The impact of mindfulness-based interventions on brain activity: A systematic review of functional magnetic resonance imaging studies

Highlights

• Systematic review of the impact of mindfulness based interventions on brain activity.

• We found changes in insula reactivity following 8 week interventions.

• Included studies were with heterogeneous populations and tasks.

Abstract

Mindfulness-based interventions are increasingly used in the treatment and prevention of mental health conditions. Despite this, the mechanisms of change for such interventions are only beginning to be understood, with a number of recent studies assessing changes in brain activity. The aim of this systematic review was to assess changes in brain functioning associated with manualised 8-session mindfulness interventions. Searches of PubMed and Scopus databases resulted in 39 papers, 7 of which were eligible for inclusion. The most consistent longitudinal effect observed was increased insular cortex activity following mindfulness-based interventions. In contrast to previous reviews, we did not find robust evidence for increased activity in specific prefrontal cortex sub-regions. These findings suggest that mindfulness interventions are associated with changes in functioning of the insula, plausibly impacting awareness of internal reactions ‘in-the-moment’. The studies reviewed here demonstrated a variety of effects across populations and tasks, pointing to the need for greater consistency in future study design.

Introduction

The popularity of mindfulness interventions has exploded over the past decade. Used both to prevent and treat, mindfulness programmes can be found in various healthcare settings, the workplace as well as in schools and community organisations. Despite this surge in uptake, scientific understanding of the mechanisms of mindfulness interventions is in the early stages. There have been numerous calls for the use of neuroscience to better understand the mechanisms underlying effective psychological treatment (Craske, 2014, Holmes et al., 2014) and investigation of the neural mechanisms associated with effective mindfulness interventions is on the rise. Neuroimaging offers a tool for: i) expanding our understanding of why and how treatments work, ii) providing insight into adaptations of existing treatments to better target underlying mechanisms and iii) to assess factors associated with an individual’s likelihood of responding to a particular treatment (Wise et al., 2014).

Mindfulness involves practicing the awareness of ‘in-the-moment’ experiences, as well as an attitude of non-judgemental acceptance of current experiences (Kabat-Zinn, 2013). The overarching aim of mindfulness practice is to “maintain awareness of moment-to moment experiences, disengaging oneself from strong attachment to beliefs, thoughts or emotions thereby developing a greater sense of emotional balance and well-being” (Ludwig and Kabat-Zinn, 2008, p.1350). Two standardised forms of mindfulness training, mindfulness-based stress reduction (MBSR) and mindfulness-based cognitive therapy (MBCT) have garnered most support for their efficacy. MBSR and MBCT are manualised, group-based 8-week training programs that teach mindfulness skills through a range of formal and informal practices.

MBSR was originally developed in 1979 as a mental health training intervention for chronic health conditions (Kabat-Zinn, 2013). It has been shown to be effective for a range of conditions with meta-analyses demonstrating efficacy in the treatment of anxiety (effect size, Cohen’s d = 0.47), depression (0.26), pain (0.33) and stress (0.55; Bohlmeijer et al., 2010, Khoury et al., 2015). MBCT is an adaptation of MBSR, developed as a preventative treatment for recurrent depression (Segal et al., 2002). It combines psychological educational components of cognitive behavioural therapy (CBT) for depression with meditation elements of MBSR (Chiesa and Malinowski, 2011, Williams et al., 2014). MBCT has been shown to halve the risk of relapse (Kuyken et al., 2016) and is currently a recommended treatment for relapse prevention of recurrent depression in a number of national clinical guidelines, such as the UK National Institute for Health and Clinical Excellence (NICE).

While evidence for the efficacy of MBSR/MBCT is accumulating, the mechanisms through which these treatments act are yet to be fully elucidated. Mindfulness practice can be considered to cultivate two broad aspects of mental processing: awareness of cognitive, emotional and somatic processes (‘present-moment awareness’) and the ability to experience these processes with a non-judgemental and non-reactive attitude (‘non-judgemental acceptance’). It has been proposed that these abilities help to enhance psychological flexibility and reduce engagement in maladaptive habits and reactions (Kabat-Zinn, 2013, Segal et al., 2013). Additionally, MBCT is specifically theorized to decrease depressive recurrence by i) developing the ability to recognize, decentre and disengage from self-devaluing ruminative thought-patterns, ii) developing meta-awareness i.e. becoming able to observe thoughts and feelings as temporary and automatic events in the mind instead of as facts or true descriptions, and by iii) fostering self-compassion (Segal et al., 2013).

Recent systematic reviews of the psychological mechanisms of change point to self-reported mindfulness as the most consistent mediator of improved outcomes (Alsubaie et al., 2017, Gu et al., 2015, van der Velden et al., 2015). Other mechanisms may play a role, including: cognitive and emotional reactivity (Gu et al., 2015), compassion, meta-awareness and rumination (MBCT; van der Velden et al., 2015). Each of these reviews, however, have highlighted methodological shortcomings in existing work. In particular, reliance on self-report measures may limit the capacity to dissociate effects of training with participants’ beliefs and expectations after learning about the theoretical rationale of the treatment (discussed during the psychoeducation portion of the intervention). One method to address this issue is by incorporating objective methods in addition to self-report, such as neuroimaging, into rigorous trial designs supporting investigations of causal pathways (van der Velden et al., 2015).

There has been a number of important reviews of the neural correlates of meditation and mindfulness to date. Existing work includes narrative reviews (Creswell, 2017, Rubia, 2009, Tang et al., 2015), a meta analytic review of four common meditation techniques (focused attention, mantra recitation, open monitoring, and compassion/loving-kindness; Fox et al., 2016) and a systematic review of structural and functional changes associated with MBSR and closely related mindfulness-based interventions (MBIs; Gotink et al., 2016). The overarching finding has been that practicing meditation is associated with increased neural activation in the insula, prefrontal and anterior cingulate cortices (Fox et al., 2016, Gotink et al., 2016, Tang et al., 2015). The insula has been ascribed a number of roles in higher order cognitive functioning, including awareness of interoceptive experiences. Regions of prefrontal cortex (PFC) and dorsal anterior cingulate cortex (dACC) have been associated more broadly with a range of ‘higher-order’ cognitive processes, including attentional control and emotion regulation, both of which are implicated in the balance of awareness of in-the-moment experiences with non-judgemental acceptance.

Neuroscientific reviews to date have commented on limitations in previous work and made recommendations for improved methodologies. These include calling for well-controlled prospective studies and use of more advanced analytical tools. In addition, much of the literature on neural mechanisms of meditation has been based on i) cross-sectional studies comparing expert with novice meditators, ii) diverse and heterogenous meditation techniques, iii) limited discussion of subregions of the prefrontal cortex. We discuss these limitations in turn below.

Much of the existing knowledge on neural mechanisms of meditation comes from cross-sectional studies comparing expert with novice meditators. Those who engage in long-term meditation, like other types of long-term training such as musicianship, often differ from the general population in key ways. These include socio-economic status, personality and educational levels (Luders et al., 2013), and could plausibly extend to any number of internal mental processes. Cross-sectional studies comparing expert with novice meditators might therefore be influenced by any of these confounding variables. To enable causal inferences, there is a need for studies using a within-subjects design, affording greater experimental control and greater confidence that observed differences are associated with the intervention.

The scope of meditative approaches included in recent neuroscientific reviews has been typically broad (e.g. loving kindness and compassion meditation, open monitoring, mantra recitation, focussed attention, integrative mind-body training, zen and insight meditations). This approach is useful for assessing effects common across different types of meditation, but is less useful for assessing the specific effects associated with manualised interventions that include meditation exercises. Comparing interventions with core similarities (particularly in terms of duration, course content and home practice) reduces the impact of extraneous variables on outcome measures, with the potential to demonstrate more specific effects. As MBSR and MBCT are complex psychotherapeutic interventions that include components of meditation exercises, but also involve psycho-education, dialogue, and exercises from other treatments (e.g. CBT), there is a need to specifically investigate the neural correlates and putative mechanisms of MBSR, MBCT and closely-related MBIs, and how these relate to the prediction of clinical and well-being outcomes.

Finally, previous reviews have tended to group findings from different subregions of the PFC together. The PFC is a large and multifaceted area of the frontal lobe of the brain that has well-established links to ‘higher-order’ cognitive processing including: attention direction, stimulus appraisal, reasoning and decision-making. Advances in cognitive and affective neuroscience have clearly demonstrated evidence for localisation of specific processes to subregions or networks of regions within the PFC (Badre and Wagner, 2002, Etkin et al., 2011, Ridderinkhof et al., 2004). Further specificity on which particular subregions are implicated would be informative with regards to an understanding of neural mechanisms implicated.

In this review, we investigate changes in task-relevant neural functioning associated with manualised mindfulness-based interventions. We target our search to focus specifically on longitudinal within-subjects designs where participants undergo fMRI both before and after treatment. This is an important study design because participants serve as their own internal control, thereby addressing some of the concerns with previous work synthesising cross-sectional studies. We include studies of manualised mindfulness-based interventions to specifically assess changes in neural activity associated with a ‘standard dose’ of MBCT/MBSR or closely derived interventions. In contrast to previous work, we examine individual subregions of PFC to provide more specificity on which areas of PFC might be implicated.