The neuroscience of mindfulness meditation
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
REVIEWS
The neuroscience of mindfulness
meditation
Yi‑Yuan Tang1,2*, Britta K. Hölzel3,4* and Michael I. Posner2
Abstract | Research over the past two decades broadly supports the claim that mindfulness
meditation — practiced widely for the reduction of stress and promotion of health —
exerts beneficial effects on physical and mental health, and cognitive performance. Recent
neuroimaging studies have begun to uncover the brain areas and networks that mediate
these positive effects. However, the underlying neural mechanisms remain unclear, and it is
apparent that more methodologically rigorous studies are required if we are to gain a full
understanding of the neuronal and molecular bases of the changes in the brain that
accompany mindfulness meditation.
Longitudinal studies Meditation can be defined as a form of mental training types of meditation. However, it is important to note that
Study designs that compare that aims to improve an individual’s core psychological other styles of meditation may operate via distinct neural
data from one or more groups capacities, such as attentional and emotional self-reg- mechanisms15,16.
at several time points and that
ulation. Meditation encompasses a family of complex
ideally include a (preferably
active) control condition and
practices that include mindfulness meditation, mantra Challenges in meditation research
random assignment to meditation, yoga, tai chi and chi gong 1. Of these prac- Findings on the effects of meditation on the brain are
conditions. tices, mindfulness meditation — often described as non- often reported enthusiastically by the media and used by
judgemental attention to present-moment experiences clinicians and educators to inform their work. However,
Cross-sectional studies
(BOX 1) — has received most attention in neuroscience most of the findings have not yet been replicated. Many
Study designs that compare
data from an experimental research over the past two decades2–8. researchers are enthusiastic meditators themselves.
group with those from a control Although meditation research is in its infancy, a Although their insider perspective may be valuable for a
group at one point in time. number of studies have investigated changes in brain deep understanding of meditation, these researchers must
activation (at rest and during specific tasks) that are ensure that they take a critical view of study outcomes. In
associated with the practice of, or that follow, training fact, for meditation studies there is a relatively strong bias
in mindfulness meditation. These studies have reported towards the publication of positive or significant results,
changes in multiple aspects of mental function in begin- as was shown in a meta-analysis17.
1
Department of Psychological
ner and advanced meditators, healthy individuals and The methodological quality of many meditation
Sciences, Texas Tech
University, Lubbock, Texas patient populations9–14. research studies is still relatively low. Few are actively
79409, USA. In this Review, we consider the current state of controlled longitudinal studies, and sample sizes are small.
2
Department of Psychology, research on mindfulness meditation. We discuss the As is typical for a young research field, many experiments
University of Oregon, Eugene, methodological challenges that the field faces and point are not yet based on elaborated theories, and conclusions
Oregon 97403, USA.
3
Department of
to several shortcomings in existing studies. Taking into are often drawn from post-hoc interpretations. These
Neuroradiology, Technical account some important theoretical considerations, we conclusions therefore remain tentative, and studies must
University of Munich, 81675 then discuss behavioural and neuroscientific findings in be carefully replicated. Meditation research also faces
Munich, Germany. light of what we think are the core components of medi- several specific methodological challenges.
4
Massachusetts General
tation practice: attention control, emotion regulation
Hospital, Charlestown,
Massachusetts 02129, USA. and self-awareness (BOX 1). Within this framework, we Cross-sectional versus longitudinal studies. Early medi-
*These authors contributed describe research that has revealed changes in behaviour, tation studies were mostly cross-sectional studies: that is,
equally to this work. brain activity and brain structure following mindfulness they compared data from a group of meditators with
Correspondence to Y.‑Y.T. meditation training. We discuss what has been learned data from a control group at one point in time. These
e‑mail: yiyuan.tang@ttu.edu
doi:10.1038/nrn3916
so far from this research and suggest new research studies investigated practitioners with hundreds or
Published online strategies for the field. We focus here on mindfulness thousands of hours of meditation experience (such as
18 March 2015 meditation practices and have excluded studies on other Buddhist monks) and compared them with control
NATURE REVIEWS | NEUROSCIENCE VOLUME 16 | APRIL 2015 | 213
© 2015 Macmillan Publishers Limited. All rights reservedREVIEWS
Box 1 | Mindfulness meditation
Different styles and forms of meditation are found in almost all cultures and religions. Mindfulness meditation originally
stems from Buddhist meditation traditions3. Since the 1990s, mindfulness meditation has been applied to multiple mental
and physical health conditions, and has received much attention in psychological research2,4–7. In current clinical and
research contexts, mindfulness meditation is typically described as non-judgemental attention to experiences in the
present moment3–8. This definition encompasses the Buddhist concepts of mindfulness and equanimity158 and describes
practices that require both the regulation of attention (in order to maintain the focus on immediate experiences, such as
thoughts, emotions, body posture and sensations) and the ability to approach one’s experiences with openness and
acceptance2–10,53,158–161. Mindfulness meditation can be subdivided into methods involving focused attention and those
involving open monitoring of present-moment experience9.
The mindfulness practices that have been the subject of neuroscientific research comprise a broad range of methods
and techniques, including Buddhist meditation traditions, such as Vipassana meditation, Dzogchen and Zen, as well as
mindfulness-based approaches such as integrative body–mind training (IBMT), mindfulness-based stress reduction
(MBSR) and clinical interventions based on MBSR4–6. Both
MBSR and IBMT have adopted mindfulness practices from a Mindfulness meditation
the Buddhist traditions and aim to develop
moment-to‑moment, non-judgemental awareness through
various techniques8-11,53,73. IBMT has been categorized in
Attention Emotion Self-
the literature as open-monitoring mindfulness control regulation awareness
meditation 9,10,161
, whereas MBSR includes both focused
attention and open-monitoring practices8.
It has been suggested that mindfulness meditation
Self-regulation
includes at least three components that interact closely to
constitute a process of enhanced self-regulation:
enhanced attention control, improved emotion regulation b
and altered self-awareness (diminished self-referential Early stage Middle stage Advanced stage
processing and enhanced body awareness)10 (see the
figure, part a). Mindfulness meditation can be roughly
Effort to
divided into three different stages of practice — early, Effortful Effortless
reduce mind
middle (intermediate) and advanced — that involve doing being
wandering
different amounts of effort11 (see the figure, part b).
Nature Reviews | Neuroscience
groups of non-meditators matched on various dimen- group suggests that meditation has caused the observed
sions9,18. The rationale was that any effects of meditation changes, especially when other potentially confounding
would be most easily detectable in highly experienced variables are controlled for properly20–22.
practitioners.
A number of cross-sectional studies revealed differ- Novice meditators versus expert meditators. Although
ences in brain structure and function associated with most cross-sectional studies included long-term medi-
meditation (see below). Although these differences may tators9,17, longitudinal studies are often conducted in
constitute training-induced effects, a cross-sectional beginners or naive subjects. Thus, differences in the
study design precludes causal attribution: it is possible results of cross-sectional and longitudinal analyses
that there are pre-existing differences in the brains of might be attributed to the different brain regions used
meditators, which might be linked to their interest in during learning of meditation versus those used during
meditation, personality or temperament 2,19. Although the continued practice of an acquired skill. It would be
correlational studies have attempted to discover whether interesting to follow subjects over a long-term period
more meditation experience is related to larger changes of practice to determine whether changes induced by
in brain structure or function, such correlations still meditation training persist in the absence of continued
cannot prove that meditation practice has caused the practice. However, such long-term longitudinal studies
changes because it is possible that individuals with these would be compromised by feasibility constraints, and
particular brain characteristics may be drawn to longer it is likely that future longitudinal studies will remain
meditation practice. restricted to relatively short training periods2.
More recent research has used longitudinal designs,
which compare data from one or more groups at sev- Control groups and interventions. It is important to con-
eral time points and ideally include a (preferably active) trol for variables that may be confounded with meditation
control condition and random assignment to condi- training, such as changes in lifestyle and diet that might
Correlational studies
Studies that assess the tions11–14,20–25. In meditation research, longitudinal stud- accompany the meditation practice or the expectancy
co‑variation between two ies are still relatively rare. Among those studies, some and intention that meditation beginners bring to their
variables: for example, have investigated the effects of mindfulness training practice. Researchers must carefully determine which
co‑variation of functional or over just a few days, whereas others have investigated variables are integral aspects of the meditation training
structural properties of the
brain and a behavioural
programmes of 1 to 3 months. Some of these studies and which can be controlled for. Some earlier studies only
variable, such as reported have revealed changes in behaviour, brain structure and controlled for the length of time that the individual has
stress. function11–14,20–25. A lack of similar changes in the control practised meditation and the effects of repeated testing,
214 | APRIL 2015 | VOLUME 16 www.nature.com/reviews/neuro
© 2015 Macmillan Publishers Limited. All rights reservedREVIEWS
Blood-oxygen-level- but more recent studies have developed and included used to investigate effects on both grey and white matter.
dependent contrasts active interventions in control groups — such as stress Studies have captured cortical thickness32,51, grey-matter
(BOLD contrasts). Signals that management education26, relaxation training 14,23,27 or volume and/or density 33,40, fractional anisotropy and axial
can be extracted with health enhancement programmes20–22 — that can control and radial diffusivity38,39. These studies have also used dif-
functional MRI and that reflect
the change in the amount of
for variables such as social interaction with the group and ferent research designs. Most have made cross-sectional
deoxyhaemoglobin that is teachers, amount of home exercise, physical exercise and comparisons between experienced meditators and con-
induced by changes in the psychoeducation. These studies are therefore better able trols32–34; however, a few recent studies have investigated
activity of neurons and their to extract and delineate the meditation-specific effects. longitudinal changes in novice practitioners38–40. Some
synapses in a region of the
For example, one study investigating short-term medita- further studies have investigated correlations between
brain. The signals thus reflect
the activity in a local brain
tion training used a ‘sham meditation’ condition in which brain changes and other variables related to mindfulness
region. participants thought they were meditating, but did not practice, such as stress reduction40, emotion regulation39
receive proper meditation instructions, which allowed or increased well-being 47. Most studies include small
Arterial spin labelling the researchers to control for factors such as expec- sample sizes of between 10 and 34 subjects per group31-42.
(ASL). An MRI technique that is
capable of measuring cerebral
tancy, body posture and attention from the teacher 28. Because the studies vary in regard to study design,
blood flow in vivo. It provides Mechanistic studies ideally need to use interventions that measurement and type of mindfulness meditation, it is
cerebral perfusion maps are as effective as mindfulness meditation in producing not surprising that the locations of reported effects are
without requiring the the beneficial effects on target variables but that allow diverse and cover multiple regions in the brain31–34,36–52.
administration of a contrast
for assessment of the unique mechanism underlying the Effects reported by individual studies have been found
agent or the use of ionizing
radiation because it uses
mindfulness practice23,29. in multiple brain regions, including the cerebral cortex,
magnetically labelled subcortical grey and white matter, brain stem and cer-
endogenous blood water as a Control conditions in functional imaging. Although all ebellum, suggesting that the effects of meditation might
freely diffusible tracer. functional neuroimaging studies must use appropriate involve large-scale brain networks. This is not surprising
Brain state
comparison conditions, this challenge is particularly because mindfulness practice involves multiple aspects
The reliable patterns of brain important when imaging meditative states (BOX 2). The of mental function that use multiple complex interactive
activity that involve the comparison condition should be one in which a state networks in the brain. TABLE 1 summarizes the main find-
activation and/or connectivity of mindfulness meditation is not present. Many studies ings of structural neuroimaging studies on mindfulness
of multiple large-scale brain
use resting comparison conditions, but a problem with meditation (grey and white matter).
networks.
this is that experienced practitioners are likely to enter An activation likelihood estimation meta-analysis, which
Fractional anisotropy into a state of meditation when at rest. However, other also included studies from traditions other than mind-
A parameter in diffusion tensor active tasks introduce additional brain activity that ren- fulness meditation, was conducted to investigate which
imaging, which images brain ders the comparison difficult to interpret. Using imaging regions were consistently altered in meditators across
structures by measuring the
diffusion properties of water
protocols that do not rely on blood-oxygen-level-dependent studies17. The findings demonstrated a global medium
molecules. It provides contrasts (BOLD contrasts), such as arterial spin labelling, effect size, and eight brain regions were found to be
information about the might be a possible solution for this problem30. consistently altered in meditators: the frontopolar
microstructural integrity of cortex, which the authors suggest might be related to
white matter.
Changes in brain structure enhanced meta-awareness following meditation prac-
In the past decade, 21 studies have investigated altera- tice; the sensory cortices and insula, areas that have
tions in brain morphometry related to mindfulness been related to body awareness; the hippocampus, a
meditation17. These studies varied in regard to the region that has been related to memory processes; the
exact mindfulness meditation tradition under investi- anterior cingulate cortex (ACC), mid-cingulate cortex
gation31–34,36–52, and multiple measurements have been and orbitofrontal cortex, areas known to be related to
self and emotion regulation; and the superior longitu-
dinal fasciculus and corpus callosum, areas involved in
Box 2 | Imaging the meditative state intra- and inter-hemispherical communication17.
A brain state can be defined as a reliable pattern of activity and/or connectivity in Thus, some initial attempts have been undertaken to
multiple large-scale brain networks11,73. Meditation training involves obtaining a investigate the brain regions that are structurally altered
meditative state, and measurements of behaviour and/or brain activity can be made by the practice of meditation. However, our knowledge
while participants are thought to be in such a state15,76,162,163. These studies can elucidate of what these changes actually mean will remain trivial
how the state influences the brain and behaviour2,10,14,73. To identify brain regions until we gain a better understanding of how such struc-
activated during the state of meditation (compared to a baseline state) across multiple
tural changes are related to the reported improvements
studies in experienced healthy meditators, an activation likelihood estimate
in affective, cognitive and social function. Very few
meta-analysis of 10 studies with 91 subjects published before January 2011 was
performed108. This revealed three areas in which there were clusters of activity: the studies have begun to relate findings in the brain to self-
caudate, which is thought (together with the putamen) to have a role in attentional reported variables and behavioural measures34,39,47,48,51.
disengagement from irrelevant information, allowing a meditative state to be achieved Future studies therefore need to replicate the reported
and maintained; the entorhinal cortex (parahippocampus), which is thought to control findings and begin to unravel how changes in the neural
the mental stream of thoughts and possibly stop mind wandering; and the medial structure relate to changes in well-being and behaviour.
prefrontal cortex, which is thought to support the enhanced self-awareness during Growing evidence also demonstrates changes in the
meditation108 (also see REF. 162). It was suggested that these regions of activity might functional properties of the brain following meditation.
represent a core cortical network for the meditative state, independent of the Below, we summarize such findings in the context of
meditation technique. It is important to note, however, that this meta-analysis included
the framework of core mechanisms of mindfulness
mostly papers from traditions other than mindfulness.
meditation (BOX 1; FIG. 1).
NATURE REVIEWS | NEUROSCIENCE VOLUME 16 | APRIL 2015 | 215
© 2015 Macmillan Publishers Limited. All rights reservedREVIEWS
Table 1 | Structural changes in the brain associated with mindfulness meditation
Meditation Control Sample size of Type of measurement Key areas affected‡ Refs
tradition* meditation (M) and
control (C) groups
Cross-sectional studies (non-clinical samples)
Insight Non-meditators M: 20, C: 15 Cortical thickness Right anterior insula and right middle and 32
superior frontal sulci
Zen Non-meditators M: 13, C: 13 Grey-matter volume Less age-related decline at left putamen 34
Insight Non-meditators M: 20, C: 20 Grey-matter density Right anterior insula, left inferior temporal gyrus 31
and right hippocampus
Tibetan Non-meditators M: 10, C: 10 Grey-matter density Medulla oblongata, left superior and inferior 33
Dzogchen frontal gyrus, anterior lobe of the cerebellum
(bilateral) and left fusiform gyrus
Zen Non-meditators M: 17, C: 18 Cortical thickness Right dorsal anterior cingulate cortex and 51
secondary somatosensory cortices (bilateral)
MBSR Non-meditators M: 20, C: 16 Grey-matter volume Left caudate nucleus 52
Zen Non-meditators M: 10, C: 10 DTI: mean diffusivity and Lower mean diffusivity in left posterior parietal 37
fractional anisotropy white matter and lower fractional anisotropy in
left primary sensorimotor cortex grey matter
Longitudinal studies (non-clinical samples)
IBMT Active control: M: 22, C: 23 DTI: FA and grey-matter FA increased for left anterior corona radiata, 38
(4 weeks) relaxation training volume superior corona radiata (bilateral), left superior
longitudinal fasciculus, genu and body of corpus
callosum. No effect on grey-matter volume
MBSR Individuals on a M: 16, C: 17 Grey-matter density Left hippocampus, left posterior cingulate gyrus, 40
waiting list cerebellum and left middle temporal gyrus
IBMT Active control: M: 34, C: 34 DTI: FA, radial diffusivity Decrease of axial diffusivity in corpus callosum, 39
(2 weeks) relaxation training and axial diffusivity corona radiata, superior longitudinal fasciculus,
posterior thalamic radiation and sagittal striatum
Longitudinal studies (clinical samples)
MBI Usual care M: 14, C: 13 Grey-matter density Caudate (bilateral), left inferior temporal lobe, 42
(patients with hippocampus (bilateral), left occipital cuneus
Parkinson disease) and other small clusters; anterior cerebellum
increased in usual care group
MBSR Waiting list M: 8, C: 5 Hippocampal volume Trend towards less hippocampal atrophy 41
(patients with (region of interest analysis)
mild cognitive
impairment)
DTI, diffusion tensor imaging; FA, fractional anisotropy; IBMT, integrative body–mind training; MBI, mindfulness-based intervention; MBSR, mindfulness-based
stress reduction. *Studies that include meditators from traditions other than mindfulness or studies only investigating correlations with other variables are not
listed. ‡Meditators show increased values, unless otherwise noted.
Mindfulness and attention neural areas, also referred to as executive attention)55,56.
Many meditation traditions emphasize the necessity to Other distinctions between types of attention refer to
cultivate attention regulation early in the practice9,53. combinations of these three components. For example,
Axial and radial diffusivity A sufficient degree of attentional control is required to sustained attention refers to the sense of vigilance during
Derived from the eigenvalues stay engaged in meditation, and meditators often report long continued tasks and may involve both tonic alerting
of the diffusion tensor, their improved attention control as an effect of repeated prac- and orienting, whereas selective attention may involve
underlying biophysical
properties are associated with
tice10,14. Multiple studies have experimentally investigated either orienting (when a stimulus is present) or executive
axonal density and such effects54. function (when stored information is involved).
myelination, respectively. Performance in these three basic domains can be
Components of attention. Attention is often subdivided measured with the attention network test (ANT)57. This
Activation likelihood
into three different components: alerting (readiness in test uses as a target an arrow pointing left or right. The
estimation meta-analysis
A technique for preparation for an impending stimulus, which includes target is surrounded by flankers, and subtracting reac-
coordinate-based tonic effects that result from spending time on a task tion times to congruent stimuli (that is, those on the side
meta-analysis of neuroimaging (vigilance) and phasic effects that are due to brain of the screen indicated by the arrow) from reaction times
data. It determines the changes induced by warning signals or targets); orienting to incongruent stimuli produces a measure of the time to
convergence of foci reported
from different experiments,
(the selection of specific information from multiple sen- resolve conflict. The inclusion of cues that indicate when
weighted by the number of sory stimuli); and conflict monitoring (monitoring and or where the target will occur allows the measurement
participants in each study. resolution of conflict between computations in different of alerting and orienting. These measures are used to
216 | APRIL 2015 | VOLUME 16 www.nature.com/reviews/neuro
© 2015 Macmillan Publishers Limited. All rights reservedREVIEWS
Anterior cingulate Posterior cingulate performance task that measured aspects of tonic alertness,
cortex cortex/precuneus but did show some improvement in orienting 22. We do
Multiple not know whether the differences in the findings of these
prefrontal studies are due to the type of training, type of control or
regions other subtle factors.
A systematic review that compiled the findings of
Medial these studies (as well as the effects on other measures of
prefrontal
cortex
cognition) concluded that early phases of mindfulness
meditation might be associated with improvements in
Striatum
conflict monitoring and orienting, whereas later phases
Amygdala Insula might be mainly associated with improved alerting 60. It is
currently still unclear how different meditation practices
differentially affect the specific attentional components2,9,53.
Medial view Lateral view In addition, the length of practice needs to be defined more
consistently in future research.
Figure 1 | Brain regions involved in the components of mindfulness meditation.
Nature Reviews | Neuroscience
Schematic view of some of the brain regions involved in attention control (the anterior
Neural mechanisms of enhanced attention control.
cingulate cortex and the striatum), emotion regulation (multiple prefrontal regions, limbic
regions and the striatum) and self-awareness (the insula, medial prefrontal cortex and
Several functional and structural MRI studies on mind-
posterior cingulate cortex and precuneus). fulness training have investigated neuroplasticity in brain
regions supporting attention regulation. The brain region
to which the effects of mindfulness training on atten-
quantify efficiency in each of the three networks that tion is most consistently linked is the ACC11,23,38,39,73–76.
support the individual components of attention. Alerting The ACC enables executive attention and control77–79 by
involves the brain’s noradrenaline system, which origi- detecting the presence of conflicts emerging from incom-
nates in the locus coeruleus. Orienting involves frontal patible streams of information processing. The ACC and
and parietal areas, including the frontal eye fields and the fronto-insular cortex form part of a network that
inferior and superior parietal lobe. The executive net- facilitates cognitive processing through long-range con-
work involved in conflict resolution involves the ACC, nections to other brain areas11,80. Cross-sectional studies
anterior insula and basal ganglia58,59. have reported enhanced activation of regions of the ACC
in experienced meditators compared to controls during
Effects of mindfulness meditation on attention. The focused attention meditation76 or when mindfully antici-
ANT and other experimental paradigms have been pating delivery of a painful stimulus81. Greater activation
used to investigate the effects of meditation on atten- of the ventral and/or rostral ACC during the resting state
tional performance60. Improved conflict monitoring was following 5 days of IBMT was also found in an actively
reported in several studies14,61–64. For example, a longitu- controlled, randomized, longitudinal study 23. Although
dinal study showed that only 5 days (20 min per day) of ACC activation may be enhanced in earlier stages of
integrative body–mind training (IBMT) led to improved mindfulness meditation, it might decrease with higher
conflict monitoring 14. In addition, cross-sectional stud- levels of expertise, as demonstrated in a cross-sectional
ies of 3 months of mindfulness meditation showed a study 18. Structural MRI data suggest that mindfulness
reduced attentional blink (a lapse in attention following meditation might be associated with greater cortical
a stimulus within a rapid stream of presented stimuli thickness51 and might lead to enhanced white-matter
that has been related to executive function65,66) follow- integrity in the ACC38,39.
ing training 64 (also see REF. 67). Better performance in Other attention-related brain regions in which func-
conflict monitoring has also been demonstrated in expe- tional changes have been observed following mindful-
rienced meditators in cross-sectional studies68. However, ness meditation include the dorsolateral prefrontal cortex
although altered attention is a common finding in these (PFC), where responses were enhanced during executive
well-designed meditation studies, some studies investi- processing 82, as revealed by a randomized longitudi-
gating mindfulness-based stress reduction (MBSR) have nal study, and parietal attention regions, which showed
not observed effects on conflict monitoring 69,70. greater activation following an MBSR course in people
Most of the studies on the effects of short-term with social anxiety, as demonstrated by an uncontrolled
(1 week) mindfulness meditation on alerting have longitudinal study 83. Furthermore, a diminished age-
not found significant effects, but studies investigating related decline of grey-matter volume in the putamen as
long-term meditators (ranging from months to years) well as diminished age-related decline in sustained atten-
did detect changes in alerting 27,70–72. Enhanced orienting tion performance were found in a cross-sectional study of
has been reported in some cross-sectional studies using Zen meditation practitioners34.
longer periods of training. For example, 3 months of Although there is evidence that brain regions rele-
Shamatha mindfulness training improved tonic alert- vant for the regulation of attention show functional and
ness (the ability to remain alert over time) and allowed structural changes following mindfulness meditation
for improved orienting towards a visual target in com- practice, it has not yet been determined whether these
parison to controls71. However, 8 weeks of MBSR did not changes are actually related to the improved attentional
improve measures of sustained attention in a continuous performance. Longitudinal studies that use measures
NATURE REVIEWS | NEUROSCIENCE VOLUME 16 | APRIL 2015 | 217
© 2015 Macmillan Publishers Limited. All rights reservedREVIEWS
of attentional performance along with functional MRI is important, with beginners showing a different pat-
(fMRI) are needed. If supported by more rigorous tern from expert meditators. However, although several
future research, the evidence of improved attention reg- studies have pointed to the involvement of fronto-limbic
ulation and strengthened brain activity in the regions regions, very few studies have begun to relate changes
underlying attentional control following mindfulness in these regions to changes in measures of behaviour or
meditation might be promising for the treatment of well-being 10.
psychiatric disorders in which there are deficiencies in A frequently reported finding is that mindfulness
these functions74,84,85. practice leads to (or is associated with) a diminished acti-
vation of the amygdala in response to emotional stimuli
Mindfulness and emotion regulation during mindful states83,94,95 as well as in a resting state93,
Enhanced emotion regulation has been suggested to suggesting a decrease in emotional arousal. However,
underlie many of the beneficial effects of mindfulness although such results have been reported for meditation
meditation. Emotion regulation refers to strategies that beginners, they have less consistently been detected in
can influence which emotions arise and when, how long experienced meditators95 (but see REF. 18).
they occur, and how these emotions are experienced and Prefrontal activations are often enhanced as an effect
expressed. A range of implicit and explicit emotion regu- of mindfulness meditation in novice meditators (but see
lation processes has been proposed86, and mindfulness- REF. 29): for example, greater dorsolateral PFC responses
based emotion regulation may involve a mix of these were found during executive processing within an emo-
processes, including attentional deployment (attending tional Stroop task in healthy individuals after 6 weeks
to mental processes, including emotions), cognitive of mindfulness training 82. Enhanced dorsomedial and
change (altering typical patterns of appraisal regarding dorsolateral PFC activation was also detected when par-
one’s emotions) and response modulation (decreasing ticipants expected to see negative images while engaging
tonic levels of suppression). in a mindful state94. Moreover, after an MBSR course, an
enhanced activation in the ventrolateral PFC in people
Effects of mindfulness meditation on emotion regulation. suffering from anxiety was found when they labelled the
Improvements in emotion regulation associated with affect of emotional images97. By contrast, experienced
mindfulness meditation have been investigated through meditators have been found to show diminished acti-
various approaches, including experimental studies, self- vation in medial PFC regions95. This finding could be
reporting studies, measurement of peripheral physiology interpreted as indicating reduced control (disengagement
and neuroimaging 10. These studies have reported various of elaboration and appraisal) and greater acceptance of
positive effects of mindfulness meditation on emotional affective states.
processing, such as a reduction in emotional interference Neuroimaging studies of ameliorated pain process-
by unpleasant stimuli87, decreased physiological reac- ing through mindfulness meditation have also pointed
tivity and facilitated return to emotional baseline after to expertise-related differences in the extent of cogni-
response to a stressor film88, and decreased self-reported tive control over sensory experience. Meditation begin-
difficulties in emotion regulation89. Consequently, low- ners showed increased activity in areas involved in the
ered intensity and frequency of negative affect 90,91 and cognitive regulation of nociceptive processing (the ACC
improved positive mood states14,91,92 are reported to be and anterior insula) and areas involved in reframing the
associated with mindfulness meditation. evaluation of stimuli (the orbitofrontal cortex), along with
reduced activation in the primary somatosensory cortex
Neural mechanisms of improved emotion regulation. in a 4‑day longitudinal study with no control group30,
Neuroimaging studies that have probed the enhanced whereas meditation experts were characterized by
emotion regulation associated with mindfulness medita- decreased activation in dorsolateral and ventrolateral PFC
tion in an attempt to identify the underlying brain acti- regions and enhancements in primary pain processing
vation patterns typically present study participants with regions (the insula, somatosensory cortex and thalamus)
emotional pictures82,93–97, words and/or statements29,98 compared with controls in two cross-sectional studies35,81.
and instruct them to encounter these with a state of These findings are in line with the assumption that
mindfulness or a simple baseline state. the process of mindfulness meditation is characterized
The hypothesis that drives many of these studies is as an active cognitive regulation in meditation begin-
that mindful emotion regulation works by strengthen- ners, who need to overcome habitual ways of internally
ing prefrontal cognitive control mechanisms and thus reacting to one’s emotions and might therefore show
downregulates activity in regions relevant to affect pro- greater prefrontal activation. Expert meditators might
cessing, such as the amygdala. Present-moment awareness not use this prefrontal control. Rather, they might have
and non-judgemental acceptance through mindfulness automated an accepting stance towards their experience
meditation8,10 are thought to be crucial in promoting and thus no longer engage in top-down control efforts
cognitive control because they increase sensitivity to but instead show enhanced bottom‑up processing 100.
affective cues that help to signal the need for control99. In the early stages of meditation training, achieving
Studies have therefore investigated whether mindfulness the meditation state seems to involve the use of atten-
training exerts its effects through enhanced top-down tional control and mental effort; thus, areas of the lat-
control or facilitated bottom‑up processing 100. The find- eral prefrontal and parietal cortex are more active than
ings (outlined below) suggest that the level of expertise before training 11,16,100,101. This may reflect the higher level
218 | APRIL 2015 | VOLUME 16 www.nature.com/reviews/neuro
© 2015 Macmillan Publishers Limited. All rights reservedREVIEWS
of effort often found when participants struggle to obtain Importantly, this study also investigated the correlation
the meditation state in the early stages11,73,98,102. However, between neural and self-reported findings and demon-
in the advanced stages, prefrontal–parietal activity is strated that the changes in PFC–amygdala connectivity
often reduced or eliminated, but ACC, striatum and were correlated with anxiety symptom improvement.
insula activity remains9,10,53,73,76,101–103. Whether effort has Further studies are needed to elucidate the complex
a key role in PFC and ACC activation during or following interplay between regions of the fronto-limbic network
meditation needs further investigation. in mindfulness meditation.
Analysis of functional connectivity between regions Although the proposed similarities between mind-
of the fronto-limbic network could help to further eluci- fulness and the reappraisal strategy of emotion regula-
date the regulatory function of executive control regions. tion have been much debated, there is some evidence
Only a few studies have included such analyses. One that mindfulness also shares similarities with extinction
cross-sectional study on pain processing in meditators processes (BOX 3).
demonstrated decreased connectivity of executive and Brain regions involved in motivation and reward pro-
pain-related brain regions35, and one study of mindful- cessing also show functional alterations that are related to
ness-naive smokers demonstrated reduced connectivity mindfulness training, such as stronger activity of the puta-
between craving-related brain regions during a mindful- men and caudate during a resting state following mindful-
ness condition compared to passive viewing of smoking- ness training23 and lower activation in the caudate nucleus
related images during cigarette craving 96, suggesting during reward anticipation in experienced meditators106.
a functional decoupling of involved regions. Another These studies might indicate altered self-regulation in
longitudinal, randomized study reported that people the motivational realm, with possibly reduced suscepti-
suffering from anxiety showed a change from a negative bility to incentives and enhanced reward-related activity
correlation between the activity of frontal regions and during rest.
that of the amygdala before intervention (that is, negative Brain regions involved in the regulation of emotions
connectivity) to a positive correlation between the activity have also shown structural changes following mindful-
of these regions (positive connectivity) after a mindful- ness meditation31,32,38–41,48,51. Although these findings pro-
ness intervention97. Because such a negative correlation vide some initial evidence that these brain regions are
will occur when prefrontal regions downregulate limbic related to mindfulness practice, the question of whether
activation104,105, it was speculated that the positive coupling they are involved in mediating the beneficial effects of
between the activity of the two regions after mindfulness mindfulness meditation remains largely unanswered.
intervention might indicate that meditation involves
monitoring of arousal rather than a downregulation or Mindfulness and self-awareness
suppression of emotional responses, and that it might According to Buddhist philosophy, the identification with
be a unique signature of mindful emotion regulation. a static concept of ‘self ’ causes psychological distress. Dis-
identification from such a static self-concept results in
the freedom to experience a more genuine way of being.
Box 3 | Mindfulness meditation as exposure therapy
Through enhanced meta-awareness (making awareness
itself an object of attention), mindfulness meditation is
Exposure therapy aims for patients to extinguish a fear response and instead to thought to facilitate a detachment from identification
acquire a sense of safety in the presence of a formerly feared stimulus by exposing with the self as a static entity 3,10,107 and a tendency to iden-
them to that stimulus and preventing the usual response164. Mindfulness meditation tify with the phenomenon of ‘experiencing’ itself is said
resembles an exposure situation because practitioners ‘turn towards their emotional
to emerge15,108–112. Currently, empirical research into this
experience’, bring acceptance to bodily and affective responses, and refrain from
engaging in internal reactivity towards it. Research on fear conditioning has helped to
area is only just emerging111,113, and the few interpretations
identify a network of brain regions that are crucial for the extinction of conditioned of connections between neuroimaging findings and self-
fear responses and the retention of extinction165. This network includes the reported data — which we will summarize briefly below
ventromedial prefrontal cortex (vmPFC), which is important for a successful recall of — are suggestive at best.
the extinction; the hippocampus166, which is related to signalling the extinguished
context (contextual safety); and the amygdala, which has a crucial role during the Self-referential processing. Altered self-representation has
acquisition and expression of conditioned fear167 and is thought to be downregulated been investigated with questionnaire studies. Early stud-
by the vmPFC and the hippocampus105,168. Activation in the vmPFC (subgenual anterior ies reported mindfulness training to be associated with
cingulate cortex) is primarily linked to the expression of fear learning during a delayed a more positive self-representation, higher self-esteem,
test of extinction and is critical for the retention of extinction169.
higher acceptance of oneself 114 and styles of self-concept
There is emerging evidence from MRI studies that the aforementioned brain regions
show structural and functional changes following mindfulness meditation training
that are typically associated with less-severe pathological
(see main text). This overlap of involved brain regions, as well as the conceptual symptoms115. Meditators have also been shown to score
similarity between mindfulness and an exposure situation, suggest that mindfulness higher than non-meditators on a scale that measures
training might enhance the ability to extinguish conditioned fear by structurally and non-attachment 116: a construct that is based on insight
functionally affecting the brain network that supports safety signalling. The capacity into the constructed and impermanent nature of mental
for successful extinction memories reliably differentiates healthy from pathological representations. Although such concepts are not easy to
conditions170,171, and is crucial in order to overcome maladaptive states. It helps capture in experimental and neuroscientific studies, find-
individuals to learn to have no fear response to neutral stimuli when there is no ings from a few recent studies seem to suggest that brain
adaptive function for a fear response. Instead, individuals can experience a sense of structures supporting self-referential processing might be
safety and can flexibly elicit other emotions and behaviours.
affected by mindfulness meditation98,117,118.
NATURE REVIEWS | NEUROSCIENCE VOLUME 16 | APRIL 2015 | 219
© 2015 Macmillan Publishers Limited. All rights reservedREVIEWS
Although there is much debate about its exact function, affective or subjective self-referential value. Furthermore,
a widespread view holds that the default mode network a preliminary analysis from a study of a state of ‘non-dual
(DMN)119,120 is involved in self-referential processing. This awareness’ (a state of awareness in which perceived duali-
network includes midline structures of the brain, such as ties, such as the distinction between subject and object, are
areas of the medial PFC, posterior cingulate cortex (PCC), absent) showed a decreased functional connectivity of the
anterior precuneus and inferior parietal lobule121,122. These central precuneus with the dorsolateral PFC. The author
structures show high activity during rest, mind wandering speculates that this finding might be indicative of a state in
and conditions of stimulus-independent thought 121 and which awareness is itself the subject of awareness111.
have been suggested to support diverse mechanisms by Together, the findings from these studies have been
which an individual can ‘project’ themselves into another taken to suggest that mindfulness meditation might alter
perspective123. fMRI studies have investigated activity the self-referential mode so that a previous narrative,
in the DMN in association with mindfulness practice. evaluative form of self-referential processing is replaced
Regions of the DMN (the medial PFC and PCC) showed by greater awareness98,111. We suggest that this shift in
relatively little activity in meditators compared to con- self-awareness is one of the major active mechanisms
trols across different types of meditation, which has of the beneficial effects of mindfulness meditation.
been interpreted as indicating diminished self-referential However, because these interpretations are built on a
processing 117. Functional connectivity analysis revealed still-fragmentary understanding of the function of the
stronger coupling in experienced meditators between the involved brain regions, future research will need to test
PCC, dorsal ACC and dorsolateral PFC, both at baseline and elaborate these assumptions.
and during meditation, which was interpreted as indi- Across the functional and structural MRI studies that
cating increased cognitive control over the function of have been published to date, especially those based on
the DMN117. Increased functional connectivity was also the longitudinal, randomized, controlled studies with
found between DMN regions and the ventromedial PFC active control groups and meta-analyses, the ACC, PFC,
in participants with more compared to less meditation PCC, insula, striatum (caudate and putamen) and amyg-
experience118. It has been speculated that this increased dala seem to show consistent changes associated with
connectivity with ventromedial PFC regions supports mindfulness meditation9–11,13,17,23,34,73,108,130 (FIG. 1; TABLE 2).
greater access of the default circuitry to information about We consider these areas to be the core regions involved
internal states because this region is highly interconnected in self-regulation of attention, emotion and awareness
with limbic regions118. following mindfulness training. However, we acknowl-
edge that many other brain areas are also involved in
Awareness of present-moment experiences. Evaluative mindfulness practice and warrant further investigation
self-referential processing is assumed to decrease as an using rigorous randomized and controlled designs.
effect of mindfulness meditation, whereas awareness of
present-moment experiences is thought to be enhanced. Future questions
Mindfulness practitioners often report that the practice of Mechanisms of mindfulness-induced changes. A num-
attending to present-moment body sensations results in ber of studies seem to suggest that mindfulness medita-
an enhanced awareness of bodily states and greater per- tion induces changes in brain structure and function,
ceptual clarity of subtle interoception. Empirical findings raising the question of which underlying mechanisms
to support this claim are mixed. Although studies that support these processes. It is possible that engaging the
assessed performance on a heartbeat detection task — a brain in mindfulness affects brain structure by inducing
standard measure of interoceptive awareness — found dendritic branching, synaptogenesis, myelinogenesis or
no evidence that meditators had superior performance to even adult neurogenesis. Alternatively, it is possible that
non-meditators124,125, other studies found that meditators mindfulness positively affects autonomic regulation and
showed greater coherence between objective physiological immune activity, which may result in neuronal preser-
data and their subjective experience in regard to an emo- vation, restoration and/or inhibition of apoptosis14,23,131.
tional experience126 and the sensitivity of body regions127. It is well known that mindfulness-based techniques are
Multiple studies have shown the insula to be impli- highly effective in stress reduction, and it is possible
cated in mindfulness meditation: it shows stronger acti- that such stress reduction may mediate changes in brain
vation during compassion meditation128 and following function14,48,132–137 (BOX 4). A combination of all of these
mindfulness training 23,52,98, and has greater cortical thick- mechanisms may even occur.
ness in experienced meditators32. Given its known role in It is also important to realize that the direction of
awareness129, it is conceivable that enhanced insula activ- the observed effects of mindfulness meditation has
ity in meditators might represent the amplified awareness not been consistent across all studies. Although larger
of present-moment experience. values in meditators compared to controls are predomi-
Similarly, a study reported an uncoupling of the right nantly reported, a cross-sectional study also revealed
insula and medial PFC and increased connectivity of the smaller fractional anisotropy and cortical thickness
right insula with dorsolateral PFC regions in individuals values in meditators in some brain regions, including
after mindfulness training 98. The authors interpret their the medial PFC, postcentral and inferior parietal cor-
findings as a shift in self-referential processing towards tices, PCC and medial occipital cortex 138. Along these
a more self-detached and objective analysis of interocep- lines, mindfulness-induced increases are predominantly
tive and exteroceptive sensory events, rather than their observed in longitudinal studies. However, it was also
220 | APRIL 2015 | VOLUME 16 www.nature.com/reviews/neuro
© 2015 Macmillan Publishers Limited. All rights reservedREVIEWS
Table 2 | Evidence for changes in the core brain regions after mindfulness meditation
Brain region Study design Findings* Refs
ACC Cross-sectional, Vipassana meditators (N = 15) Enhanced ACC activation during breath awareness (focused 76
(self-regulation of versus controls (N = 15) attention) meditation
attention and emotion)
Longitudinal, IBMT versus active control Enhanced ACC activity in resting state 23
(relaxation training) (N = 23 each group)
PFC Longitudinal, mindfulness training (N = 30) Greater dorsolateral PFC activation during emotional Stroop 82
(attention and emotion) versus active control (N = 31) executive processing
Longitudinal, patients with generalized anxiety Enhanced activation of ventrolateral PFC, enhanced 97
disorder, MBSR (N = 15) versus active control connectivity of several PFC regions with amygdala
(N = 11)
Longitudinal, uncontrolled, before and after Anxiety relief following mindfulness training was related to 157
mindfulness training (N = 15) ventromedial PFC and ACC activation (along with insula)
PCC Cross-sectional, expert meditators (N = 12) PCC deactivation during different types of meditation, increased 117
(self-awareness) versus controls (N = 13) coupling with ACC and dorsolateral PFC
Cross-sectional, expert meditators (N = 14) Reduced connectivity between left PCC and medial PFC and 118
divided into high and low practice groups ACC at rest in high practice group
Longitudinal, IBMT, active control (relaxation Enhanced right PCC activity at resting state 23
training) (N = 23 each group)
Insula Cross-sectional, MBSR (N = 20) and waiting list Greater anterior insula activation and altered coupling between 52
(awareness and control (N = 16) dorsomedial PFC and posterior insula during interoceptive
emotional processing) attention to respiratory sensations
Cross-sectional, expert Tibetan Buddhist Enhanced insula activation when presented with emotional 128
meditators (N = 15) and novices (N = 15) sounds during compassion meditation
Longitudinal, IBMT, active control (relaxation Enhanced left insula activity at resting state 23
training) (N = 23 each group)
Striatum Longitudinal, IBMT, active control (relaxation Enhanced caudate and putamen activity at resting state 23
(regulation of attention training) (N = 23 each group)
and emotion)
Cross-sectional, expert meditators (N = 34) and Lower activation in the caudate nucleus during reward 106
controls (N = 44) anticipation
Amygdala Longitudinal, mindful attention training (N = 12), Decreased activation in right amygdala in response to emotional 93
(emotional processing) compassion training (N = 12) and active control pictures in a non-meditative state
(N = 12)
Longitudinal, uncontrolled, patients with social Diminished right dorsal amygdala activity during reacting to 83
anxiety disorder before and after MBSR (N = 14) negative self-belief statements
Cross-sectional, beginner (N = 10) and expert Downregulation of the left amygdala when viewing emotional 95
Zen meditators (N = 12) pictures in a mindful state in beginner but not expert meditators
Exemplary studies for each region support its involvement in mindfulness (the list is not comprehensive). Future research will need to test the hypothesized
functions by relating behavioural and neuroimaging findings. ACC, anterior cingulate cortex; IBMT, integrative body–mind training; MBSR, mindfulness-based
stress reduction; PCC, posterior cingulate cortex; PFC, prefrontal cortex. *Meditators show increased values, unless otherwise noted.
reported, for example, that as a consequence of medi- Decoding mental states. Mindfulness meditation
tation, larger decreases in perceived stress were associ- approaches can be divided into those involving focused
ated with larger decreases in grey-matter density in the attention and those involving open monitoring. Even
amygdala48. Thus, the underlying mechanisms seem to within the same meditation style, practitioners can be
be more complex than currently assumed, and further at different stages of mindfulness practice2. Investigating
research is necessary. the distinction between these different stages in terms of
Although neuroimaging has advanced our under- brain function will require new advanced tools and meth-
standing of the individual brain regions involved in ods. For instance, simultaneous multi-level recording
mindfulness meditation, most evidence supports the — using fMRI and electrophysiology — could provide
idea that the brain processes information through the information on how the brain and body interact to sup-
dynamic interactions of distributed areas operating in port the meditation practice143. Electroencephalography
large-scale networks139,140. Because the complex mental feedback has been used to aid training and study medita-
state of mindfulness is probably supported by alterations tion by providing the practitioners with information on
in large-scale brain networks, future work should con- the brain waves they are producing. Similarly, real-time
sider the inclusion of complex network analyses, rather fMRI has been used to provide subjects with feedback
than restricting analyses to comparisons of the strength of the brain activity they are producing and allows the
of activations in single brain areas. Recent studies have experimenter to examine pain, cognitive control, emo-
explored functional network architecture during the tion regulation and learning of meditation. This dynamic
resting state using these new tools141,142. recording and feedback technique may help to train the
NATURE REVIEWS | NEUROSCIENCE VOLUME 16 | APRIL 2015 | 221
© 2015 Macmillan Publishers Limited. All rights reservedYou can also read
