Fnhum 07 00566

ORIGINAL RESEARCH ARTICLE
published: 18 September 2013

HUMAN NEUROSCIENCE doi: 10.3389/fnhum.2013.00566
Executive control and felt concentrative engagement

following intensive meditation training
Anthony P. Zanesco 1,2*, Brandon G. King 1,2 , Katherine A. MacLean 3 and Clifford D. Saron 2,4
1
Department of Psychology, University of California, Davis, CA, USA
2
Center for Mind and Brain, University of California, Davis, CA, USA
3
Department of Psychiatry and Behavioral Sciences, Johns Hopkins University School of Medicine, Baltimore, MA, USA
4
The M.I.N.D. Institute, UC Davis Medical Center, Sacramento, CA, USA
Edited by: Various forms of mental training have been shown to improve performance on
Wendy Hasenkamp, Mind and Life cognitively demanding tasks. Individuals trained in meditative practices, for example, show
Institute, USA
generalized improvements on a variety of tasks assessing attentional performance. A
Reviewed by:
central claim of this training, derived from contemplative traditions, posits that improved
Michael Lifshitz, McGill University,
Canada attentional performance is accompanied by subjective increases in the stability and
Sara Lazar, Massachusetts General clarity of concentrative engagement with one’s object of focus, as well as reductions
Hospital, USA in felt cognitive effort as expertise develops. However, despite frequent claims of
*Correspondence: mental stability following training, the phenomenological correlates of meditation-related
Anthony P. Zanesco, Center for Mind
attentional improvements have yet to be characterized. In a longitudinal study, we
and Brain, University of California,
267 Cousteau Place, Davis, CA assessed changes in executive control (performance on a 32-min response inhibition
95618, USA task) and retrospective reports of task engagement (concentration, motivation, and effort)
e-mail: azanesco@ucdavis.edu following one month of intensive, daily Vipassana meditation training. Compared to
matched controls, training participants exhibited improvements in response inhibition
accuracy and reductions in reaction time variability. The training group also reported
increases in concentration, but not effort or motivation, during task performance.
Critically, increases in concentration predicted improvements in reaction time variability,
suggesting a link between the experience of concentrative engagement and ongoing
fluctuations in attentional stability. By incorporating experiential measures of task
performance, the present study corroborates phenomenological accounts of stable, clear
attentional engagement with the object of meditative focus following extensive training.
These results provide initial evidence that meditation-related changes in felt experience
accompany improvements in adaptive, goal-directed behavior, and that such shifts may
reflect accurate awareness of measurable changes in performance.
Keywords: executive control, meditation, response inhibition, sustained attention, task engagement
Goal-directed behaviors requiring sustained concentration are (Hölzel et al., 2011), the extent to which observed improvements
ubiquitous in daily life. As a consequence, the ability to vol- are accompanied by corresponding changes in phenomenologi-
untarily control attention is essential for promoting academic cal aspects of attention is unknown. In the present study, we aim
and professional success, maintaining mental and physical health, to characterize training-related changes in phenomenal aware-
and building adaptive interpersonal skills (Tangney et al., 2004). ness that accompany improvements in sustained, goal-directed
But there are limits on the overall capacity to direct and con- attention following intensive meditative practice.
trol attentional resources (Kaplan and Berman, 2010). Perhaps The fluctuating nature of attention has long been acknowl-
unsurprisingly, individuals commonly find sustaining their con- edged by several Buddhist contemplative traditions (Wallace,
centration during simple tasks to be stressful and effortful (Warm 1999, 2006). These contemplative traditions have developed
et al., 2008; Langner and Eickhoff, 2013), and momentary lapses complex mental training techniques for cultivating stable atten-
can disrupt the stability of attention as the mind drifts on and off tion, developing introspective and meta-cognitive abilities, and
task over time (Weissman et al., 2006). Furthermore, individuals increasing one’s capacity for behavioral and emotional regula-
are often unaware that their attention has lapsed at all (Smallwood tion (Lutz et al., 2008). Attention training through meditative
and Schooler, 2006). There is increasing evidence suggesting practice can thus be conceived as a method for developing
that directed mental training, including meditation, may serve central attentional resources for the adaptive regulation of cog-
as one potential method to attenuate deficits in attentional sta- nition and behavior. During meditative practice, practitioners
bility (Slagter et al., 2011; Mrazek et al., 2013). Although these may employ specific focused-attention techniques to selectively
studies provide evidence that meditation training may impact maintain attention on an object of concentration, typically the
neural and behavioral markers of attention and executive control sensations of the breath and body, while monitoring the quality
Frontiers in Human Neuroscience www.frontiersin.org September 2013 | Volume 7 | Article 566 | 1

Zanesco et al. Concentrative engagement following meditation training
of ongoing awareness (e.g., clear or dull, focused or distracted). et al., 2002; Bellgrove et al., 2004; Weissman et al., 2006,
Other monitoring techniques involve awareness, introspection, 2009; Kelly et al., 2008; Prado and Weissman, 2011; Prado
and discriminative analysis of the contents of phenomenological et al., 2011). Task-unrelated cognitive processing (e.g., mind-
experience (e.g., discerning bare sensations from associated eval- wandering; Smallwood and Schooler, 2006) may also underlie
uations and judgments). Together, these focused-attention and attentional variability, as recent studies suggest an association
monitoring techniques comprise the basic methods for training in between response time variability and instances of task unrelated
Vipassana meditation (Goldstein, 1976; Goldstein and Kornfield, thought (Cheyne et al., 2009; Mrazek et al., 2012; Seli et al., 2013).
2001), from which a number of contemplative-based therapies Taken together, the available evidence suggests that ongoing vari-
are derived (e.g., Mindfulness-Based Stress Reduction; Grossman ations in attentional state may contribute significantly to observed
et al., 2004; Williams and Kabat-Zinn, 2011). behavioral variability when goal-directed attention is maintained
In line with the notion that meditative training may sup- over time.
port generalized improvements in executive control (Slagter et al., Recent studies pairing subjective measures of attentional
2011), recent longitudinal studies suggest that the ability to coor- engagement to behavioral and physiological markers of lapses
dinate one’s attention and behavior in response to task demands in attention have contributed to our understanding of ongoing
may be improved through intensive practice of meditation (Jha fluctuations in performance and sensory processing (Smallwood
et al., 2007; MacLean et al., 2010; Sahdra et al., 2011; Allen et al., 2004; Christoff et al., 2009; Kam et al., 2010; Macdonald
et al., 2012; Moore et al., 2012). For example, compared to et al., 2011). By probing whether participants were in a focused
wait-list controls, meditators who practiced 3 months of focused- (on-task) or unfocused (off-task) state, Kam et al. (2010)
attention meditation (Shamatha) demonstrated improved perfor- observed attenuated modulation of visual event-related poten-
mance accuracy in sustained-attention tasks requiring perceptual tials on trials preceding reports of unfocused states. Furthermore,
discrimination of rare targets (MacLean et al., 2010) and were bet- Macdonald et al. (2011) observed that trial-by-trial ratings
ter able to inhibit habitual pre-potent responses (Sahdra et al., of participants’ depth of focus (more vs. less focused) pre-
2011). These findings indicate improvements in two constituent dicted target discrimination and were negatively related to pre-
processes that underlie human executive control: (1) the main- stimulus alpha oscillatory power, an electrophysiological marker
tenance of attentional or perceptual resources over extended of attention commonly implicated in stimulus detection. In
periods of time, as indicated by a moderation of the rate of decline line with these findings, Lutz et al. (2002) previously demon-
in perceptual sensitivity to target stimuli (Parasuraman, 1979; strated that verbal descriptions of participants’ preparedness to
Nuechterlein et al., 1983; See et al., 1995), and (2) the ability to perceive 3D “popouts” in random dot stereograms were associ-
withhold inappropriate pre-potent response tendencies in tasks ated with increased bilateral synchronization of frontal EEG and
requiring behavioral inhibition to rare targets (Robertson et al., improved behavioral performance. By quantifying phenomeno-
1997; Ridderinkhof et al., 2004). Such findings are consistent logical aspects of attention and awareness, these studies illustrate
with a larger body of evidence suggesting that intensive med- how subjective indices of attention may be used to clarify the cog-
itation training positively impacts component processes which nitive and neural processes that contribute to overall performance
contribute to poor performance in tasks requiring sustained outcomes.
executive control (Slagter et al., 2011). Efforts to understand the experiential correlates of cognition
The maintenance of goal-directed attention over time and the may benefit from the investigation of mental training regimens
inhibition of task-inappropriate behavioral responses are thought incorporating meditative introspection. Introspective monitoring
to involve the coordinated effort of a network of brain regions techniques form a core component of training in Vipsassana med-
within frontal and parietal cortices (Miller and Cohen, 2001; itation, which may facilitate more accurate reporting of subjective
Corbetta and Shulman, 2002; Peterson and Posner, 2012; Langner mental states than would likely be obtained from individuals
and Eickhoff, 2013). Together, these component processes place untrained in the observation of internal phenomena (Varela,
considerable processing demands on attentional systems when 1996; Lutz and Thompson, 2003). Though results have been
executive control must be maintained over time (Parasuraman, mixed (Nielsen and Kaszniak, 2006; Khalsa et al., 2008), there is
1979; Nuechterlein et al., 1983; See et al., 1995). This sustained some evidence supporting the efficacy of these techniques in facil-
attentional demand leads to a decline in performance known as itating introspective accuracy (Sze et al., 2010; Fox et al., 2012).
the vigilance decrement, which is thought to reflect the depletion For instance, Fox et al. (2012) reported that meditative experience
of information processing resources that cannot be immediately is related to increased accuracy between self-reported, neural,
replenished under the constraints of the current task demands and behavioral markers of tactile sensitivity. Further, using a tac-
(Warm et al., 2008; Kaplan and Berman, 2010). tile detection task, Mirams et al. (2013) observed fewer tactile
Ongoing fluctuations in attention on a moment-to-moment misperceptions near participants’ individual sensory threshold
basis may also impact the maintenance of goal-directed attention after a 6-day brief intervention of body-scan mindfulness med-
over time. This notion is supported by an emerging consen- itation. There is also evidence that meditation training moderates
sus from behavioral and neurophysiological studies that have dynamic activation in primary sensory cortices during attentional
linked changes in functional connectivity between brain regions orienting to tactile stimuli, suggesting a role for the modula-
underlying attentional control and sensory processing to lapses tion of alpha-band oscillatory activity in processing and filtering
of attention, as reflected in variability of reaction times (RT) sensory information (Kerr et al., 2011, 2013). Thus, meditation
during ongoing task performance (Manly et al., 2000; West practice may promote meta-cognitive and interoceptive capacities

that aid practitioners in observing and describing internal mental et al., 2012). No study of intensive contemplative training has
states and experiences. With increased experience, these reports yet addressed this concern by assessing performance motivation.
should more closely mirror processes inferred from externally Behavioral improvements observed following meditation training
observable measures. In turn, increased meta-cognitive aware- interventions may thus reflect changes in a number of underlying
ness of attentional states may allow practitioners to better regulate factors, including both attention-specific processes (e.g., endoge-
their performance by recognizing and disengaging from distrac- nous focus and concentration) as well as motivational processes.
tions and endogenously moderating the stability of their ongoing The incorporation of first-person information about an individ-
attention. ual’s attentional and motivational state in studies of meditation
A central claim of contemplative training posits that improved may help clarify the relative impact of these factors on train-
attentional performance is accompanied by subjective increases ing outcomes. Taken together, the examination of these aspects
in the stability and clarity of concentrative engagement with of task engagement—concentration, effort, and motivation—
the object of meditative focus (Wallace, 1999, 2006). However, may prove useful in clarifying the experiential consequences of
there is little direct evidence detailing the potential correspon- directed mental training.
dence between states of felt concentration and improvements in In the present longitudinal study, participants completed a
executive control and attentional stability. Despite contempla- sustained response inhibition task (RIT) (Sahdra et al., 2011) and
tive (Goldstein, 1976; Wallace, 1999) and psychological (Mrazek reported on mental states related to task engagement, a construct
et al., 2012) accounts suggesting that increases in experiential encompassing felt concentration, effort, and motivation, before
concentration may parallel reductions in unwanted and intrusive (pre-test) and after (post-test) an intensive 1-month Vipassana
thoughts, feelings, and sensations, self-reported concentrative (Insight Meditation) retreat. A group of matched control par-
engagement has not been directly linked to observed performance ticipants completed identical longitudinal assessments but did
outcomes. Contemplative accounts also suggest that less cognitive not undergo training. Our first aim was to assess the effects of
effort should be required for directing and maintaining attention Vipassana meditation training on both response inhibition accu-
as expertise develops (Wallace, 1999). In line with this conception, racy and reaction time variability, measured within the same
several researchers have interpreted patterns of neural activity experimental paradigm. Our second aim was to examine changes
during attentional tasks (Lutz et al., 2009) and meditative states in task engagement as a result of training. Our third aim was to
(Brefczynski-Lewis et al., 2007; Saggar et al., 2012) as reflecting examine the contribution of potential training-related changes
decreased task-related effort and demand following meditative in subjective task experience to improvements in both executive
training. These studies, however, did not attempt to directly assess control and attentional stability.
task demand or effort.
In addition to receiving instruction in specific attentional METHODS
practices, meditation practitioners are also encouraged to cul- PARTICIPANTS
tivate an enduring motivation to engage with the teachings, Training participants underwent a 1-month intensive residential
techniques, and principles of contemplative practice globally meditation retreat held at Spirit Rock Meditation Center (SRMC)
(Goldstein, 1976; Wallace, 1999; Goldstein and Kornfield, 2001; in Woodacre, California. Twenty-eight self-selected individuals
Wallace and Shapiro, 2006). Although these conative factors likely were assessed at the beginning and end of the retreat. A compar-
promote effective training, they may also contribute to differ- ison group of 27 control participants (matched on demographic
ences in motivation between experimental conditions, which may variables and estimated lifetime and daily meditation experience;
confound interpretations of attentional improvements (Jensen see Table 1 for final participant sample) were recruited from
Table 1 | Group matching on demographic and experience variables.
Measure Control group Training group All participants t-value (df ) p-value
Age (years) 54.70 (23–72) 49.62 (25–70) 52 1.345 (47) 0.19

Sex 5 Male, 19 Female 8 Male, 18 Female 13 Male, 37 Female − −
Education 4.74 (2–6) 5.08 (4–6) 4.92 0.994 (47) 0.33
Income 7.38 (1–11) 8.23 (1–11) 7.82 0.905 (48) 0.37
Mean meditation (min/day) 31.72 (6–120) 41.52 (0–320) 36.92 0.733 (47) 0.47
Lifetime meditation (hours) 1767.46 (76–9265) 3311.52 (165–15000) 2556.64 1.753 (43) 0.09
Years of experience 9.91 (1–30) 13.67 (3–39) 11.91 1.410 (47) 0.17
Mean values and ranges are provided for demographic and meditation experience variables for the final behavioral sample (control n = 24, training n = 26). Between-
group t-tests revealed no significant differences on reported variables (all ps > 0.05) at initial assessment. Education was scored on the following scale: 1, less than
high school diploma; 2, high school diploma; 3, some college; 4, college degree; 5, some graduate study; 6, graduate degree. Total annual household income was
reported on the following scale: 1, $10,000 or less; 2, $10,001–20,000; 3, $20,001–30,000; 4, $30,001–40,000; 5, $40,001–50,000; 6, $50,001–60,000; 7, $60,001–
70,000; 8, $70,001–80,000; 9, $80,001–90,000; 10, $90,001–100,000; 11, More than $100,000. Estimated meditation experience variables included average daily
minutes of formal meditation practice during the past month, total number of lifetime hours of formal meditation practice, and number of years practicing meditation.

SRMC community meditation classes and were tested before and have a mean of 1850 ms and a range not exceeding 1550–2150 ms.
after an interval of ∼1 month (M = 27.65 days, SD = 3.51 days) A variable ISI was used to minimize the potential performance
onsite at SRMC. Control participants had previous experience benefit gained from a predictable stimulus, thereby increas-
with meditation but did not undergo intensive training during overall task demand (MacLean et al., 2009). Participants
ing the time between assessments and had not completed any responded as quickly and accurately as possible with the left
retreats up to 4 weeks prior to beginning the study. All study mouse button (right index finger) to frequent long lines (70% of
details were approved by the University of California, Davis insti- stimuli) while withholding responses to rare short lines (30% of
tutional review board. Participants gave informed consent at the stimuli) and received sound feedback through headphones (Sony
first study assessment and were debriefed at the end of the sec- MDR-V150). Auditory feedback consisted of a ding when partici-
ond assessment. Participants were compensated $120 for their pants correctly withheld their response to the short line target and
participation. a woosh when participants incorrectly withheld their response,
or failed to respond, to a long line non-target. The length of
MEDITATION TRAINING the short line was adjusted according to Parameter Estimation
Training involved a collection of techniques known as Vipassana through Sequential Testing (PEST) until converging on an overall
meditation, drawn from the Theravadan Buddhist tradi- accuracy of 75% (see details in MacLean et al., 2009).
tion (Goldstein and Kornfield, 2001). Instruction during the
retreat was provided by multiple experienced SRMC teachers. RIT
Meditation techniques involved the repeated application of atten- Next, participants completed the 32-min RIT (960 trials in total)
tion to the physical sensations of the breath, the observation with the short target line length set to each participant’s indi-
and identification of sensations, thoughts, desires, intentions, and vidual threshold. At both assessments, the length of the short
emotions, and the meta-cognitive monitoring of the quality of line was set to the participant’s pre-test threshold in order to
attention and diverse mental states. Furthermore, participants equate task parameters across assessments. Stimulus and response
engaged in a number of aspirational and emotion-generative parameters for the RIT were the same as for the threshold pro-
meditation practices emphasizing the cultivation of compassion cedure, except that target lines occurred less frequently (10% of
and loving-kindness (Salzberg, 2002) to supplement the primary all stimuli totaling 96 target lines), the length of the target line
training. Participants maintained silence during the duration remained the same throughout the task, and there was no sound
of the retreat and typically attended thirteen 45-min medita- feedback.
tion sessions each day (seven sitting sessions and six walking
sessions). Analysis
Response inhibition accuracy was quantified using the non-
TESTING PROCEDURES parametric index of perceptual sensitivity, A . When hit rate
Training group participants were tested on the morning of the is greater than false alarm rate, A is calculated as A = 0.5 +
(H − F)(1 + H − F)
first and last day of the retreat. Testing sessions took place 4H(1 − F) ; when hit rate is less than false alarm rate, A
in participants’ individual dormitory rooms. Each participant is calculated as A = 0.5 − (F − H )(1 + F − H)
4F(1 − H) (see Stanislaw and
was provided with a box containing an IBM T-40 ThinkPad Todorov, 1999; hits were defined as correct inhibitions to tar-
laptop equipped with Presentation software (Neurobehavioral gets and false alarms were defined as incorrect inhibitions to
Systems, http://www.neurobs.com) to control stimulus deliv- non-targets). This index commonly ranges from 0.5 to 1, with
ery and record behavioral responses, as well as materials and the former value reflecting chance performance and the latter
instructions for assembling the testing station. Instructions were value perfect performance. RT variability for each participant was
included for setting dim ambient lighting (e.g., blocking window quantified as the reaction time coefficient of variability (RT CV =
light and using a low-wattage lamp) and maintaining a viewing standard deviation RT/mean RT) for non-target trials. For each
distance of 57-cm from the computer screen. Control group par- participant at each assessment, A and RT CV were calculated for
ticipants underwent identical testing procedures in the same dor- the overall task and for each of eight contiguous trial blocks. Each
mitories. At each testing session, participants completed the RIT block contained 120 trials and lasted 4 min.
immediately followed by retrospective questionnaire measures of We analyzed training-related changes in A and RT CV using
task engagement. The RIT was the second of six behavioral tasks multi-level models with SAS PROC MIXED version 9.3 in order
completed at each assessment. to examine linear trajectories of growth across the eight blocks of
the RIT. Fixed effects in these models are interpreted as regression
RESPONSE INHIBITION TASK coefficients (i.e., a parameter estimate represents the expected dif-
Threshold ference in the dependent variable given a one-unit increase in
Participants first completed an ∼10-min threshold procedure to the independent variable while holding the other variables con-
calibrate task difficulty for each individual in order to equate task stant). For all analyses, independent variables representing group
demand across participants. Participants maintained eye gaze fix- (control = 0, training = 1) and assessment (pre-assessment = 0,
ation on a small dot at the center of the screen while they viewed post-assessment = 1) were treated as dummy variables. Block was
single gray vertical lines appear one at a time against a black centered to the first 4-min block (block 1 = 0) and this parame-
background. Each stimulus was presented for 150 ms. The inter- ter represents the linear trajectory (slope) of performance across
stimulus interval (ISI) varied randomly but was constrained to each 4-min segment of the RIT.

SELF-REPORT MEASURES OF TASK ENGAGEMENT p = 0.169], and no significant interaction between assessment
Immediately following the RIT, participants completed two self- and group [F(1, 48) = 0.009, p = 0.924]. These results are consis-
report measures from the Dundee Stress State Questionnaire tent with general practice effects and suggest that the groups had
(DSSQ; Matthews et al., 2002) to retrospectively assess motiva- comparable threshold values across assessments (pre control tar-
tion (14 items; e.g., “I wanted to succeed on the task,” “I felt get visual angle M = 3.51◦ , SD = 0.66◦ ; post control M = 3.94◦ ,
apathetic about my performance”) and concentration (7-item SD = 0.57◦ ; pre training M = 3.72◦ , SD = 0.63◦ ; post training
sub-scale from the 30-item thinking style questionnaire; e.g., “I M = 4.17◦ , SD = 0.48◦ ).
found it hard to maintain my concentration for more than a
short time,” “My mind wandered a great deal”) experienced dur- Mean reaction time
ing task performance. Each item was rated from 0 (“not at all”) to In order to rule out response time slowing as a factor influ-
4 (“extremely”), reverse scored items were corrected, and items encing performance improvements we examined whether groups
were summed to obtain scale scores. Thus, larger scores indi- differed in their overall mean reaction time across all non-
cate high levels of concentration or motivation. Cronbach’s alpha target trials. A repeated measures ANOVA on the within-subjects
reliability coefficients indicated acceptable levels of consistency effects of assessment (pre-assessment and post-assessment), the
among items in the concentration scale (α = 0.79 for pre- and between subjects effects of group (control and training), and
α = 0.85 for post-test), and marginally acceptable levels for the their interaction, demonstrated no significant main effects for
motivation scale (α = 0.65 for pre- and α = 0.62 for post-test). In group [F(1, 48) = 0.611, p = 0.438], assessment [F(1, 48) = 0.013,
addition, participants were asked to report: the amount of mental, p = 0.911], or their interaction [F(1, 48) = 0.368, p = 0.547].
physical, and temporal demand experienced during task perfor- These analyses show that the groups did not systematically dif-
mance; effort devoted to task performance; the degree to which fer in overall RT for the long-line stimuli (i.e., non-target trials)
they achieved their performance goals; and the perceived frustra- across assessments (pre control M = 531.52 ms, SD = 110.48 ms;
tion induced by the task by rating six independent items taken post control M = 525.89 ms, SD = 93.02 ms; pre training M =
from the NASA-TLX (Hart and Staveland, 1988) on a scale from 501.70 ms, SD = 133.25 ms; post training M = 509.87 ms, SD =
0 (“low”) to 10 (“high”). Following Matthews et al. (2002), these 102.29 ms).
six items were averaged to obtain an overall measure of task effort
and demand. Accuracy
Multi-level models were used to examine changes in response
RESULTS inhibition accuracy (A ) over the 32-min task as a function of
There were no significant differences between groups on demo- the fixed effects of task block (centered to the first 4-min block),
graphic or meditation-experience variables at pre-test (see assessment (centered to the pre-assessment), and group (cen-
Table 1). Two participants (1 training) were excluded from anal- tered to the control group). We included random effects on the
yses because performance was below or near chance at one of intercept and slope across blocks to allow for individual differ-
the two assessments (>3 SD lower than mean A of the sample, ences in initial A and the slope of A across blocks of the task.
M = 0.895, SD = 0.072), strongly suggesting these participants We first tested a model including the effects of block, group,
did not comply with or understand the task instructions, resulting and assessment.1 This model predicted a significant effect of
in an uninterpretable change in performance across assessments. block (β = −0.006, p < 0.001), indicating a decline in A over
One additional participant (control) was excluded due to near the course of the task. This parameter (β = −0.006) reflects the
chance level performance in overall A at both assessments (>3 amount of linear decline observed in mean A across each of the
SD lower than mean A ). Finally, two participants were excluded eight contiguous 4-min blocks of the task. A significant effect
due to an interruption of the testing session (control) and explicit of assessment was also found (β = 0.048, p < 0.001), suggest-
failure to comply with task instructions (training). Thus, the ing that participants improved in overall perceptual sensitivity at
final behavioral sample included 26 training and 24 compari- the second assessment. There was no effect of group (β = 0.013,
son group participants. Among these participants, individuals p = 0.387).
with incomplete questionnaire data were excluded from respec- Next we included the interaction term between assessment
tive analyses on concentration (training: n = 23; control: n = and group to investigate training-related changes in overall A .
24), effort (training: n = 23; control: n = 24), and motivation This model revealed a significant interaction between assess-
(training: n = 22; control: n = 23). ment and group (β = 0.021, p = 0.029; see Table 2 for parameter
estimates), consistent with our hypothesis that meditation train-
RIT PERFORMANCE ing would improve response inhibition accuracy. This parameter
Threshold estimate (β = 0.021) reflects the mean increase in A across
The purpose of the threshold procedure was to maintain constant assessments for the training group over and above the change
task difficulty across participants. Although we did not antici- observed for control participants. Thus, although a significant
pate training-specific changes in threshold, we tested for possible
effects of group (training vs. control) and assessment (pre-test vs. 1 Demographic variables, including participant age, estimated hours of life-
post-test) using repeated measures analysis of variance (ANOVA). time meditation experience, and years of experience, were also examined in
We found a main effect of assessment [F(1, 48) = 48.850, p < this model, but were not significant predictors of A’ or RT CV (all ps > 0.05)
0.001, η2p = 0.504], no significant effect of group [F(1, 48) =1.948, and were excluded from this and subsequent analyses.

Table 2 | Parameter estimates from models of RIT performance. assessment and group (β = −0.045, p < 0.001; see Table 2 for
parameter estimates and test statistics), consistent with our
Model and parameter Estimate (SE) Test statistic BIC
hypothesis that meditation training would reduce fluctuations in
ACCURACY (A’ ) −1826 response time. This parameter (β = −0.045) reflects the mean
Fixed effects decrease in RT CV across assessments for the training group
β0–intercept 0.890 (0.012) 76.82*** over and above the change observed for the control group. Thus,
β1–block −0.006 (0.001) 3.78*** although a significant effect of assessment was observed for con-
β2–assessment 0.037 (0.007) 5.22*** trol participants (β = −0.026, p < 0.001; pre M = 0.326 RT CV,
β3–group 0.002 (0.016) 0.88 post M = 0.299 RT CV), the overall decrease for training par-
β4–group × assessment 0.021 (0.009) 2.19* ticipants (β = −0.071, p < 0.001; pre M = 0.306 RT CV, post
Random effects M = 0.234 RT CV) was significantly greater (see Figure 1). As in
σ02 (intercept) 0.002 (0.001) 3.33*** the analysis of A , a third model including all two-way and three-
σ0, 1 (covariance) 0.001 (0.001) 0.98 way interactions between block, assessment, and group revealed
σ12 (slope) 0.001 (0.001) 2.45** a significant three-way interaction (β = −0.007, p = 0.037), sug-
σe2 (residual variance) 0.005 (0.000) 18.71*** gesting that meditation training moderated the increase in vari-
REACTION TIME VARIABILITY (RT CV) −2051 ability over blocks of the task. The addition of these variables,
Fixed effects however, did not improve model fit (BIC = −2045) over the
β0–intercept 0.304 (0.015) 19.82*** second model (BIC = −2051).
β1–block 0.005 (0.001) 3.83***
TASK ENGAGEMENT
β2–assessment −0.026 (0.006) 4.46***
β3–group −0.020 (0.023) 0.89
Repeated measures ANOVAs were used to test the effects of assess-
ment (pre-test vs. post-test) and group (training vs. control) on
β4–group × assessment −0.045 (0.008) 5.44***
each of the three measures of felt task engagement (see Table 3
Random effects
for descriptive statistics).
σ02 (intercept) 0.005 (0.001) 4.35***
σ0, 1 (covariance) 0.001 (0.001) 1.25
Concentration
σ12 (slope) 0.001 (0.001) 2.39**
For self-reported concentration, the ANOVA revealed a main
σe2 (residual variance) 0.003 (0.000) 18.71***
effect of assessment [F(1, 45) = 7.151, p = 0.010, η2p = 0.137], a
Full maximum likelihood estimates are reported for the best fitting models of non-significant effect of group [F(1, 45) = 2.102, p = 0.154], and
change in A’ and RT CV across the fixed effects of block (first 4-min block = a significant interaction between assessment and group [F(1, 45) =
0), group (control = 0, training = 1), and assessment (pre-assessment = 0, 9.445, p = 0.004, η2p = 0.173]. Post-hoc comparisons revealed a
post-assessment = 1) in the RIT (n = 50). Standard errors are reported in paren- significant increase from pre- to post-test in the training group
theses. Test statistics are reported as t-values for fixed effects estimates and [t(22) = −4.374, p < 0.001, d = 0.78], but not the control group
z-values for random effects estimates. Bayesian information criterion (BIC) is [t(23) = 0.266, p = 0.79]. The training group participants also
reported with lower values (more negative) indicating a better model fit. *p < reported greater concentration at post-test than did controls
0.05; **p < 0.01; ***p < 0.001. [t(45) = −2.662, p = 0.011, d = 0.73].
effect of assessment was observed for control participants (β = Effort and demand
0.037, p < 0.001; pre M = 0.867 A , post M = 0.904 A ), sug- For self-reported effort and demand, the ANOVA revealed no sig-
gesting probable test-retest effects, the overall change for training nificant effects of assessment [F(1, 45) = 1.997, p = 0.164], group
participants (β = 0.058, p < 0.001; pre M = 0.876 A , post M = [F(1, 45) = 0.225, p = 0.637], or their interaction [F(1, 45) =
0.934 A ) was significantly greater than for control participants 0.289, p = 0.594]. Thus, the groups did not differ in their self-
(see Figure 1). A third model including all two-way and three- reported effort during the RIT at either assessment.
way interactions between block, assessment, and group revealed
a significant three-way interaction (β = 0.009, p = 0.039), sug- Motivation
gesting that training moderated the decline in performance over For self-reported motivation, the ANOVA revealed no signifi-
blocks of the task. This final model (BIC = −1820), however, did cant effects of assessment [F(1, 43) = 0.085, p = 0.772], group
not fit the data better than the second model (BIC = −1826). [F(1, 43) = 0.197, p = 0.660], or their interaction [F(1, 43) =
0.726, p = 0.399]. Thus, levels of motivation did not differ
Reaction time variability between groups at either assessment.
A similar analysis was used to assess changes in reaction time
variability (RT CV) over the 32-min RIT. The first model demon- TASK ENGAGEMENT AS A PREDICTOR OF RIT PERFORMANCE
strated a significant effect of block (β = 0.005, p < 0.001), such Next, we examined the relation between measures of task engage-
that RT CV increased over the course of the 32-min task. In addi- ment and measures of performance on the RIT across partici-
tion, a significant effect of assessment was found (β = −0.050, pants. Individual measures of task engagement (concentration,
p < 0.001), but there was no effect of group (β = −0.033, p = effort, and motivation) were included as predictors of RIT per-
0.111). After including the interaction between assessment and formance at both pre- and post-assessment in a series of multiple
group in the model, we found a significant interaction between regressions. Predictors were entered simultaneously as a set in

1.00
0.40
Reaction Time Coefficient of Variability
0.35
●
●
0.95
●
Perceptual Sensitivity (A')

● ● ● ●
● ●
●
0.30
● ●
● ●
●
●
0.90
0.25
●
●
● ●
● ●
0.20
● ●
●
● ● ●
0.85
●
●
0.15
●
0.80
0.10
1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8
Time (4 Minute Block) Time (4 Minute Block)
● Control Pre−Testing ● Training Pre−Testing Control Post−Testing Training Post−Testing
FIGURE 1 | Mean response inhibition accuracy and reaction time variability across eight contiguous 4-min blocks of the RIT by group and
assessment.
Table 3 | Means and standard deviations for self-reported task

p = 0.072, sr2 = 0.077), whereas effort (B = −0.015, p = 0.297,
engagement.
sr2 = 0.027) and motivation (B = −0.001, p = 0.625, sr2 =
0.006) did not.
Pre-test Mean (SD) Post-test Mean (SD) In the next series of multiple regressions, we examined whether
post-test measures of task engagement explained a significant
TRAINING
amount of variance in post-test measures of RIT performance.
Concentration 20.65 (4.56) 23.65 (2.21)
At post-test, the set of task engagement predictors explained
Effort and Demand 6.02 (1.20) 5.68 (1.35)
a significant amount of variance in post-test A [R2 = 0.411,
Motivation 28.29 (7.16) 28.95 (9.93)
F(3, 41) = 9.535, p < 0.001]. Concentration, however, was the
CONTROL
only significant parameter in the model (B = 0.007, p < 0.001,
Concentration 20.50 (4.84) 20.29 (5.65)
sr2 = 0.237), whereas motivation (B = 0.001, p = 0.183, sr2 =
Effort and Demand 5.92 (0.96) 6.23 (0.91)
0.026) and effort (B = −0.006, p = 0.387, sr2 = 0.011) were
Motivation 28.21 (9.09) 26.87 (9.91) not. As with post-test A , we observed that task engagement
explained a significant amount of variance in RT CV at post-
order to explore whether overall task engagement was a predictor test [R2 = 0.443, F(3, 41) = 10.890, p < 0.001]. Once again, con-
of RIT performance. We also examined whether each individ- centration was the only significant individual parameter in the
ual predictor uniquely explained variance in measures of RIT model (B = −0.011, p < 0.001, sr2 = 0.273), while motivation
performance while controlling for the other measures of task (B = −0.001, p = 0.403, sr2 = 0.010) and effort (B = 0.013, p =
engagement. 0.168, sr2 = 0.027) were not significant predictors of RT CV.
In the first multiple regression, we examined whether pre-test Concentration was the only measure of task engagement
task engagement explained a significant amount of variance in found to be a consistent predictor of performance after control-
pre-test A . As a whole, task engagement (concentration, effort, ling for other measures of task engagement across the multiple
and motivation) was not a significant predictor of pre-test A regression analyses. In another set of regressions, we therefore
[R2 = 0.128, F(3, 41) = 1.998, p = 0.129]. The examination of followed-up these analyses examining the relation between con-
individual parameters revealed that pre-test concentration was centration and measures of RIT performance alone. At pre-test,
the only significant predictor of pre-test A in the model (B = concentration predicted a significant amount of variance in both
0.005, p = 0.049, sr2 = 0.088). Effort (B = −0.003, p = 0.782, A [R2 = 0.113, F(1, 45) = 5.707, p = 0.021] and RT CV [R2 =
sr2 = 0.002) and motivation (B = 0.001, p = 0.546, sr2 = 0.008) 0.083, F(1, 45) = 4.097, p = 0.049]. Similarly, post-test concen-
did not uniquely explain any variance in pre-test A when control- tration predicted a significant amount of variance in both post-
ling for the other predictors. Next, we examined whether pre-test test A [R2 = 0.368, F(1, 45) = 26.229, p < 0.001] and post-test
task engagement was a significant predictor of pre-test RT CV. RT CV [R2 = 0.403, F(1, 45) = 30.436, p < 0.001]. Thus, across
Task engagement did not predict a significant amount of vari- all analyses, individuals who reported more felt concentration
ance in RT CV at the first assessment [R2 = 0.116, F(3, 41) = demonstrated greater RIT accuracy and lower RT variability. The
1.787, p = 0.165]. Pre-test concentration was the only parameter addition of predictors representing group and the interaction
of task engagement that approached significance (B = −0.006, between group and concentration did not add significantly to

the explained variance in any of these models, suggesting that DISCUSSION

the relation between concentration and measures of RIT per- The present longitudinal study of intensive Vipassana med-
formance was not moderated by group membership at either itation adds to a growing body of evidence indicating that
assessment. the capacity for executive control and attentional stability may
be improved through directed mental training. We observed
PREDICTING CHANGES IN RIT PERFORMANCE FROM CHANGES IN training-related improvements in performance accuracy and
CONCENTRATIVE ENGAGEMENT decreased reaction time variability on a sustained RIT, as well as
Our third aim was to relate changes in task engagement to increases in participant-reported concentrative task engagement.
RIT performance following meditative training. As the results Critically, training-related increases in self-reported concentra-
indicated significant training-related increases in concentrative tion predicted reductions in RT variability. This suggests that
engagement, we used hierarchical multiple regressions to inves- the experience of clear and unwavering concentration may be a
tigate whether increased felt concentration predicted improve- phenomenological correlate of stable attention, reported and felt
ments in RIT performance among training group participants. in aggregate by individuals undergoing introspective meditative
Specifically, we tested the unique variance explained by changes training.
in concentration on mean A and RT CV for the training group Theories of sustained attention propose that the vigilance
(n = 23). decrement reflects the consumption of executive resources, which
In the first hierarchical multiple regression, post-test A are depleted as attention is maintained over time (Warm et al.,
served as the dependent variable. The first step included pre- 2008; MacLean et al., 2009). Consistent with prior research, we
test A and pre-test concentration as predictors in order to observed a decline in performance over the course of the 32-min
account for initial performance and concentration levels prior RIT. We also observed an increase in reaction time variabil-
to training. These predictors explained a significant amount ity as the task progressed. Thus, in addition to decrements in
of variance [R2 = 0.355, F(2, 20) = 5.493, p = 0.013], in post- perceptual sensitivity (A ), increases in reaction time variability
test A . In the second step, the addition of post-test concen- over the course of task performance may reflect an additional
tration did not add significantly to the explained variance of feature of resource depletion during sustained response inhibi-
the model [R2 = 0.027, F(1, 19) = 0.831, p = 0.374]. This tion. This depletion of executive resources over time may lead
suggests that changes in self-reported concentration did not to increased behavioral variability, as fewer resources are avail-
predict training-related improvements in response inhibition able for maintaining attention on the task set and regulating
accuracy. behavior. Reaction time variability may also result in part from
A second hierarchical multiple regression was conducted on fluctuations in the stability of attention as awareness is drawn to
post-test RT CV using the same analytic strategy. Inclusion of task-unrelated thoughts (Seli et al., 2013). Accordingly, this deple-
pre-test RT CV and pre-test concentration explained a significant tion of attentional resources might impact the ability to resist task
amount of variance in post-test RT CV [R2 = 0.623, F(2, 20) = disengagement and subsequent mind-wandering by consuming
16.503, p < 0.001]. The addition of post-test concentration to necessary executive resources.
the model explained a significant amount of unique variance in The observed improvements in response inhibition accu-
post-test RT CV [R2 = 0.079, F(1, 19) = 5.019, p = 0.037], racy are consistent with previous findings following intensive,
indicating that increases in felt concentration were predic- focused-attention (Shamatha) meditation training. Using an
tive of reductions in RT CV following training (B = −0.009, identical RIT, Sahdra et al. (2011) reported improvements in per-
p = 0.037; see Table 4 for parameter estimates). Figure 2 depicts ceptual sensitivity (overall A ) and sustained performance (slope
the magnitude and trajectory of changes in concentration of A over task duration) after ∼1.5 months of full-time medi-
and RT CV plotted separately for both training and control tation practice. While often articulated as conceptually distinct,
groups. focused-attention and monitoring techniques likely engage many
of the same attentional and executive processes (Lutz et al., 2008;
Slagter et al., 2011). In the present study, both techniques were
Table 4 | Changes in RT CV predicted by changes in felt concentration employed as components of Vipassana training. Thus, our find-
(dependent variable is RT CV at post-test). ings of improved attentional stability and response inhibition
cannot be attributed solely to the incorporation of either focused-
Step 1 (R 2 = 0.623) B SE β t-value (df ) p-value attention or monitoring techniques in the training undertaken by
our study participants. Furthermore, training appeared to moder-
Constant 0.147 0.055 − 2.701 (20) 0.014
ate the typical linear decline in ongoing performance (A and RT
RT CV at pre-test 0.435 0.100 0.670 4.364 (20) 0.001
CV) over the duration of the task. Although this result is theoret-
Concentration at pre-test −0.002 0.002 −0.215 1.398 (20) 0.177
ically important, the statistical models including the interaction
Step 2 (R 2 = 0.702) of training and performance over blocks did not improve model
fit compared to models considering only overall training-related
Constant 0.320 0.092 − 3.489 (19) 0.002 changes. It remains unclear, however, whether these behavioral
RT CV at pre-test 0.394 0.093 0.606 4.242 (19) 0.001 results are mediated by training-related changes in some other
Concentration at pre-test 0.001 0.002 0.070 0.369 (19) 0.717 attentional or cognitive process, such as visual working memory
Concentration at post-test −0.009 0.004 −0.424 2.240 (19) 0.037 capacity which was not measured in the current study.

Control Group (n = 24) Training Group (n = 23)
0.6
0.6
●

●
0.5
0.5
●
0.4
0.4
●
● ● ●
● ●
● ●
0.3
0.3
●
●
● ●
●
●
●
●
●
● ●
● ●
●
● ● ●
● ● ● ● ●
● ● ●
0.2
0.2
● ● ●
● ●
●
● ● ●
● ●
0.1
0.1
5 10 15 20 25 30 5 10 15 20 25 30
Self−Reported Concentration Self−Reported Concentration
FIGURE 2 | Plot of change across pre- and post-assessments for reaction time variability and concentration for training and control group
participants. Each vector represents the trajectory from pre- to post-assessment scores for a given individual. Post-assessment scores are represented by a
dot. Thus, each vector represents the direction and magnitude of change in these variables for each individual.
Although reaction time variability fails to encode the full com- that meditation-related improvements in tasks requiring directed
plexity of ongoing sensorimotor dynamics over the course of an attention and executive control may be supported by cortical
experiment, it nonetheless reflects a useful measure of variability plasticity and increased connectivity in regions associated with
in attention on a trial-to-trial basis. The current findings are con- sensory processing, interoceptive awareness, and the voluntary
sistent with previous studies in which reductions in RT variability control and maintenance of attention (Farb et al., 2013; Kerr et al.,
were observed following intensive meditation interventions (Lutz 2013). Of particular relevance to the study of phenomenological
et al., 2009; van Vugt and Jha, 2011). These studies, however, correlates of attentional states, connectivity and cortical plastic-
did not report evidence for improved performance accuracy ity in regions supporting executive and interoceptive processes,
alongside reductions in variability. Notably, Lutz et al. (2009) such as insular cortex, may support greater subjective awareness
investigated the neural correlates of improvements in attentional of mental states and bodily experiences (Craig, 2009).
stability using an auditory sustained-attention paradigm. They The present findings are also consistent with studies of
demonstrated reductions in RT variability and enhanced phase non-meditative interventions demonstrating improvements in
consistency of oscillatory neural responses in the theta band executive control in individuals trained to regulate habitual
following 3 months of Vipassana training. Increased theta phase- behavior. For example, Muraven (2010) reported improvements
locking to stimulus onset over anterior scalp regions predicted in response inhibition following 2 weeks of practiced self-
the reduction in RT variability. Thus, the ongoing engagement of control over common urges in daily life (e.g., avoiding tempting,
attention may entrain oscillatory neural responses to task-related unhealthy food). Intensive meditation retreats typically involve
sensory input, suggesting a potential neural correlate of observed continuous behavior regulation, such as walking and eating
improvements in attentional stability. slowly, maintaining meditative posture, and refraining from con-
Meditation experience has been linked to increased func- versation and eye contact with one’s peers. Thus, learning to reg-
tional coupling between the posterior cingulate, dorsal anterior ulate habitual daily behaviors may facilitate an enhanced capacity
cingulate, and dorsolateral prefrontal cortex, which are regions to inhibit pre-potent response tendencies, independent of mental
that have been consistently implicated in processes of attention training exercises. Other non-specific factors, such as sustained
and executive control (Brewer et al., 2011). Similarly, increased exposure to the natural wilderness setting of the residential retreat
connectivity has been observed between bilateral regions of the center, may have provided additional cognitive benefits (e.g.,
dorsolateral prefrontal cortex and the right insula in meditation Berman et al., 2008). One possibility is that attention may be
practitioners (Farb et al., 2007; Hasenkamp and Barsalou, 2012). regulated exogenously by the modest stimulation of the nat-
Finally, both cross-sectional (e.g., Lazar et al., 2005) and longi- ural environments, allowing attentional resources to replenish
tudinal (e.g., Hölzel et al., 2011) studies have demonstrated that (Kaplan, 1995; Berman et al., 2008; Kaplan and Berman, 2010).
Vipassana meditation is associated with increased cortical thick- Open, or non-directed, monitoring of internal experiences, as
ness in prefrontal, anterior cingulate, insular, and somatosensory might be practiced during Vipassana meditation, may have simi-
cortices. These neuroimaging results provide indirect evidence lar recuperating effects, as attention is lightly engaged and drawn

to the natural internal milieu. An important avenue for future stability. This interpretation is supported by training group par-
research will be to clarify the contribution of non-specific fac- ticipants reporting higher mean levels of self-reported concentra-
tors of the meditation-training environment and regimen to the tion than controls following training. Furthermore, the strength
improvement of executive control. and direction of the association between concentration and RT
By incorporating experiential measures of task engagement, variability did not differ between groups at post-assessment. The
the present study corroborates contemplative phenomenological question of increased introspective accuracy, however, cannot be
accounts of stable, clear concentration while directing atten- definitely answered with the current study design. Future studies
tion following extensive training (Goldstein, 1976; Wallace, 1999; should determine if training-related improvements in introspec-
Goldstein and Kornfield, 2001). At both assessments and for both tive accuracy are found independent of corresponding improve-
groups, the experience of concentrative focus accounted for sub- ments in performance by including measures not sensitive to
stantial variance in accuracy and behavioral variability in the training.
sustained RIT. Furthermore, although self-reported concentra- There is increasing concern that studies of contemplative
tion related to performance, we observed no group difference in training may be susceptible to motivational confounds (Jensen
this pattern, contrary to what might be expected given a poten- et al., 2012). It is often presumed that self-selected samples are
tial role for meditation training in facilitating the introspection of highly motivated to perform well and report personal bene-
mental states. We speculate that such a difference might have been fits of training. In particular, practitioners of meditation may
observed more readily had we compared participants trained in hold beliefs about the efficacy of practice that bias their perfor-
meditation to those inexperienced with the techniques, consis- mance by encouraging them to devote more cognitive resources
tent with findings from previous cross-sectional studies (Sze et al., to accomplish performance goals. Although self-selection was a
2010; Fox et al., 2012). This pattern, observed across all sample potential confound in the present study, we observed no dif-
participants, suggests a strong association between subjective feel- ferences in task-specific motivation and task effort between the
ings of concentration and task performance in the overall sample training group and the meditation-experience-matched control
of experienced meditators. This indicates that mental states of group. Furthermore, motivation and cognitive demand were not
concentration may comprise a potentially important explanatory consistent predictors of performance outcomes, challenging the
construct for understanding individual differences in attentional notion that these factors have strong influence on behavioral
and executive lapses. outcomes among experienced meditators. This is in contrast
Training appeared to increase the quality of concentrative to cross-sectional studies comparing naïve and novice partici-
engagement, which in turn predicted improved behavioral sta- pants (see Jensen et al., 2012). Our use of demographic and
bility. Nonetheless, increases in felt concentration did not predict meditation-experience-matched training and control groups may
improvements in response inhibition accuracy. Thus, when retro- have allowed us to equate levels of motivation, as both groups
spectively considering aggregate levels of felt concentration, one’s presumably shared similar convictions and biases regarding the
sense of moment-by-moment fluctuations in attention over trials beneficial effects of meditation practice.
may be more phenomenologically accessible than other experien- The groups were adequately matched on both pre-test perfor-
tial features of ongoing awareness. In contrast, increases in target mance and overall meditation experience, however, the measures
discrimination and inhibition of pre-potent responses presum- of meditation-experience used here to match participants may
ably reflect the more indirect consequences of task attentiveness. not be reliable indicators of participants’ lifetime experience.
Therefore, levels of performance accuracy may serve as a less Some participants reported difficulty in attempting to estimate
salient cue when inferring one’s attentional state than the ebb and their lifetime meditation experience. Failure to adequately match
flow of awareness across trials. However, despite training-related participants on past meditation experience may complicate the
differences in subjective and behavioral markers of concentra- interpretation of findings from longitudinal studies of medita-
tion, there were no training-related changes in the felt experience tion training, as more experienced practitioners may have higher
of mental effort and demand during task performance. Because levels of performance at pre-assessment and may be differentially
reports of felt concentration, motivation, and effort were col- affected by the training itself. Furthermore, without empirically
lected retrospectively, after completion of the task, our findings delineating the type, quality, and meaning of estimates of life-
do not speak directly to participants’ moment-to-moment expe- time hours of meditation practice, the utility of these measures in
rience. It will be interesting to examine whether meditation informing research investigations of meditation training remains
practitioners are aware of ongoing fluctuations in attentional and uncertain.
motivational states using online reporting and experience sam- An intriguing direction for future research may be to utilize
pling techniques, as opposed to aggregated retrospective reports. alternative cognitive training regimes as active comparison con-
It is unclear, however, whether increased introspective accu- ditions in studies of meditation training. Although task-specific
racy contributes to the observed association between changes training rarely leads to generalizable improvements in atten-
in felt concentration and attentional stability. It is possible that tion and executive control, some forms of cognitive training
training participants’ self-reports of concentration at the post- such as musical training, action video games, working memory
assessment more accurately reflected their measures of perfor- training, and meditation practice have each emerged as poten-
mance than did their reports at pre-assessment. In contrast, tial paradigms shown to impact general cognitive processes not
we interpret our findings as reflective of concomitant increases explicitly trained during practice (Green and Bavelier, 2008;
in phenomenological and behavioral correlates of attentional Jaeggi et al., 2008; Bavelier and Davidson, 2013). It will be

important for researchers to begin to compare the effectiveness advance our understanding of the phenomenological correlates
of different training regimes in improving attention and meta- of cognitive processes, and in particular, change resulting from
cognitive awareness. Furthermore, the inclusion of wait-list and intensive mental training.
active control comparison conditions in longitudinal studies of
meditation may be useful in examining the unique or overlapping ACKNOWLEDGMENTS
effects of meditation practice while experimentally controlling We thank the Spirit Rock Meditation Center staff and teachers:
non-specific factors such as the training environment. Kevin Duggan, Jack Kornfield, Sylvia Boorstein, Guy Armstrong,
Taken together, our findings support the view that Vipassana Gil Fronsdal, John Travis, James Baraz, Sally Armstrong, Diana
meditation training facilitates the efficient management of atten- Winston, Tony Bernhard; and our research participants. We
tional resources and is accompanied by experiential changes in thank Stephen Aichele, Tonya Jacobs, Anahita Hamidi, and Erika
feelings of attentional stability and clarity that correspond to mea- Rosenberg for their continued feedback and help in implement-
sureable changes in sensorimotor performance (RT variability). ing this program of research.
Further, these attentional benefits occur without concomitant
increases in self-perceived effort or motivation. It will be use- FUNDING
ful for future studies to further explore the utility of meditative This study was funded by Mind and Life Institute Varela
introspection as a means of refining phenomenological investiga- Contemplative Science Research Grant Award No. 09-000107
tions of cognition (Lutz et al., 2002; Lutz and Thompson, 2003). to Anthony P. Zanesco and Brandon G. King, gifts from the
Combining phenomenological, behavioral, and electrophysiolog- Baumann Foundation and Tan Teo Foundations to Clifford D.
ical approaches is crucial to advancing our understanding of the Saron, Fetzer Institute Grant No. 2191 and John Templeton
brain mechanisms underlying attentional stability and the impact Foundation Grant No. 39970 to Clifford D. Saron, and a National
of meditative training on cognition. The present study under- Science Foundation Predoctoral Fellowship to Katherine A.
scores the importance of examining felt experience in order to MacLean.
REFERENCES Sci. U.S.A. 108, 20254–20259. doi: modes of self-reference. Soc. Cogn. Front. Hum. Neurosci. 6:38. doi:
Allen, M., Dietz, M., Blair, K. S., van 10.1073/pnas.1112029108 Affect. Neurosci. 2, 313–322. doi: 10.3389/fnhum.2012.00038
Beek, M., Rees, G., Vestergaard- Cheyne, J. A., Solman, G. J. F., 10.1093/scan/nsm030 Hölzel, B. K., Lazar, S. W., Gard, T.,
Poulsen, P., et al. (2012). Carriere, J. S. A., and Smilek, Fox, K. C. R., Zakarauskas, P., Dixon, Schuman-Olivier, Z., Vago, D. R.,
Cognitive-affective neural plasticity D. (2009). Anatomy of an error: M., Ellamil, M., Thompson, E., and and Ott, U. (2011). How does
following active-controlled mind- a bidirectional state model of Christoff, K. (2012). Meditation mindfulness meditation work.
fulness intervention. J. Neurosci. task engagement/disengagement experience predicts introspective Proposing mechanisms of action
32, 15601–15610. doi: 10.1523/ and attention-related errors. accuracy. PLoS ONE 7:e45370. doi: from a conceptual and neural
JNEUROSCI.2957-12.2012 Cognition 111, 98–113. doi: 10.1371/journal.pone.0045370 perspective. Perspect. Psychol. Sci. 6,
Bavelier, D., and Davidson, R. J. (2013). 10.1016/j.cognition.2008.12.009 Goldstein, J. (1976). The Experience of 537–559. doi: 10.1177/1745691611
Brain training: games to do you Christoff, K., Gordon, A. M., Insight: A Simple and Direct Guide 419671
good. Nature 494, 425–426. doi: Smallwood, J., Smith, R., and to Buddhist Meditation. Boston, Jaeggi, S. M., Buschkuehl, M.,
10.1038/494425a Schooler, J. W. (2009). Experience MA: Shambhala. Jonides, J., and Perrig, W. J.
Bellgrove, M. A., Hester, R., and sampling during fMRI reveals Goldstein, J., and Kornfield, J. (2001). (2008). Improving fluid intelli-
Garavan, H. (2004). The func- default network and executive Seeking the Heart of Wisdom: the gence with training on working
tional neuroanatomical correlates system contributions to mind Path of Insight Meditation. Boston, memory. Proc. Natl. Acad. Sci.
of response variability: evidence wandering. Proc. Natl. Acad. MA: Shambhala. U.S.A. 105, 6829–6833. doi:
from a response inhibition task. Sci. U.S.A. 106, 8719–8724. doi: Green, C. S., Bavelier, D. (2008). 10.1073/pnas.0801268105
Neuropsychologia 42, 1910–1916. 10.1073/pnas.0900234106 Excercising your brain: a review Jensen, C. G., Vangkilde, S., Frokjaer,
doi: 10.1016/j.neuropsychologia. Corbetta, M., and Shulman, G. L. of human brain plasticity and V., and Hasselbach, S. G. (2012).
2004.05.007 (2002). Control of goal-directed training-induced learning. Psychol. Mindfulness training affects
Berman, M. G., Jonides, J., and Kaplan, and stimulus-driven attention in Aging 23, 692–701. doi: 10.1037/ attention—or is it attentional effort.
S. (2008). The cognitive benefits of the brain. Nat. Rev. Neurosci. 3, a0014345 J. Exp. Psychol. Gene. 141, 106–123.
interacting with nature. Psychol. Sci. 201–216. doi: 10.1038/nrn755 Grossman, P., Niemann, L., Schmidt, doi: 10.1037/a0024931
19, 1207–1212. doi: 10.1111/j.1467- Craig, A. D. (2009). How do you feel – S., and Walach, H. (2004). Jha, A. P., Krompinger, J., and Baime,
9280.2008.02225.x now. The anterior insula and human Mindfulness-based stress reduction M. J. (2007). Mindfulness train-
Brefczynski-Lewis, J. A., Lutz, A., awareness. Nat. Rev. Neurosci. 10, and health benefits: a meta-analysis. ing modifies subsystems of
Schaefer, H. S., Levinson, D. B., 59–70. doi: 10.1038/nrn2555 J. Psychosom. Res. 57, 35–43. doi: attention. Cogn. Affect. Behav.
and Davidson, R. J. (2007). Neural Farb, N. A. S., Segal, Z. V., 10.1016/S0022-3999(03)00573-7 Neurosci. 7, 109–119. doi:
correlates of attentional expertise and Anderson, A. K. (2013). Hart, S. G., and Staveland, L. E. (1988). 10.3758/CABN.7.2.109
in long-term meditation prac- Mindfulness meditation training “Development of NASA-TLX (Task Kam, J. W. Y., Dao, E., Farley, J.,
titioners. Proc. Natl. Acad. Sci. alters cortical representations of Load Index): results of empirical, Fitzpatrick, K., Smallwood, J.,
U.S.A. 104, 11483–11488. doi: interoceptive attention. Soc. Cogn. and theoretical research,” in Human Schooler, J. W., et al. (2010). Slow
10.1073/pnas.0606552104 Affect. Neurosci. 8, 15–26. doi: Mental Workload, eds P. A. Hancock fluctuations in attentional con-
Brewer, J. A., Worhunsky, P. D., Gray, 10.1093/scan/nss066 and N. Meshkati (Amsterdam: trol of sensory cortex. J. Cogn.
J. R., Tang, Y. Y., Weber, J., and Farb, N. A. S., Segal, Z. V., Mayberg, North Holland), 139–183. Neurosci. 23, 460–470. doi:
Kober, H. (2011). Meditation expe- H., Bean, J., McKeon, D., Fatima, Hasenkamp, W., and Barsalou, L. 10.1162/jocn.2010.21443
rience is associated with differences Z., et al. (2007). Attending to W. (2012). Effects of meditation Kaplan, S. (1995). The restorative ben-
in default mode network activity the present: mindfulness med- experience on functional connec- efits of nature: toward an integra-
and connectivity. Proc. Natl. Acad. itation reveals distinct neural tivity of distributed brain networks. tive framework. J. Environ. Psychol.

15, 169–182. doi: 10.1016/0272- (2008). Attention regulation Mrazek, M. D., Franklin, M. S., Phillips, Robertson, I. H., Manly, T., Andrade,
4944(95)90001-2 and monitoring in meditation. D. T., Baird, B., and Schooler, J. J., Baddeley, B. T., and Yiend, J.
Kaplan, S., and Berman, M. G. Trends Cogn. Sci. 12, 163–169. doi: W. (2013). Mindfulness train- (1997). “Oops!”: performance cor-
(2010). Directed attention as a 10.1016/j.tics.2008.01.005 ing improves working memory relates of everyday attentional fail-
common resource for executive Lutz, A., Slagter, H. A., Rawlings, N. capacity and GRE performance ures in traumatic brain injured and
functioning and self-regulation. B., Francis, A. D., Greischar, L. L., while reducing mind wandering. normal subjects. Neuropsychologia
Perspect. Psychol. Sci. 5, 43–57. doi: and Davidson, R. J. (2009). Mental Psychol. Sci. 24, 776–781. doi: 35, 747–758. doi: 10.1016/S0028-
10.1177/1745691609356784 training enhances attentional stabil- 10.1177/0956797612459659 3932(97)00015-8
Kelly, A. M. C., Uddin, L. Q., Biswal, ity: neural and behavioral evidence. Mrazek, M. D., Smallwood, J., and Saggar, M., King, B. G., Zanesco,
B. B., Castellanos, F. X., and J. Neurosci. 29, 13418–13427. doi: Schooler, J. W. (2012). Mindfulness A. P., MacLean, K. A., Aichele,
Milham, M. P. (2008). Competition 10.1523/JNEUROSCI.1614-09.2009 and mind wandering: finding con- S. R., Jacobs, T. L., et al. (2012).
between functional brain networks Macdonald, J. S. P., Mathan, S., and vergence through opposing con- Intensive training induces lon-
mediates behavioral variability. Yeung, N. (2011). Trial-by-trial structs. Emotion 12, 442–448. doi: gitudinal changes in meditation
Neuroimage 39, 527–537. doi: variations in subjective attentional 10.1037/a0026678 state-related EEG oscillatory activ-
10.1016/j.neuroimage.2007.08.008 state are reflected in ongoing Muraven, M. (2010). Building self- ity. Front. Hum. Neurosci. 6:256.
Kerr, C. E., Jones, S. R., Wan, Q., prestimulus EEG alpha oscilla- control strength: Practicing doi: 10.3389/fnhum.2012.00256
Pritchett, D. L., Wasserman, R. tions. Front. Psychol. 2:82. doi: self-control leads to improved Sahdra, B. K., MacLean, K. A., Ferrer,
H., Wexler, A., et al. (2011). 10.3389/fpsyg.2011.00082 self-control performance. J. Exp. E., Shaver, P. R., Rosenberg, E. L.,
Effects of mindfulness meditation MacLean, K. A., Aichele, S. R., Bridwell, Soc. Psychol. 46, 465–468. doi: Jacobs, T. L., et al. (2011). Enhanced
training on anticipatory alpha D. A., Mangun, G. R., Wojciulik, 10.1016/j.jesp.2009.12.011 response inhibition during inten-
modulation in primary somatosen- E., and Saron, C. D. (2009). Nielsen, L., and Kaszniak, A. W. sive meditation training predicts
sory cortex. Brain Res. Bull. 85, Interactions between endoge- (2006). Awareness of subtle emo- improvements in self-reported
96–103. doi: 10.1016/j.brainresbull. nous and exogenous attention tional feelings: a comparison of adaptive socioemotional function-
2011.03.026 during vigilance. Attent. Percept. long-term meditators and nonmed- ing. Emotion 11, 299–312. doi:
Kerr, C. E., Sacchet, M. D., Lazar, S. Psychophys. 71, 1042–1058. doi: itators. Emotion 6, 392–405. doi: 10.1037/a0022764
W., Moore, C. I., and Jones, S. 10.3758/APP.71.5.1042 10.1037/1528-3542.6.3.392 Salzberg, S. (2002). Lovingkindness:
R. (2013). Mindfulness starts with MacLean, K. A., Ferrer, E., Aichele, Nuechterlein, K. H., Parasuraman, The Revolutionary art of Happiness.
the body: somatosensory atten- S. R., Bridwell, D. A., Zanesco, R., and Jiang, Q. (1983). Visual Boston, MA: Shambhala.
tion and cortical alpha modula- A. P., Jacobs, T. L., et al. (2010). sustained attention: Image degra- See, J. E., Howe, S. R., Warm, J. S.,
tion in mindfulness meditation. Intensive meditation training dation produces rapid sensitivity and Dember, W. N. (1995). Meta-
Front. Hum. Neurosci. 7:12. doi: improves perceptual discrimi- decrement over time. Science 220, analysis of the sensitivity decre-
10.3389/fnhum.2013.00012 nation and sustained attention. 327–329. doi: 10.1126/science. ment in vigilance. Psychol. Bull.
Khalsa, S. S., Rudrauf, D., Damasio, A. Psychol. Sci. 21, 829–839. doi: 6836276 117, 230–249. doi: 10.1037/0033-
R., Davidson, R. J., Lutz, A., and 10.1177/0956797610371339 Parasuraman, R. (1979). Memory load 2909.117.2.230
Tranel, D. (2008). Interoceptive Manly, T., Davison, B., Heutink, J., and event rate control sensitiv- Seli, P., Cheyne, J. A., and Smilek,
awareness in experienced medita- Galloway, M., and Robertson, I. ity decrements in sustained atten- D. (2013). Wandering minds
tors. Psychophysiology 45, 671–677. H. (2000). Not enough time or tion. Science 205, 924–927. doi: and wavering rhythms: linking
doi: 10.1111/j.1469-8986.2008. not enough attention. Speed, error 10.1126/science.472714 mind wandering and behavioral
00666.x and self-maintained control in the Peterson, S. E., and Posner, M. I. variability. J. Exp. Psychol. Hum.
Langner, R., and Eickhoff, S. B. (2013). Sustained Attention to Response (2012). The attention system of Percept. Perform. 39, 1–5. doi:
Sustaining attention to simple tasks: Test (SART). Clin. Neuropsychol. the human brain: 20 years after. 10.1037/a0030954
a meta-analytic review of the neu- Assess. 3, 167–177. Annu. Rev. Neurosci. 35, 73–89. Slagter, H. A., Davidson, R. J., and
ral mechanisms of vigilant atten- Matthews, G., Campbell, S. E., doi: 10.1146/annurev-neuro- Lutz, A. (2011). Mental training as
tion. Psychol. Bull. 139, 870–900. Falconer, S., Joyner, L. A., Huggins, 062111-150525 a tool in the neuroscientific study
doi: 10.1037/a0030694 J., Gilliland, K., et al. (2002). Prado, J., Carp, J., and Weissman, D. of brain and cognitive plasticity.
Lazar, S. W., Kerr, C., Wasserman, R. H., Fundamental dimensions of subjec- H. (2011). Variations of response Front. Hum. Neurosci. 5:17. doi:
Gray, J. R., Greve, D., Treadway, M. tive state in performance settings: time in a selective attention 10.3389/fnhum.2011.00017
T., et al. (2005). Meditation expe- task engagement, distress, and task are linked to variations of Smallwood, J., Davies, J. B., Heim,
rience is associated with increased worry. Emotion 2, 315–340. doi: functional connectivity in the D., Finnigan, F., Sudberry, M.
cortical thickness. Neuroreport 16, 10.1037/1528-3542.2.4.315 attentional network. Neuroimage V., O’Connor, R. C., et al.
1893–1897. doi: 10.1097/01.wnr. Miller, E. K., and Cohen, J. D. (2001). 54, 541–549. doi: 10.1016/ (2004). Subjective experience
0000186598.66243.19 An integrative theory of pre- j.neuroimage.2010.08.022 and the attentional lapse. Task
Lutz, A., Lachaux, J. P., Martinerie, J., frontal cortex functioning. Annu. Prado, J., and Weissman, D. H. (2011). engagement and disengage-
and Varela, F. J. (2002). Guiding Rev. Neurosci. 24, 167–202. doi: Spatial attention influences trial- ment during sustained attention.
the study of brain dynamics by 10.1146/annurev.neuro.24.1.167 by-trial relationships between Conscious. Cogn. 13, 657–690. doi:
using first-person data: synchrony Mirams, L., Poliakoff, E., Brown, response time and functional 10.1016/j.concog.2004.06.003
patterns correlate with ongoing R. J., and Lloyd, D. M. (2013). connectivity in the visual cortex. Smallwood, J., and Schooler, J.
conscious states during a sim- Brief body-scan meditation prac- Neuroimage 54, 465–473. doi: W. (2006). The restless mind.
ple visual task. Proc. Natl. Acad. tice improves somatosensory 10.1016/j.neuroimage.2010.08.038 Psychol. Bull. 132, 946–958. doi:
Sci. U.S.A. 99, 1586–1591. doi: perceptual decision making. Ridderinkhof, K. R., van den 10.1037/0033-2909.132.6.946
10.1073/pnas.032658199 Conscious. Cogn. 22, 348–359. doi: Wildenberg, W. P. M., Segalowitz, Stanislaw, H., and Todorov, N. (1999).
Lutz, A., and Thompson, E. (2003). 10.1016/j.concog.2012.07.009 S. J., and Carter, C. S. (2004). Calculation of signal detection the-
Neurophenomenology: integrating Moore, A., Gruber, T., Derose, J., and Neurocognitive mechanisms of ory measures. Behav. Res. Methods
subjective experience and brain Malinowski, P. (2012). Regular, cognitive control: the role of pre- Instr. Comp. 31, 137–149. doi:
dynamics in the neuroscience of brief mindfulness meditation prac- frontal cortex in action selection, 10.3758/BF03207704
consciousness. J. Conscious. Stud. tice improves electrophysiological response inhibition, performance Sze, J. A., Gyurak, A., Yuan, J. W., and
10, 31–52. markers of attentional control. monitoring, and reward-based Levenson, R. W. (2010). Coherence
Lutz, A., Slagter, H. A., Dunne, Front. Hum. Neurosci. 6:18. doi: learning. Brain Cogn. 56, 129–140. between emotional experience and
J. D., and Davidson, R. J. 10.3389/fnhum.2012.00018 doi: 10.1016/j.bandc.2004.09.016 physiology: does body awareness

training have an impact. Emotion Wallace, B. A. (2006). The Attention irrelevant stimuli. Neuroimage 48, Received: 01 May 2013; paper pend-
10, 803–814. doi: 10.1037/a0020146 Revolution. Boston, MA: 609–615. doi: 10.1016/j.neuro ing published: 10 July 2013; accepted:
Tangney, J. P., Baumeister, R. F., Wisdom. image.2009.06.081 25 August 2013; published online: 18
and Boone, A. L. (2004). High Wallace, B. A., and Shapiro, S. L. West, R., Murphy, K. J., Armilio, September 2013.
self-control predicts good adjust- (2006). Mental balance and M. L., Craik, F. I. M., and Stuss, Citation: Zanesco AP, King BG,
ment, less pathology, better well-being: building bridges D. T. (2002). Lapses of inten- MacLean KA and Saron CD (2013)
grades, and interpersonal suc- between Buddhism and Western tion and performance variability Executive control and felt concentra-
cess. J. Pers. 72, 271–324. doi: psychology. Am. Psychol. 61, reveal age-related increases in tive engagement following intensive
10.1111/j.0022-3506.2004.00263.x 690–701. doi: 10.1037/0003-066X. fluctuations of executive control. meditation training. Front. Hum.
van Vugt, M. K., and Jha, A. P. 61.7.690 Brain Cogn. 49, 402–419. doi: Neurosci. 7:566. doi: 10.3389/fnhum.
(2011). Investigating the impact of Warm, J. S., Parasuraman, R., and 10.1006/brcg.2001.1507 2013.00566
mindfulness meditation training on Matthews, G. (2008). Vigilance Williams, J. M. G., and Kabat-Zinn, This article was submitted to the journal
working memory: a mathematical requires hard mental work and J. (2011). Mindfulness: diverse Frontiers in Human Neuroscience.
modeling approach. Cogn. Affect. is stressful. Hum. Factors 50, perspectives on its meaning, ori- Copyright © 2013 Zanesco, King,
Behav. Neurosci. 11, 344–353. doi: 433–441. doi: 10.1518/0018720 gins, and multiple applications MacLean and Saron. This is an open-
10.3758/s13415-011-0048-8 08X312152 at the intersection of science and access article distributed under the terms
Varela, F. J. (1996). Weissman, D. H., Roberts, K. C., dharma. Contemp. Buddhism 12, of the Creative Commons Attribution
Neurophenomenology: a method- Visscher, K. M., and Woldorff, 1–18. doi: 10.1080/14639947. License (CC BY). The use, distribution
ological remedy for the hard M. G. (2006). The neural bases 2011.564811 or reproduction in other forums is per-
problem. J. Conscious. Stud. 3, of momentary lapses in attention. mitted, provided the original author(s)
330–349. Nat. Neurosci. 9, 971–978. doi: Conflict of Interest Statement: The or licensor are credited and that the
Wallace, B. A. (1999). The Buddhist 10.1038/nn1727 authors declare that the research original publication in this journal
tradition of Samatha: methods Weissman, D. H., Warner, L. M., was conducted in the absence of any is cited, in accordance with accepted
for refining and examining con- and Woldorff, M. G. (2009). commercial or financial relationships academic practice. No use, distribution
sciousness. J. Conscious. Stud. 6, Momentary reductions of atten- that could be construed as a potential or reproduction is permitted which does
175–187. tion permit greater processing of conflict of interest. not comply with these terms.

Fnhum 07 00566

Uploaded by

Copyright:

Available Formats

Fnhum 07 00566

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Fnhum 07 00566

Uploaded by

Copyright:

Available Formats

ORIGINAL RESEARCH ARTICLE

published: 18 September 2013

Executive control and felt concentrative engagement

Frontiers in Human Neuroscience www.frontiersin.org September 2013 | Volume 7 | Article 566 | 1

Frontiers in Human Neuroscience www.frontiersin.org September 2013 | Volume 7 | Article 566 | 2

Table 1 | Group matching on demographic and experience variables.

Age (years) 54.70 (23–72) 49.62 (25–70) 52 1.345 (47) 0.19

Frontiers in Human Neuroscience www.frontiersin.org September 2013 | Volume 7 | Article 566 | 3

Frontiers in Human Neuroscience www.frontiersin.org September 2013 | Volume 7 | Article 566 | 4

Frontiers in Human Neuroscience www.frontiersin.org September 2013 | Volume 7 | Article 566 | 5

Frontiers in Human Neuroscience www.frontiersin.org September 2013 | Volume 7 | Article 566 | 6

Perceptual Sensitivity (A')

Time (4 Minute Block) Time (4 Minute Block)

● Control Pre−Testing ● Training Pre−Testing Control Post−Testing Training Post−Testing

Table 3 | Means and standard deviations for self-reported task

Frontiers in Human Neuroscience www.frontiersin.org September 2013 | Volume 7 | Article 566 | 7

the explained variance in any of these models, suggesting that DISCUSSION

Frontiers in Human Neuroscience www.frontiersin.org September 2013 | Volume 7 | Article 566 | 8

Control Group (n = 24) Training Group (n = 23)

Reaction Time Coefficient of Variability

Reaction Time Coefficient of Variability

Frontiers in Human Neuroscience www.frontiersin.org September 2013 | Volume 7 | Article 566 | 9

Frontiers in Human Neuroscience www.frontiersin.org September 2013 | Volume 7 | Article 566 | 10

Frontiers in Human Neuroscience www.frontiersin.org September 2013 | Volume 7 | Article 566 | 11

Frontiers in Human Neuroscience www.frontiersin.org September 2013 | Volume 7 | Article 566 | 12

Frontiers in Human Neuroscience www.frontiersin.org September 2013 | Volume 7 | Article 566 | 13

You might also like

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.