Increased salience network connectivity following manual therapy is associated with reduced pain in chronic low back pain patients
Introduction
. Given its prevalence, combined with the lack of efficacious long-term treatments, there has been a growing interest in complementary non-pharmacological therapies to treat a variety of pain disorders
, including cLBP. Manual Therapy (MT) is a non-pharmacological therapy used to treat a range of neuromusculoskeletal and chronic pain disorders. In the last 15 years, there has been an increase in research assessing the effectiveness of MT for cLBP. A review published in 2004 included the results of 69 randomized controlled trials using MT, and found evidence that its effects are stronger than placebo and general care
. In a multi-site study conducted with 1,334 cLBP patients in the United Kingdom, those who were randomized to receive “best care” of general practice treatment in combination with spinal manipulation MT and exercise experienced a significant reduction in LBP intensity and disability scores at the three month and one year follow-up
. However, the mechanisms supporting MT are less clear and further research is needed.
. In fact, MT as a treatment for cLBP patients found that participants who received MT showed inhibited temporal summation to thermal pain post-treatment, concluding a central pain processing mechanism which suggests neuroplastic changes post-MT
. Alongside temporal summation, there are several other neurophysiological factors that have been cited as being impacted by MT (i.e. changes in pain sensitivity, somatosensory evoked potentials) suggesting a mechanism of response via the central nervous system
. Recently, neuroimaging research has suggested that cLBP is characterized by altered brain structure and function, suggesting a need for a better understanding of the underlying brain-focused neural mechanisms supporting non-pharmacological treatment modalities for this condition
;
. As such, there has been an increase in neuroimaging research to assess cortical alterations post-MT for pain. A small fMRI study assessing evoked-pain in ten healthy controls pre- and post- mid-thoracic spine thrust manipulation found reductions in brain activity in regions associated with pain processing (i.e. insular cortex, thalamus, S1, S2, etc.) as well as a significant reduction in perception of pain intensity of evoked stimuli
. In a larger cohort of experimentally-induced low back pain subjects, MT was found to immediately alter resting-state connectivity between brain regions implicated in sensory and affective components of pain processing
. Our group recently found that a single session of MT reduced the engagement of salience and social cognition brain circuitries in response to videos depicting exercises perceived as painful in cLBP patients
. This suggests that altered brain activity following MT may underlie its hypoalgesic effects and more neuroimaging research is needed to better elucidate the specific mechanisms. Furthermore, a large meta-analysis conducted on 129 Randomized Controlled Trials of MT found evidence of improvement for psychological variables following treatment (i.e. fear avoidance beliefs, depression)
. Therefore, MT, which also includes cognitive and affective components of therapy, may also impact cognitive and affective dimensions of pain, which should be further explored in mechanistic studies.
. Moreover, we found that and SLN connectivity can be modulated with maneuvers that exacerbate back pain
. Such functional connectivity analyses may also suggest how connectivity networks are modulated by MT, and which changes are associated with post-treatment reduction in back pain. Historically, the pain experience has been imaged via block-design evoked-experimental pain paradigms, which offer a way to assess hyperalgesia, an important characteristic of clinical pain. However, there are challenges that come with assessing clinical pain itself in a controlled setting, given the variability in clinical pain states from subject to subject and difficulty in experimentally controlling clinical pain severity. Our group has previously linked clinical pain intensity to intrinsic brain connectivity in chronic pain, including cLBP patients
;
;
. As such, our group and others have now shown that rs-fMRI can be a useful tool to measure the neural mechanisms underlying clinical pain and an important step in understanding the brain’s processing of chronic pain.
;
;
.
Results
[Table 1B]. ANOVA results showed a significant main effect for Time in clinical low back pain reduction (f=13.34, p=0.003), but no effect for Group x Time interaction (f=0.22 p=0.65), as low back pain intensity was reduced following both MANIP (Pre: 39.43 ± 16.5, post: 28.43 ± 16.5, p=0.005) and MOBIL (Pre: 38.83 ± 17.7, post: 31.76 ±19.4, p=0.02), while the magnitude of change did not differ between MT approaches [See Figure 2]. Additionally, a paired T-test assessing differences between pre-scan low back pain levels between the two different conditions (MANIP vs MOBIL) showed no significant differences (p=0.8). Patients ranged in low back pain bothersomeness, from 2 to 7 (0-10, visual analog scale), as rated for the week prior to the MRI session. Analyses of functional brain connectivity are highly susceptible to head motion effects, and average Root Mean Square (RMS) values were calculated for each subject and each scan. These values were used in a repeated measures ANOVA to assess potential differences in head motion between conditions. There were no significant effects of group (cLBP vs HC, p=0.37), time (post-MANIP vs pre-MANIP, p=0.69), or group x time interaction (p=0.83) for MANIP, and no significant effects of treatment (MANIP vs MOBIL, p=0.9), time (post vs pre, p=0.3), or treatment x time interaction (p=0. 7) for cLBP patients. These results suggest that head motion did not contribute significantly to our rs-fMRI connectivity findings.
Table 1BMean cLBP subject scores for expectancy for relief and credibility of treatment for both MANIP and MOBIL.
Figure 2Low back pain intensity in cLBP patients was reduced following high MANIP (p=0.005, left) and MOBIL (p=0.02, right).
Table 2ARegions showing significant SLN connectivity changes post- vs. pre-MANIP and their corresponding Z-stats for post- vs. pre-MANIP and post- vs. pre-MOBIL in cLBP patients, and for post- vs. pre-MANIP in HC.
Figure 3Whole brain results of SLN connectivity Post- vs. Pre- MANIP in cLBP patients (left) and plots showing M1 and thalamus ROIs for MANIP and MOBIL in cLBP and MANIP in HC (right).
Figure 4rmANOVA plots of SLN connectivity Post- vs. Pre-MANIP and MOBIL in cLBP patients and Post- vs. Pre-MANIP in HC in M1 (left) and Thalamus (right). ROIs were selected based on the intersection of the somatomotor network of the Yeo et. al. (2011) brain atlas and the Harvard-Oxford probabilistic atlas (Desikan, 2006).
Figure 5Whole-brain linear regression voxelwise analysis using post-intervention clinical pain change found that greater increase in SLN connectivity to lateral prefrontal cortex (LPFC) post- vs. pre-MANIP was associated with greater reduction in low back pain. A similar association was found for post-MOBIL in the same patients (bottom right).
Table 2BRegions showing significant DMN connectivity changes post- vs pre-MANIP and their corresponding Z-stats for post- vs pre-MANIP and post- vs pre-MOBIL in cLBP patients, and for post- vs pre-MANIP in HC.
Conclusions
;
. The function of the SLN is in fact essential in the processing of sensory stimuli, as the SLN plays a key role in the assessment of the inherent danger of such stimuli and how one should respond to them
. Therefore, altered response in this network could be a contributing factor to the maintenance and chronicity of pain. Our results demonstrated that cLBP patients do in fact display unique SLN in comparison to healthy controls and that a single session of grade V manipulation immediately increases the integration of the SLN with both the thalamus and the motor cortex (M1).
;
. It has implications in both sensory and cognitive components of pain processing, which contribute to the modulation of pain via an integration of sensory discrimination, attention, memories, etc.
. Additionally, the thalamus plays a role in adaptive salience responses to sensory experiences
. Therefore, post-MT increase in SLN to thalamus connectivity could reflect the modulation of brain responses in cLBP patients on both a sensory and a cognitive-affective dimension. For instance, dysrhythmia in thalamocortical connectivity has been suggested to support chronic, particularly central, pain
. Our study found that SLN connectivity was upregulated to a cluster spanning the medial dorsal (MD), Ventral Posterior Lateral (VPL), and Ventral Lateral Posterior (VLp) nuclei of the thalamus
[Figure 6]. The MD nucleus of the thalamus has been linked with the SLN
and plays a role in both limbic/emotional and cognitive (i.e. memory) processing via communication with the Prefrontal Cortex
;
;
. The VLp and VPL thalamic nuclei are known to be involved with both somatosensation and passive movement. Specifically, the VPL nucleus is important for relaying somatosensory afference to the primary somatosensory cortex (S1)
;
;
, which is consistent with somatosensory aspects of the MT intervention. The VLp, in turn, is located slightly dorsal to the VPL nucleus and has been shown to activate with passive joint movement (e.g. MT), acting as a relay center between the cerebellum and motor cortex
;
. Thus, MT may contribute to somatosensory, affective, and cognition/memory components of pain processing by modulating salience processing and attentional resource allocation via SLN connectivity to the MD, VPL, and VLp nuclei of the thalamus.
;
. In fact, the motor cortex plays a role in controlling postural changes, and cLBP patients exhibit abnormal postural control, which could be due to the reorganization of this cortex
. Several MRI studies have shown altered structure and function of M1 in chronic pain patients
;
;
. Furthermore, research assessing cortical thickness in the motor cortex post-treatment showed improvements related to both decreased pain intensity and physical disability
. Previously our group found that evoked pain modulated SLN connectivity to M1, and that sensitivity to painful stimuli was highly correlated with SLN to M1 functional connectivity changes
. Therefore, it is possible that the somatosensory component of MT, particularly of MANIP, helps alter the top-down processing of pain, immediately following the manipulation, through increased SLN connectivity to M1. Additionally, in our previous publication on this cohort, we found that both MANIP and MOBIL reduced patient’s ratings of movement-expected pain associated with the performance of back straining exercises
. Therefore, increased SLN to M1 connectivity may be related to decreased expectation of movement-induced pain.
, and has been shown to have an impact on negative cognitive processes in clinical pain populations
;
;
. Our previous results suggested a biopsychosocial mechanism underlying MT treatment for cLBP, as patients showed a post-treatment reduction in both clinical pain and fear of “back-straining” exercises, which were related to decreased BOLD response in various brain regions associated with emotion, cognition, and pain perception
. Thus, our results suggest that SLN connectivity to LPFC supports MT-induced analgesia, which could be due to cognitive top-down processing of pain via anti-nociceptive pathways.
Our results showing both somatosensory and cognitive changes post-MT implicate the importance of elucidating the psychophysiological underpinnings of this therapeutic method in cLBP patients. The single-session modulation of SLN connectivity to Thalamus and M1 suggest an immediate effect of MT on a physiological level, however the reduction in clinical pain was not associated with this change. A longitudinal trial is necessary to further assess associations between clinical pain reduction and somatosensory regions such as M1 and Thalamus. However, the association with post-MT decrease in clinical pain and increased activation in LPFC suggest that higher-level cognitive processing plays an important role in the reduction of clinical pain immediately following a single session of Manual Therapy.
. Therefore, both manipulation and mobilization play an analgesic role in the perception of low back pain. There are several challenges that still remain and need to be further examined. Firstly, our design utilized a between-groups control for only spinal manipulation, and not mobilization, which limits our capacity to fully compare the two conditions across cLBP and HC groups. “Furthermore, our analysis used a lower cluster-forming threshold (Z>2.3) than the recently recommended standard (Z>3.1). However, given the statistical power inherent to our sample size, we felt that our choice for thresholding was acceptable as it still complies with the assumptions underlying Gaussian Random Field theory. Nevertheless, our study was novel in design, and further research with larger cohorts may be needed in order to replicate our results.
future analyses could attempt to link autonomic modulation by MT with altered connectivity for specific SLN subregions. Moreover, our ROI analyses focused on the results contrasted from post- vs pre- MANIP in our cLBP patient cohort, which could have biased the results towards both the patient population and the manipulation technique. Future studies should increase power in order to focus more on interactions between both groups and techniques at a voxelwise level. A larger powered study could better assess various aspects of clinical effectiveness in both MANIP and MOBIL. Dynamic connectivity models may prove useful, as superior temporal resolution with multiband accelerated fMRI data would allow for better linkage with time series of autonomic function.
In conclusion, this study found that MT’s capacity to immediately reduce clinical low back pain, specifically after manipulation, may operate via modulation of SLN functional brain connectivity. Additionally, this modulation of salience connectivity occurs in regions associated with cognitive, affective and sensorimotor components of pain processing (i.e. thalamus, M1). Furthermore, our result showing an association between increased SLN connectivity to the LPFC for both MANIP and MOBIL suggests that MT-induced hypoalgesia may also be sensitive to higher-order cognitive processing.
Article Info
Publication History
Accepted:
November 24,
2020
Received in revised form:
October 29,
2020
Received:
June 22,
2020
Publication stage
In Press Journal Pre-Proof
Footnotes
Acknowledgement: This research was supported by the NCMIC Foundation, Inc. (M.L.L.), Norwegian Research Council / Marie Sklodowska-Curie Actions ( FRICON/COFUND – 240553/F20 to D.M.E.), the National Center for Research Resources ( P41RR14075 ; CRC 1 UL 1 RR025758 , Harvard Clinical and Translational Science Center); Martinos Computing facilities; NIH S10RR023401 ; S10RR019307 ; S10RR019254 ; S10RR023043 . 6
None of the authors have any conflicts of interest to declare.
Identification
Copyright
© 2020 by United States Association for the Study of Pain, Inc.
Comments are closed.