The neural underpinnings of third-wave cognitive behavioral therapies, such as mindfulness and acceptance-based interventions, have become a topic of burgeoning interest in recent years [1
]. Many studies have reported significant relationships between mindfulness and brain structures involved in attention regulation, body awareness, emotion regulation, and perspective taking. Such associations have been found when studying mindfulness as a trait [5
] or as a practice, in studies comparing meditators and non-meditators or in studies looking for brain changes following mindfulness-based interventions [12
]. Associations between mindfulness, brain structure (i.e. gray matter, white matter), and function have been found in multiple brain regions, including the anterior and posterior cingulate cortex/precuneus, insula, temporo–parietal junction, different areas in the prefrontal cortex, hippocampus, and the amygdala, suggesting an involvement of a large-scale neural network conveying the effects of mindfulness [3
By contrast, other psychological constructs also likely to underpin the effects of third-wave cognitive behavioral interventions on health have received considerably less attention. Psychological (in)flexibility (PF) is a key construct in Acceptance and Commitment Therapy (ACT) [15
] and mindfulness-based interventions and provides a framework for understanding the impact of chronic pain on an individual’s functioning [16
]. PF is defined as the ability to act effectively and in accordance with personal values even in the presence of difficult experiences such as negative thoughts and emotions or unpleasant body sensations (e.g. pain). This construct comprises six core processes, including mindfulness, acceptance, decentering, self as a context, accordance with personal values and committed action [15
PF appears to be an important contributor in multiple mental health and medical conditions [19
], including chronic pain [20
]. Many studies evaluating PF components have consistently found that these are highly related to functioning and quality of life in patients with chronic pain, with higher flexibility being related to better outcomes [16
]. Similarly, higher PF has also been found to be associated with less pain intensity and interference, less anxiety and depressive symptomatology, and improved physical and mental functioning in patients with FM [17
]. Furthermore, PF (and its individual components) have also been reported to mediate clinical changes in acceptance- and mindfulness-based interventions delivered to patients with chronic pain [22
] and FM [26
Although PF has been established as a relevant transdiagnostic mechanism and as a key variable in explaining the effects of third-wave cognitive behavioral interventions on health, studies assessing the brain structures underpinning this set of cognitive processes are lacking [25
]. The main objective of the present study was to explore the brain structural correlates of PF, as measured by the Psychological Inflexibility in Pain Scale (PIPS) [24
] in patients with FM. We used voxel-based morphometry, a well validated technique that facilitates exploration of volume changes throughout the brain, including cortical and subcortical areas [29
]. Here we focus on Gray Matter Volume (GMV), which is the most frequently evaluated brain structural correlate of mindfulness in cross-sectional designs (e.g., [10
]). We also examined how brain volume clusters emerging from this analysis related to other clinically relevant and third-wave cognitive variables (i.e., mindfulness and self-compassion) in FM. Finally, we also examined the effects of an intervention known to promote PF in patients with FM, namely mindfulness-based stress reduction (MBSR) [27
], compared to treatment-as-usual (TAU) on PIPS-related brain volumes, in a longitudinal study performed in a subset of participants. The clinical impact of the present research is two-fold: To broaden our knowledge on the biological correlates of cognitive processes (i.e., PF) relevant for chronic pain and offer potential new brain-based biomarkers underpinning PF that could in future be targeted with psychological interventions or be useful as prognostic/predictive factors in the field of precision medicine.
The present study explored the structural brain correlates of psychological inflexibility in pain in a sample of patients with FM. A cluster in the ventral part of the BNST was found to be significantly related with PIPS scores. Additionally, BNST volumes were also found to be related to clinical and third-wave cognitive measures which are known to be relevant in chronic pain and FM; more precisely (after applying multiple comparisons correction), highly significant associations between BNST volumes with functional impairment, depressive symptomatology, perceived stress, and the nonjudging mindfulness facet were observed. Regarding the longitudinal analyses, there was no differential effect of MBSR, compared to TAU, on BNST volumes. Participants showing higher increases in PIPS scores, indicative of a potentially negative impact on health, presented significant pre–post increases in BNST volumes, suggesting that changes in BNST volume across time may be related to changes in psychological inflexibility.
The BNST—part of “the extended amygdala”—is a key nucleus in the integration of autonomic and behavioral responses to stress; however, until recently it has been relatively overlooked compared to research examining amygdala functioning [48
]. BNST integrates information from several upstream sources and, via dense connections with the paraventricular nucleus of the hypothalamus (PVN) (the principal node of the hypothalamic–pituitary–adrenal axis in initiating stress-mediated cortisol responses) regulates neuroendocrine and behavioral responses to stress [49
]. The ventral part of the BNST has the largest density of noradrenergic fibers in the brain [50
] and most of the projections to PVN originate from this area [51
]. It has been proposed that the BNST extends the duration of fear response, enabling sustained vigilance (and so anxiety and perceived stress) characterized by temporally prolonged changes in arousal. This role contrasts with the amygdala, which regulates immediate fear responses to related to discrete and immediate threats [48
]. The BNST has also been implicated in behavioral changes associated with depression; in particular, BNST activity has been related to anhedonic behavior after chronic mild stress, whereas lesions in this structure block the onset of learned helplessness in preclinical models [52
Psychological inflexibility in pain, as measured by PIPS, comprises avoidance to pain [24
], which is known to be inherently linked to fear of movement [23
] and the anticipated threat value of intense pain. The ongoing relationship between fear of movement and pain avoidance results in perpetual vigilance and monitoring of sensations related to pain, promoting sensitization to low-intensity pain sensations, reducing mobility, and, ultimately, reduced function [53
]. Regarding the implication of the BNST in the fear-avoidance model in chronic pain, a recent study by Meier et al. [54
] reported that kinesiophobia is associated with brain activation (i.e., fMRI) in the BNST in patients with chronic low-back pain, compared to pain-free matched controls. In accord with these findings, our results support the notion of a role for BNST in conveying psychological inflexibility in chronic pain. Our finding of a positive association between BNST volume and self-reported perceived stress and depressive symptomatology, provides further support for a putative role for BNST in both stress response and allostasis [48
We identified that BNST volumes positively correlated with the impact of FM, suggesting a potential role of this structure in the pathophysiology of the syndrome. BNST volumes also correlated with pain catastrophizing, a construct previously linked to pain intensity, depressive symptoms, disability, and delayed recovery in patients with chronic pain [56
]. Pain catastrophizing can be conceptualized as a negative cognitive/affective response to anticipated or real pain; it comprises rumination upon, magnification of, and feelings of helplessness towards the pain experiences [57
]. Pain catastrophizing has a crucial role in the fear avoidance model as it fosters fear-avoidance behaviors, which in turn result in exacerbation of physical and mental symptoms [56
]. Alterations in GM morphology related to pain catastrophizing have been reported in patients with chronic pain, with associations found in brain areas involved in pain processing, emotion and motor activity, attention to pain, and top-down inhibition of pain, but not in BNST [59
]. Our study suggests that BNST functioning may also contribute to the etiology of pain catastrophizing. Whether the role of BNST is unique to FM or more generally to other chronic painful disorders remains to be determined.
We also observed a significant negative association between BNST volumes and Nonjudging scores, suggesting that this structure may be involved in the neural basis of mindfulness. Given that mindfulness involves openness to the present experience (whether positively or negatively valent), a diminished role for brain areas implied in fear-anxiety-avoidance response might be expected in individuals with higher mindfulness scores. Mindfulness is a core component of psychological flexibility [15
], accordingly, the association between Nonjudging scores and BNST volumes may also be due to a theoretical overlap between mindfulness and PIPS. The Nonjudging facet seems to be particularly related to the functional impact of FM [60
] (more so than the other FFMQ subscales) and represents the component of acceptance in mindfulness as assessed by the FFMQ [38
]. To the best of our knowledge, only one study has established before a relationship between this specific mindfulness facet and brain structure; more precisely, a positive association was found between Nonjudging and surface area in the superior prefrontal cortex in a non-clinical young sample [11
], an area linked to self-referential processing, through introspection and self-awareness. In another study, dispositional mindfulness (assessed with the Mindfulness Attention and Awareness Scale) was negatively associated with amygdala volumes, a structure which—as we explained above—is closely related to the BNST [10
Although reductions in BNST could be expected after the MBSR program, since this program promotes psychological flexibility [27
], no effects were observed after the intervention. This null finding accords with previous studies looking for brain structural changes after MBSR; none reported changes in BNST [12
]. Unexpectedly, the effect size of the intervention on PIPS scores was smaller than expected in the present sample, as significant changes with moderate-to-large effect sizes (d = 0.70) were found in the EUDAIMON study when including the whole study sample (n
= 225) [27
]. Here, PIPS scores had only a small-to-moderate effect size, (d = 0.42). It remains to be determined whether a more effective intervention (on PIPS scores) may have a greater impact on BNST volumes assessed post-intervention or at a later follow-up assessment. Interestingly, when we further explored changes in BNST volumes and their relation to changes in PIPS scores, we found that those patients experiencing higher increases in PIPS scores (i.e., increasing their psychological inflexibility), also showed increases in BNST volumes. This finding provides further support to the hypothesis that both PIPS and BNST are intrinsically related, perhaps even across time, and that an effective intervention that improves psychological flexibility may also affect BNST volumes.
Coming back to the lack of effect of MBSR on BNST volumes, it is worth bearing in mind that variations in GMV may not necessarily represent changes at the microstructure and cellular level [62
], so we must not overlook that only part of the effects of the intervention on brain structure can be evaluated with the present design. Effect on other structural brain parameters (for example, white matter integrity), functional activity and connectivity of the BNST should also be evaluated in further better powered studies for comprehensively determine the effects of mindfulness on this specific brain cluster. Such studies also provide the desirable opportunity to examine structure–function relationships in BNST and their modulation by treatment.
Further research with larger samples is needed to confirm the role of the BNST in explaining PF and other key psychological constructs in patients with chronic pain conditions. Future studies should also ascertain whether the relationship between PF and BNST is exclusive of patients with FM or if this association is transdiagnostic. It is however surprising that the only brain cluster related to PIPS was found in the BNST and not in other brain areas frequently associated with PF-related constructs such as mindfulness facets or acceptance, for example, the anterior and posterior cingulate cortex/precuneus, insula, temporo–parietal junction, prefrontal cortex, hippocampus, and amygdala). Further studies with other clinical and non-clinical populations should also explore whether additional brain areas could also be related to PF. Likewise, case-control studies assessing whether clinical samples with greater psychological inflexibility (e.g. fibromyalgia) show indeed increased BNST volumes would be also welcome. Finally, longitudinal neuroimaging studies including interventions specifically targeting PF, such as Acceptance and Commitment Therapy, should also evaluate changes in BNST volumes to provide further evidence on the relevance of this area in conveying the effects of psychological treatments in chronic pain conditions.
Several limitations of the present study should be acknowledged. Firstly, the relatively small sample size limited our sensitivity to detect significant correlations, both at baseline and when seeking effects of treatment. Secondly, in our sample, MBSR did not produce a significant improvement in PIPS scores, reducing the likelihood of identifying corresponding changes in BNST volumes. Although the study focused in exploring neural correlates of PF in a sample of patients with FM, further studies should also include pain-free and/or clinical control participants to determine whether the relationship between BNST volumes and PF could be generalized to other chronic pain conditions or even to pain-free healthy individuals.