparisons at the cluster level as p_FDR<0.05. For the second strategy, because previous studies suggest that the insula, medial prefrontal cortex (mPFC), anterior cingulate cortex (ACC), middle cingulate cortex (MCC), orbital prefrontal cortex (OPFC), amygdala (AMY), and hippocampus (HIP) are involved in reward and pain modulation networks (23,46–48), we were particularly interested in whether acupuncture and / or VGAIT could modulate these brain regions. Therefore, we predefined a mask consisting of these brain regions (ROIs) using the Anatomical Automatic Labeling tool (http://www.fil.ion.ucl.ac.uk/spm/ext/#AAL). To correct for multiple comparisons, Monte Carlo simulations using the 3dFWHMx and 3dClustSim (as part of the AFNI program, https://afni.nimh.nih.gov, released in July 2017) were applied for the ROIs (i.e. pain encoding and modulation regions listed above, where for each region, the minimum voxel size required for p<0.05 cluster level p-value correction is indicated as a k value in the results presented below) (46,49). All results derived from ROI analysis were followed by minimum k value.

Statistical Analysis’

To explore the association between the rsFC z-score changes and pain-threshold changes, we extracted the average z-score values of the significantly altered rsFC clusters before and after the intervention (post- minus pre-) for real versus sham acupuncture, and for VGAIT versus VGAIT control. Correlation analysis was used to examine associations between the significantly altered rsFC clusters and pain threshold changes. To reduce multiple comparisons, we selected one of the four pain-threshold measures which was modulated most by real acupuncture (i.e., heat pain threshold on arm) and VGAIT (i.e., pressure pain threshold on thumbnail), respectively, for the correlation analysis; p values were Bonferroni corrected (see Results for details).

Results

Demographic and neuropsychologic results

Of the 27 participants enrolled, 24 participants (mean ± standard error [SE], 25.21±0.77 years of age; 16 female) completed the study and were included in the analyses. We found that the heat and pressure threshold changes were significantly different among the four interventions. Real acupuncture and VGAIT significantly increased participants’ pain thresholds, compared to sham acupuncture and VGAIT control. Detailed pain threshold results can be found in our previous publication (23) and in Table 1.

Table 1.

Pain threshold changes between post- and pre-intervention

Pain threshold (N=24)	heat arm pre	heat arm post	heat arm diff*	heat leg pre	heat leg post	heat leg diff*	pressure leg pre	pressure leg post	pressure leg diff*#	pressure nail pre	pressure nail post	pressure nail diff*#
real acu	41.04±0.64	42.25±0.58	1.21±0.24	42.76±0.71	43.73±0.70	0.97±0.36	6.10±0.45	6.84±0.38	0.74±0.27	5.99±0.37	7.13±0.37	1.14±0.17
sham acu	40.61±0.57	40.79±0.61	0.18±0.29	42.49±0.60	42.39±0.63	−0.10±0.29	5.56±0.41	5.29±0.37	−0.28±0.19	5.65±0.40	5.42±0.33	−0.22±0.18
VGAIT	41.31±0.69	41.88±0.68	0.57±0.33	42.44±0.73	43.31±0.77	0.87±0.34	5.38±0.36	6.38±0.39	1.00±0.20	5.49±0.33	6.61±0.34	1.11±0.22
VGAIT control	41.01±0.67	41.20±0.65	0.19±0.17	42.04±0.67	42.26±0.74	0.21±0.32	5.63±0.37	5.26±0.38	−0.36±0.10	5.87±0.32	5.81±0.36	−0.06±0.14

Notes: Pain thresholds reported using mean±standard error (SE). Comparison between real and sham acupuncture,

^*p<0.05 pairwise t-test; comparison between VGAIT and VGAIT control,

^#p<0.05 pairwise t-test. Abbreviations: real acu, real acupuncture; sham acu, sham acupuncture; VGAIT, video-guided acupuncture imagery treatment; heat arm, heat pain threshold on arm; heat leg, heat pain threshold on leg; pressure leg, pressure pain threshold on leg; pressure nail, pressure pain threshold on nail; pre, pre-intervention; post, post-intervention; diff, pain threshold change (post- minus pre-intervention).

Imaging results

PAG resting-state functional analysis results

Compared with sham acupuncture, real acupuncture produced significant rsFC decreases between the PAG and bilateral precuneus (Pcu) (Table 2, Figure 2A). Comparison between VGAIT and VGAIT control revealed that VGAIT significantly decreased the PAG rsFC with right paracentral lobule (PCL), posterior cingulate cortex (PCC), bilateral Pcu, MCC (k=22) and right mPFC (k=18) (Table 2 and Figure 2C); and increased PAG rsFC with visual associated brain regions.

Table 2.

Comparisons of the PAG based rsFC change between different interventions

Comparisons (N=24)	Brain Regions	Cluster	Peak	MNI coordinates
		Size	T	x	y	z
Sham acu > Real acu	Right Pcu	114	4.35	12	−62	30
	Left Pcu	132	3.94	−4	−70	46
Real acu > Sham acu	No regions survive the threshold
VGAIT > VGAIT control	Left IOG/OFG	195	3.67	−28	−94	−12
	Left OFG		3.43	−26	−88	−10
	Right OFG/IOG	152	4.81	22	−92	−14
	Right IOG		3.81	34	−92	−8
	Left LG	92	3.55	−12	−98	−14
	Left OP		3.47	−20	−100	0
	Left MOG	94	4.18	−38	−92	12
VGAIT control > VGAIT	Right PCL	438	6.17	8	−34	54
	Left PCC		4.58	−4	−30	42
	Right Pcu		3.41	10	−44	50
	Left Pcu		3.28	−12	−44	48
	Left MCC*	101	4.72	−4	22	28
	Right mPFC*	55	4.18	16	66	8
(Real acu - Sham acu)
>	Right PCL/MCC	190	4.53	8	−34	56
(VGAIT - VGAIT control)
(VGAIT - VGAIT control)	Left OP	426	5.24	−16	−104	−8
>	Left IOG		3.23	−26	−98	−8
(Real acu - Sham acu)	Right LG	157	4.38	6	−92	−12

Notes: We performed pairwise t-tests and report two-sided p values. Statistical maps were threshold at p < 0.005 at voxel level and p_FDR< 0.05 at the cluster level.

^*indicates results that are significant at cluster p < 0.05 after 3dFWHMx and 3dClustSim correction for pre-defined ROI. Other results are significant at cluster p_FDR< 0.05 corrected at the whole brain level. Abbreviations: acu, acupuncture; rsFC, resting state functional connectivity; VGAIT, video-guided acupuncture imagery treatment; PAG, periaqueductal gray; Pcu, precuneus; IOG, inferior occipital gyrus; OFG, occipital fusiform gyrus; LG, lingual gyrus; OP, occipital pole; MOG, middle occipital gyrus; PCL, paracentral lobule; PCC, posterior cingulate cortex; MCC, middle cingulate cortex; mPFC, medial prefrontal cortex.

Figure 2. PAG based connectivity results.

A. Compared to sham acupuncture, real acupuncture had significantly lower rsFC with Pcu. B. Positive correlation in sham versus real acupuncture was observed between the PAG-left Pcu rsFC and heat pain on arm. Bonferroni correction was applied, and the significance threshold was adjusted to p < 0.025 since two significant clusters were identified. C. Compared to VGAIT control, VGAIT had significantly lower rsFC with the MCC, PCL/PCC/Pcu and mPFC. D. Compared to VGAIT, real acupuncture had significantly higher rsFC with the PCL/MCC. Abbreviations: PAG, periaqueductal gray; Pcu, precuneus; PCL, paracentral lobule; PCC, posterior cingulate cortex; MCC, middle cingulate cortex; mPFC, medial prefrontal cortex.

We also compared PAG connectivity changes between real acupuncture and VGAIT ([real minus sham acupuncture] versus [VGAIT minus VGAIT control]), and found that, compared to VGAIT, real acupuncture produced greater rsFC increases between PAG and right PCL and MCC (Table 2, Figure 2D), as well as significant decreases between PAG and visual areas.

VTA resting-state functional analysis results

Using bilateral VTA as the seed, we found that real acupuncture increased rsFC between VTA and right amygdala (AMY), hippocampus (HIP) (Table 3, Figure 3A), and right middle / superior temporal gyrus (MTG / STG), and decreased rsFC between VTA and left supramarginal gyrus (SMG) and right middle / inferior frontal gyrus (MFG / IFG), compared to sham acupuncture (Table 3).

Table 3.

Comparisons of the VTA based rsFC change between different interventions

Comparisons (N=24)	Brain Regions	Cluster	Peak	MNI coordinates
		Size	T	x	y	z
Real acu > Sham acu	Right AMY	145	4.59	28	−4	−26
	Right HIP		4.31	24	−8	−22
	Right MTG	311	4.54	64	−16	−8
	Right STG		4.46	60	−18	−2
Sham acu > Real acu	Left SMG	182	3.94	−62	−40	40
	Right MFG	146	3.99	46	18	30
	Right IFG		3.50	48	30	18
VGAIT > VGAIT control	No regions survive the threshold
VGAIT control > VGAIT	Left CAU	406	5.18	−8	16	6
	MCC	166	4.55	0	−2	40
	Right SMA		3.31	10	−8	46
	Left SMA		3.15	−8	6	44
(Real acu - Sham acu) > (VGAIT - VGAIT control)	Right SPL	167	5.30	24	−44	66
	Left SMG	86	5.18	−48	−28	40
	Left PoCG		3.53	−38	−32	42
	Right CAU	546	4.12	6	14	2
	Right ACC		3.81	6	30	−10
	Right MFC		3.70	6	36	−20
	Right NAc		3.48	6	8	−8
	Right STG	119	3.57	54	−34	6
	Right MCC	149	4.29	8	4	38
	Right SMA		3.99	12	10	42
(VGAIT - VGAIT control)	No regions survive the threshold
>
(Real acu - Sham acu)

Notes: We performed pairwise t-tests and report two-sided p values. Statistical maps were threshold at p < 0.005 at voxel level and p_FDR< 0.05 at the cluster level. Abbreviations: acu, acupuncture; rsFC, resting state functional connectivity; VGAIT, video-guided acupuncture imagery treatment; VTA, ventral tegmental area; AMY, amygdala; HIP, hippocampus; MTG, middle temporal gyrus; STG, superior temporal gyrus; SMG, supramarginal gyrus; MFG, middle frontal gyrus; IFG, inferior frontal gyrus; CAU, caudate; MCC, middle cingulate cortex; SMA, Supplementary motor area; SPL, superior parietal lobule; PoCG, postcentral gyrus; ACC, anterior cingulate cortex; MFC, medial frontal cortex; NAc, nucleus accumbens.

Figure 3. VTA based connectivity results.

A. Compared to sham acupuncture, real acupuncture had significantly higher rsFC with right AMY/HIP. B. A significant negative correlation for real versus sham acupuncture was observed between the VTA-HIP rsFC and heat pain on arm. Bonferroni correction was applied and the significance threshold was adjusted to p < 0.0125 because four significant clusters were identified. C. Compared to VGAIT control, VGAIT had significantly lower rsFC with the MCC and CAU. D. Compared to VGAIT, real acupuncture had significantly higher rsFC with the MCC and CAU/ACC/NAc. Abbreviations: VTA, ventral tegmental area; AMY, amygdala; HIP, hippocampus; CAU, caudate; MCC, middle cingulate cortex; ACC, anterior cingulate cortex; NAc, nucleus accumbens.

Comparison between VGAIT and VGAIT control showed that VGAIT decreased rsFC between VTA and left caudate (CAU), MCC, and bilateral supplementary motor area (SMA), compared to VGAIT control (Table 3, Figure 3C).

Comparison between real acupuncture and VGAIT revealed that VGAIT induced significant rsFC decreases between the VTA and right superior parietal lobule (SPL), left SMG, postcentral gyrus (PoCG), right CAU, anterior cingulate cortex (ACC), medial frontal cortex (MFC), nucleus accumbens (NAc), right STG, MCC, and right SMA (Table 3, Figure 3D).

Association between rsFC differences and pain threshold changes

To explore the association between rsFC differences (between real versus sham acupuncture and between VGAIT versus VGAIT control) and corresponding pain threshold changes, we extracted the average z-values of the cluster and performed correlation analysis with corresponding pain threshold changes (post-intervention minus pre-intervention). We found a significant positive correlation between PAG - left Pcu rsFC and heat pain threshold changes administered to the arm for real versus sham acupuncture (r=0.49, p=0.016, Figure 2B; Bonferroni correction was applied, and the significance threshold was adjusted to p < 0.025 because two significant clusters were identified (see Table 2 for details). A significant negative correlation was observed between VTA - HIP rsFC and heat pain administered to the arm in real versus sham acupuncture (r=−0.77, p<0.001, Figure 3B, Bonferroni correction was applied, and the significance threshold was adjusted to p<0.0125 because four significant clusters were identified, see Table 3 for details).

Association between rsFC differences and imagery vividness

In this study, we measured participants’ imagery vividness using Betts’ Questionnaire upon Mental Imagery (BQMI) (50). The BQMI mean rating (mean ± SE) was 81.13 ± 4.13. We also explored the association between rsFC changes and the imagery vividness. No significant association was found between BQMI rating and PAG / VTA rsFC change differences between the VGAIT and VGAIT control groups (p values ranging from 0.06 to 0.96).

Discussion

The present study found that real, compared with sham, acupuncture decreased rsFC between PAG and Pcu, and increased rsFC between VTA and AMY and HIP. VGAIT significantly decreased rsFC between PAG and paracentral lobule, PCC, Pcu, mPFC, and MCC, as well as between VTA and CAU and MCC, compared to VGAIT control. Furthermore, compared to real acupuncture, VGAIT significantly decreased PAG rsFC with paracentral lobule and MCC, and VTA rsFC with CAU, ACC, NAc, and MCC. Our findings suggest that analgesic effects may be modulated by distinct mechanisms when coupled with different intervention modalities and support important roles for DPMS and the reward system in acupuncture and VGAIT analgesia.

Pain disorders are highly prevalent and disabling with unsatisfactory treatment options (51). Despite the controversies regarding the effectiveness of opioids for relieving pain, the prescription rates of opioids have increased in the USA (52), accompanied with increased opioid addiction and overdoses (53,54). Recently, acupuncture has been recommended as a first-line option for treating pain disorders, but is limited by its cost and inconvenience (55). The COVID-19 pandemic further limited the availability of acupuncture. The advantages of VGAIT are 1) it does not require involvement of the acupuncturist and can be administered by staff (or even patients after training); and 2) it is cost-effective and can be easily applied remotely as part of e-medicine. These characteristics support the potential benefits of VGAIT as an effective pain management method (23,24).

In this study, we comparatively investigated the modulatory effects of different treatment modalities on PAG / VTA rsFC. Real and sham acupuncture produce specific somatosensory afferences as caused by the real and sham acupuncture needle stimuli, respectively. An important difference is that real acupuncture stimulates muscle and tendon, whereas sham acupuncture needle primarily stimulates skin. Although there is no direct somatosensory afference, VGAIT and VGAIT control contain visual sensory afferences as well as mental imagery of sensory stimulation. Interestingly, we found VGAIT and VGAIT control can produce mild deqi sensations, which is believed to be a key for acupuncture treatment. This also suggests that somatosensory sensations may be produced by visual needle execution and / or sensory imagery without direct peripheral somatosensory stimulation, which can further modulate multiple brain networks including the descending pain modulation and reward networks.

We found acupuncture can significantly modulate PAG functional connectivity, which is consistent with previous findings suggesting an important role of the descending pain modulation system in acupuncture analgesia (5,56,57). Specifically, we found that real acupuncture decreased rsFC between PAG and Pcu, a key region of the default mode network (DMN). Studies suggest that the DMN can modulate pain perception of acute noxious stimuli experienced by oneself or when observing another person through autonomic and antinociceptive DPMS (58,59). Research suggests that the DMN may be a key network affected by chronic pain (60,61). Our results suggest a possible neural substrate that may be involved in acupuncture treatment of chronic pain.

We also found that, compared to VGAIT control, VGAIT induced a significant decrease in rsFC between PAG and mPFC. Findings from previous studies suggest that vivid imagining while thinking about oneself or a close friend can activate mPFC (62,63). mPFC is an important region that is activated during imagining (64,65) and was likely activated when our participants vividly imagined acupuncture sensations while watching the video of the acupuncturist manipulating needles on their leg. The decreased rsFC between PAG and mPFC may indicate that VGAIT modulated participants’ perception and mentalizing. We also found that VGAIT produced decreased rsFC with PCL and MCC, compared to real acupuncture. These significant connectivity changes suggest the possibility that VGAIT may modulate key regions involved in the somatosensory system.

Accumulating evidence suggests that the reward system also plays an important role in pain modulation (29). The dopamine reward network includes neurons of VTA and their projections to many brain regions (66), which can modulate the perception of nociceptive information, the affective symptoms of chronic pain, as well as the efficacy of pain medications (67–69).

The limbic system, which contains many of the brain areas underlying emotion and reward, includes the hippocampus and amygdala and their interconnections. We found that rsFC between VTA and limbic regions (HIP, AMY) increase in response to real, compared to sham, acupuncture. This result is consistent with a recent finding indicating that acupuncture can increase VTA rsFC with amygdala in chronic low back pain (35). It is possible that this increase in mesocorticolimbic connectivity represents the activation of VTA dopaminergic projections to the limbic system during acupuncture analgesia. This modulation effect of acupuncture treatment on the dopamine reward system is also consistent with previous human and animal studies (34,70).

We also found that rsFC between VTA and MCC and CAU were significantly decreased after VGAIT, compared to real acupuncture and VGAIT control. The CAU is a key structure that makes up the dorsal striatum, which is also part of the reward system (71). Task-related fMRI studies have demonstrated that CAU is activated during reward-based learning, good decision-making, and during listening to highly pleasurable musical pieces (72,73). Accumulating findings indicate that the reward system is involved in imagery processing (74,75), our result is consistent with these findings.

Studies suggest that heat-evoked pain is predominantly mediated by small, nonmyelinated peripheral nociceptive nerve fibers (C-fibers), while pressure-evoked pain is predominantly mediated by small, myelinated peripheral nociceptive nerve fibers (A-delta fibers) (39,40). Our previous study revealed that real acupuncture produced segmental (leg) and suprasegmental (arm/thumbnail) heat and mechanical pain threshold increases, while VGAIT produced segmental (leg) and suprasegmental (thumbnail) mechanical pain threshold increases and only a trend for segmental heat pain threshold, in contrast to the sham and VGAIT control interventions. This suggests that both real acupuncture and VGAIT can increase participants’ pain thresholds, thereby achieving analgesic effects. However, VGAIT may be more effective on the A-delta fibers (23). Acupuncture has been shown to induce an ascending antinociceptive response in the orofacial formalin test in rodents, which can be blocked by sciatic C-fiber depletion (76). In the current study, we found that acupuncture-induced general heat pain threshold changes were associated with decreases in rsFC between PAG and Pcu, as well as an increase in rsFC between VTA and HIP. These findings suggested a role for the endogenous opioid pathway and mesocorticolimbic reward pathway in acupuncture analgesia on C-fibers. Further study is needed to verify these findings.

In this study, the video played in VGAIT sessions for each participant was real acupuncture administered on each participant’ own body. Literature suggests that observation of oneself or others’ experience can be distinguished and that self-observation ability varies significantly across contexts (77,78). Further study is needed to investigate effects of VGAIT using the acupuncture videos administered to another person’s (third party) body.

Limitations of the current work include a relatively small sample size, necessitating that the findings be validated in larger scale studies. Also, the current study only examined a single administration of each intervention and did not examine the extent to which pain relief persisted over weeks or months. Finally, we performed this study on healthy participants; it is unclear whether similar results and conclusions will extend to clinical pain populations.

In conclusion, we found that both acupuncture and VGAIT can modulate several key brain regions involved in the DPMS and reward / motivation systems, but that they might target different circuits. Elucidating different ways to modulate DPMS and reward networks could facilitate the development of new pain management therapies.

Acronyms:

VGAIT

video-guided acupuncture imagery treatment

Acu

acupuncture

rsFC

resting-state functional connectivity

PAG

periaqueductal gray

VTA

ventral tegmental area

DPMS

descending pain modulatory system

Pcu

precuneus

AMY

amygdala

HIP

hippocampus

PCL

paracentral lobule

PCC

posterior cingulate cortex

MCC

middle cingulate cortex

mPFC

medial prefrontal cortex

CAU

caudate

ACC

anterior cingulate cortex

NAc

nucleus accumbens

QST

Quantitative Sensory Testing

ANOVA

analysis of variance

SE

standard error

MTG

middle temporal gyrus

STG

superior temporal gyrus

SMG

supramarginal gyrus

MFG

middle frontal gyrus

IFG

inferior frontal gyrus

SMA

supplementary motor area

SPL

superior parietal lobule

PoCG

postcentral gyrus

MFC

medial frontal cortex

DMN

default mode network