The effect of a meditation retreat on two distinct modes of the self:
The narrative and the experiential.
Annabel Gilbert 12830984
University of Amsterdam 31.07.2020
Supervisor: Dr. Ruben Laukkonen Examiner: Dr. Filip van Opstal
Table of Contents Abstract... 4 Covid-19 Constraints... 5 Introduction... 6 Meditation... 7 Vipassanā... 8 Self-Referential Processing... 9 Previous Research... 11
Aim of this Study... 11
Materials and Methods... 14
Participants... 14
Design and Procedure... 14
Meditation Retreat... 14
Questionnaires... 15
EEG Experiment (FTR Paradigm)...16
EEG Data... 18 Acquisition... 18 Pre-processing... 18 Results... 19 Decoding... 19 ... 21
... 22
Questionnaires... 23
Rumination Reflection Questionnaire (RRQ)...23
Multidimensional Assessment of Interoceptive Awareness (MAIA)...24
Discussion... 25 Limitations... 27 Future Direction... 28 Conclusion... 29 References... 31 Appendix... 37
Abstract
Within the traditional Buddhist framework, the self is suggested to be a set of dynamic and transformable processes rather than a stable entity. This notion is the foundation of many Buddhist practices such as Vipassanā meditation. Thus, to what extent can the practice of Vipassanā influence the ability to access distinctive self-referential processing states volitionally? Previous research has identified two distinct modes of self-referencing known as the narrative and the experiential self. In the current study, we recruited experienced meditators to investigate meditation-related changes, objectively, through recording electroencephalographic (EEG) activity and, subjectively, through self-report questionnaires. Participants took part in an intensive 6-day meditation retreat (Vipassanā) and the impact of the retreat on the ability to volitionally access the narrative and experiential state were assessed. We used EEG as the research tool and further applied multivariate pattern analysis (MVPA), commonly referred to as decoding. MVPA has the ability to reveal dynamic differences in global mental representational states. By utilizing MVPA, the classifier was able to accurately discriminate between narrative and experiential states significantly above-chance level. However, contrary to our hypothesis, decoding accuracy did not increase post-retreat. Interestingly, decoding accuracy significantly increased post-retreat in the experiential and rest states, indicating the mind to engage in those distinct states more readily. Finally, we also discussed the possible limitations of this study and suggestions for future research, stressing the importance of a solid experimental design in meditation-related studies.
Covid-19 Constraints
This research project has been affected by Covid-19 and therefore I would like to briefly explain the circumstances which led to modifications from the original plan and structure of this project.
As the Covid-19 outbreak hit in the midst of conducting our research for this study, several aspects and goals have been restricted. Prior to the outbreak, EEG data from a total of 21 retreat participants and 6 control participants had been collected, however the pandemic constrained us from gaining access to the already collected control data. Hence, throughout this report I will only refer to the data which we were able to access (i.e. 21 retreat subjects), while keeping in mind that we did in fact collect data from 6 control subjects as well. Further, the aim was to obtain control data until a total of 21 subjects had been tested, however data collection was prohibited due to Covid-19. Therefore, the analysis will be limited to 21 retreat subjects, which leads to some limitations, which will be discussed.
Introduction
With the proposed benefits of meditation on physical and mental well-being, it is crucial to gain a more precise understanding of the underlying neurophysiological processes. Contemporary interventions such as the mindfulness-based stress reduction (MBSR) program founded by Kabatt-Zinn in 1990, aim at better coping mechanisms for reducing stress, regulating pain and processing anxiety by practicing meditation (Gibson, 2019). This intervention thereby has the goal to positively affect mental well-being by cultivating mindfulness and improving attention regulation (Kabatt-Zinn, 2011).
Based on traditional Buddhist philosophy, the belief in a stable self is at the root of mental suffering (Harvey, 1990). Moreover, deconstructing the self is believed to permit insights into one’s mind and self that then relieve suffering (Bodhi, 2011). Somewhat consistent with these traditional notions, mindfulness has been shown to affect neural activity related to self-referential processing (see Kilpatrick et al., 2011; Farb et al., 2007; Lutz, Brühl, Scheerer, Jäncke & Herwig, 2016) and thereby positively affect mental well-being (Hölzel, Lazar, Gard, Schuman-Olivier, Vago & Ott, 2011). A shift in processing of the self is proposed to occur by reducing rumination and encouraging a more detached present-moment awareness.
Neurocognitive research in meditation may provide novel insight into the flexibility of the brain, however, results must be taken with a grain of salt, as a lack of standardized methodology and terminology in this field has contributed to overgeneralization and exaggerated interpretations of findings in popular media (see Gunderson, 2016; Goyal et al., 2014; Van Dam et al., 2018; Tang, Hölzel & Posner, 2015).
In our study we aim to elucidate neural activity patterns that underlie the narrative and experiential self in experienced meditators prior to and following a 6-day Vipassanā meditation retreat. Therefore, we utilized a multivariate pattern analysis (MVPA), in which the classifier was trained to distinguish between the narrative and experiential self, and a state of rest. We
hypothesized there to be more distinctive neural patterns following the retreat, as a result of mindfulness affecting self-referential processing in a way that potentially enables the subjects to access those mental states more readily and at ease.
This introduction is structured into a brief overview on the topic of meditation and further elaboration on the specific practice of Vipassanā, followed by describing two distinct modes of the self (i.e. narrative and experiential) and the reasoning and goal behind this current study.
Meditation
For the past two decades, meditation has been growing in popularity and is being increasingly employed in the Western world; however, there is a broad range of definitions related to this umbrella term, which varies largely amongst research, practices and cultures (Van Dam et al., 2018). One of the most commonly used descriptions is to pay attention to one’s experience moment-by-moment in a nonjudgmental and open-hearted way (Kabatt-Zinn, 2011). The more contemporary practice of mindfulness meditation is described by Lutz et al., (2016) as a set of self-enquiring training regimes based on attention and emotion regulation.
Neuroscientific research on mindfulness has been focusing largely on the Buddhist traditional framework, from which mindfulness meditation stems. Thus far, various meditation techniques such as Vipassanā, Zen and Dzogchen have been examined in neuropsychological research (Dahl, Lutz & Davidson, 2015; Tang et al., 2015). Our study will aim to contribute to this field of research further and focus on the meditation style of Vipassanā, particularly in relation to altering neural activity patterns associated with self-referential processing.
According to Buddhist philosophy, mindfulness is comprised of two components; firstly, directing one’s attention to the current moment and secondly, being accepting towards that experience (Bishop et al., 2004). The traditional purpose of meditation within the Buddhist framework is to deconstruct the ego and thereby free oneself from psychological suffering (Dahl, et al., 2015). Although more contemporary meditation practices such as mindfulness training
focus largely on attention regulation (see Lutz, Slagter, Dunne & Davidson, 2008), the more traditionally rooted Buddhist aim is to achieve states beyond self-referential processing (Dunne, 2011).
Psychopathology such as rumination in depression is suggested to be associated with increased self-referential processing (Goldin, Ramel & Gross, 2009). Further, based on the Buddhist framework, decreasing self-related processing has the potential to ‘liberate’ the mind from such distress (Desbordes et al., 2014). There is a surge in literature supporting positive effects of meditation on mental and physical health (eg. Chambers, Gullone & Allen, 2009; Jha, Krompinger & Baime 2007). Currently several mindfulness interventions exist and are also utilized in a clinical setting. For example, MBSR and mindfulness-based cognitive therapy have the goal of cultivating awareness of the present moment and viewing mental events as transitory in a judgement-free manner and unrelated to the self (Herbert, Pauli & Herbert, 2011). As the underlying mechanisms are mostly unknown, the interest in researching self-referential processing in relation to mental health disorders is increasing (Nejad, Fossati & Lemogne, 2013).
Vipassanā
As briefly discussed, meditation is a broad term referring to a number of practices which vary not only in their execution but also in their effects on cognitive processing and altering attention (Lippelt, Hommelt & Colzato, 2014). Thus, it is crucial to precisely describe the meditation practice in the context of the proposed research, to avoid misinforming the public and overgeneralizing across distinct practices (Van Dam et al., 2018). Therefore, I will elaborate on the tradition of Vipassanā, as mentioned above, this was the branch of meditation taught in this study’s retreat.
Vipassanā, one of the oldest forms of Buddhist meditation practices, is often referred to as ‘insight meditation’ or ‘mindfulness meditation’ (Emavardhana & Tori, 1997). According to the Buddhist framework, Vipassanā encourages equanimous unconditional observation to reduce
identification with self-related processes, since the constant strive for a static self is believed to be the source of unhappiness (Harvey, 1990). Meditation involves two commonly studied meditative modes, which are described as focussed attention (FA), or Samatha, and open monitoring (OM), or Vipassanā (Lutz et al., 2008).
Lutz et al., (2008) elaborated on the operationalization of FA and OM. FA describes a meditative state which involves sustained attention and focus on a chosen object; for example, following the movement of the abdomen during breathing. When the mind wanders off, the aim of FA is to redirect the attention to the chosen object. Thus, attention regulation during FA involves staying attentive to any distractions, disengaging from such and finally redirecting one’s attention to the initial object of focus (i.e. for example breathing).
Open monitoring (OM) on the other hand has the aim of cultivating present-moment awareness and dereification (Lutz, Jha, Dunne & Saron, 2015). Thus, during OM, the attentional focus shifts from a specific chosen object towards a state of open-field awareness. The aim is to let go of any distraction and return to a broad scope of bare attention where anything arising in experience is noticed and released. Before engaging in Vipassanā meditation, one often starts with FA in the beginning to focus the mind, which may require mental effort, and once stabilized, moves towards OM. During OM, the mind is aware and present without engaging in anything that may arise (Lutz et al., 2008).
Self-Referential Processing
Gallagher (2000) characterized the sense of self as something that is not fixed, but as a set of changeable processes ranging from autobiographical processes to a more embodied sense of self. The term ‘self-referential processing’ can be defined as cognitive processes in which one relates information to oneself (Nejad et al., 2013).
Gallagher (2000) also distinguished between different states of the self, known as the experiential self (‘feel’) and the narrative self (‘think’). The experiential self is characterized by
mindfulness of the present moment and awareness of one’s bodily sensations. While the narrative self refers to cognitive engagement of the self and linking subjective mental events over time (Farb et al., 2007; Lutz et al., 2016; Vago & Zeidan, 2016). On a neurophysiological level, the narrative self is characterized by decreased high-gamma (60-80Hz) activity, while the experiential self is correlated with a decrease in oscillatory activity in the beta-band (13-25Hz). These neural patterns were primarily found in left frontal regions (Dor-Ziderman, Berkovich-Ohana, Glicksohn & Goldstein, 2013).
Rumination can be defined as a form of autobiographical self-generated thinking, often characterized by a state of worrying about the self in past, present and future scenarios and memories. Cortical midline structures such as the medial prefrontal cortex (mPFC) have been shown to be activated during generated rumination through retrieval of negative memories (Kross, Davidson, Weber & Ochsner, 2008). Rumination is often symptomatic for depression due to this distressing form of self-referencing (Nejad et al., 2013). Mind-wandering is suggested to be the cause rather than the consequence of unhappiness (Killingsworth & Gilbert, 2010). Nevertheless, it should be noted that ruminating can also have positive outcomes, particularly in regard to insights and problem-solving which might stem from this so-called ‘overthinking’ (Nejad et al., 2013). Research suggests that by practicing mindfulness, a shift in self-processing states occurs, moving from a state of rumination and self-narration towards a more detached, experiential state of present-moment centred self-awareness (Farb et al., 2007; Lutz et al., 2016). Hence, down-regulation of self-referential processing may be the pillar of how mindfulness is beneficial in coping with psychological disorders and improving overall well-being (Hölzel et al., 2011).
Based on research by Farb and colleagues (2007), discussed further below, we propose that with increasing expertise in mindfulness meditation, more flexibility and freedom in accessing the narrative and experiential self appears. Meditation practice ought to allow one to stay in an experiential state without interruption from thoughts independent of the present moment, but equally possess the ability to flexibly and quickly return to thinking when it is
necessary. Mindfulness meditation is therefore proposed to have the ability to alter neurocognitive processing of the self specifically by making one more able to voluntarily switch between different self-processing states.
Previous Research
There is indeed previous research with evidence that meditators can more easily access different mental states. In 2007, Farb et al., conducted a fMRI study in which neural activity was measured while undergoing a task in which subjects were instructed to engage in the narrative or the experiential self. Half of the participants followed an 8-week intensive mindfulness retreat and subsequently showed marked reductions in the mPFC, a brain region associated with processing of the narrative self, relative to the control group. This is indicative of a more distinct neural pattern, which, however, can only be evidenced by MVPA, compared to the univariate analysis employed by Farb et al, (2007). These findings suggest that practicing meditation leads to a greater disengagement of the narrative focus, when comparing to the untrained novice group. Control subjects showed a more intertwined (correlated) pattern of neural activity.
Utilizing a similar paradigm, Lutz et al., (2016) also conducted a fMRI study in which half of the participants were given mindfulness training and then tested on a paradigm in which, again, subjects were asked to engage in the narrative or experiential focus. During the experiential focus, regions associated with the default mode network (such as the mPFC), which is active during mind-wandering (Gusnard, Akbudak, Shulman & Raichle, 2001), has been found to be more downregulated in the mindfulness group, relative to the control group. This again suggests that meditators have an enhanced ability to volitionally access the experiential and narrative self.
Aim of this Study
In the current study we aim to examine meditation-related neurophysiological changes particularly in relation to processing of the self. Through electroencephalographic (EEG)
recordings, we will investigate neural activity patterns when switching between the narrative and experiential states of self-referencing prior to and following an intensive 6-day Vipassanā meditation retreat. As previous research has shown (Farb et al., 2007; Lutz et al., 2016) that meditation has the potential to alter neural activity patterns, our current study aims to extend on those findings. During the Vipassanā retreat subjects will employ meditation in the form of focused attention (FA) and open monitoring (OM). Consistent with the findings by Farb et al., (2007) and Lutz et al., (2016), we hypothesize that the intensive mindfulness practice during the retreat will lead to a clearer distinction in brain activity patterns between the narrative and experiential self, post-retreat.
The two above mentioned studies (Farb et al., 2007; Lutz et al., 2016) utilized fMRI, therefore we decided on a different approach to investigate our research question, by recording EEG activity. EEG will then be analyzed using multivariate pattern analysis (MVPA), also known as decoding. Decoding has been widely applied in fMRI studies, however for EEG data this technique is not yet common practice, with the exception of brain computer interface research (Müller, Tangermann, Dornhege, Krauledat, Curio & Blankertz, 2008; Grootswagers, Wardle & Carlson, 2017). One of the main reasons for choosing MVPA is that the linear classifier, a backwards decoding model, can detect relevant information that would potentially get lost when applying the standard univariate analyses. Hence decoding has increased sensitivity to distinguish between different conditions (Cauchoix, Barragan-Jason, Serre & Barbeau, 2014).
Thus far, previous work studying the effects of meditation on the narrative and experiential self has only applied univariate analysis (Farb et al., 2007; Lutz et al., 2016). Univariate analysis has the ability to show that neural activity is higher or lower in specific brain regions, however, only decoding can reveal whether the global respresentational states indeed become more distinct. Our study therefore intends to extend on the previous research by applying MVPA, as it can confirm whether the changes in representational states become more different on a continuum.
We predict the decoding model to accurately classify the two self-referencing states (i.e. narrative and experiential) above chance. Further, following the intensive meditation retreat we hypothesize a clearer distinction in neural patterns, which would be indexed by higher decoding accuracy of the classifier. Increased decoding accuracy would reflect a larger degree of freedom and flexibility of the mind in accessing distinct states of the self.
Additionally, we will assess the effect of meditation on a subjective level by means of two self-report questionnaires: The Rumination Reflection Questionnaire (RRQ; Trapnell & Campbell, 1994) and the Multidimensional Asset of Interoceptive Awareness (MAIA; Mehling et al., 2012). Since practicing meditation is suggested to decrease rumination (Jain et al., 2007), we hypothesize lower scores on the trait rumination in the RRQ following the retreat. During mindfulness, the focus is often directed towards inner mind and bodily sensations (Gibson, 2019). Therefore, we expect higher scores on the MAIA post-retreat, which would reflect higher interoceptive awareness.
Materials and Methods
Participants
The current study includes data from a total of 21 subjects. Participants consisted of experienced meditators aged between 26-57 years (Mage=38, SD=8.3; 8 males) and were recruited through a self-selection process directly at the Vipassanā meditation retreat. Anyone enrolled in the retreat was able to sign up for the experiment on a voluntary basis. Participants informed us about their lifetime number of hours practicing meditation through a questionnaire. Practice hours ranged from 10-3362 (Mhours=757, SD=1131.7). Additionally, we assessed further details regarding the participants’ meditation habits such as frequency and method of their meditation practice through an online questionnaire (see:
https://docs.google.com/forms/d/e/1FAIpQLSfI2WLZox2Prp9xG1gDW7IqWi3a2z0EgmRBNUJ J-efX5Xgznw/viewform).
All participants were native English speakers or fluent in the English language. Subjects signed a written informed consent prior to starting the experiment and were made aware that they could withdraw from the study at any given time without any consequences. All participants had normal or corrected to normal vision. Further exclusion criteria was severe psychopathology.
Pre-retreat data from one subject was removed due to missing triggers in the EEG collection procedure.
The study was approved by the ethics committee of the Psychology Department at the La Sapienza University of Rome, Italy.
Design and Procedure Meditation Retreat
All subjects participated in the intervention of a 6-day Vipassanā meditation retreat. The background and method of Vipassanā meditation has been described above in the section
‘Vipassanā’, therefore I will just briefly describe the circumstances and procedures surrounding the retreat.
The retreat took place in an active monastery in Fara Sabina, Italy and was led by Prof. Dr. H.P. Barendregt, Prof. A. Raffone and M. Hartkamp. Participants practiced 10 hours of meditation per day. The subjects received instructions to focus on abdomen movement in the sitting meditation, while becoming openly aware of anything arising and focusing on sensations in the feet during the walking meditation. Meditation sessions lasted 45mins and the whole duration of the retreat took place in silence, aside from the interview sessions. During the entire time of the retreat, sitting and walking meditation alternated, only interrupted by 3 given mealtimes per day, the daily interview, and question and answer sessions.
Subjects were tested on the same tasks just before commencing the retreat and a second time just following the 6-day long retreat.
Questionnaires
Participants were required to complete self-report questionnaires at two time points; once before the meditation retreat started (T0) and a second time following the completion of the 6-day Vipassanā retreat (T1). To assess changes in rumination and self-reflection at pre- and post-retreat intervention testing times we included the Rumination Reflection Questionnaire (RRQ; Trapnell & Campbell, 1999). The RRQ consists of 24 statements (12 on rumination, 12 on reflection) which are rated based on the level of agreement.
As interoceptive awareness is suggested to be a key factor in mindfulness meditation (Gibson et al., 2019), we included the Multidimensional Assessment of Interoceptive Awareness (MAIA; see https://www.osher.ucsf.edu/maia/, Mehling et al., 2012). The MAIA is a 32-item state-trait questionnaire consisting of 8 scales (Noticing, Not-Distracting, Not-Worrying, Attention Regulation, Emotional Awareness, Self-Regulation, Body Listening and Trusting). By cultivating awareness of interoception, a shift from thinking about bodily sensations towards
simply feeling these present-moment bodily sensations may occur (Farb, Segal & Anderson, 2013).
Filling out the questionnaires took around 15mins per subject.
In order to carry out analyses on the self-report questionnaires, we performed two-tailed t-tests on the RRQ data and the MAIA data. Thereby we compared the groupwise pre- and post-retreat test scores on the individual questionnaires to determine whether there were any significant differences (p<0.05).
EEG Experiment (FTR Paradigm)
The Feel-Think-Rest (FTR) paradigm is based on a paradigm on distinctive modes of self-referential processing by Herwig, Kaffenberger, Jäncke & Brühl, 2010 (see Lutz et al., 2016). Subjects were asked to switch between the narrative (think) and experiential (feel) self and a state of rest. Simple visual stimuli of letters reflected the individual conditions (‘F’, ’T’, ’R’). Instructions on how to engage in the different conditions were given repeatedly, before and after a practice trial. Throughout the task, while engaging in any of the conditions, subjects were asked to try and remain with their eyes closed. During the think (‘T’) condition, participants were instructed to reflect and think about themselves subjectively. During the feel (‘F) condition, participants were asked to feel into themselves and their bodily sensations. During the rest (‘R’) condition, participants were asked to simply not think of or do anything specific. The task started with a 3s presentation of a fixation cross, followed by a 12s presentation of a randomly selected condition (think, feel, or rest). A short tone indicated a change in condition. This was repeated 36 times, so that every condition appeared 12 times.
Total EEG measuring time for this task was around 12mins. Stimuli were presented in a pseudo-randomized order to ensure there was no direct repetition of conditions. See Figure 1 for a schematic representation of the FTR paradigm. The exact instructions for the task can be found in the ‘Appendix A’.
Participants completed the FTR task twice, in two separate sessions. Once at testing point T0, prior to starting the Vipassanā retreat, and a second time just after the retreat at testing point T1. Both measurements were taken directly at the location of the retreat itself (at the monastery in Fara Sabina, Italy). T0 and T1 were separated by 6 consecutive days.
All scripts for the FTR task were written in PsychoPy (Version 1.85.4, www.psychopy.org) and the visual stimuli presented on an ASUS 17-inch monitor at 60Hz.
Fig. 1 Schematic overview of a trial sequence of the FTR Paradigm procedure. 12s condition
presented in a pseudo-randomized order, followed by a 3s presentation of a fixation cross. Numbers indicate triggers corresponding to the events, sent to the Cognionics EEG system.
EEG Data Acquisition
EEG was recorded at 512Hz using a Cognionics (www.cgxsystems.com) 64-channel wet EEG cap. Electrodes were placed according to the easycap-M11 system (www.easycap.de).
Pre-processing
All acquired data was pre-processed using custom written scripts in the Python environment (www.python.org; see https://github.com/dvanmoorselaar/DvM) based primarily on the MNE toolbox (www.mne.tools). Data was re-referenced to the average of all electrodes. Prior to epoching, EEG data was high-pass filtered at 0.1Hz using a Firwin filter (Gramfort et al., 2013). Individual trials of 12000ms were each segmented into epochs at -200ms to 1000ms, accounting for a total of 12 epochs for each 12 second trial. This was done to increase the trial number for improved classification since the time-dimension was not relevant for our hypotheses. To control for filter artifacts, the epochs were extended to 500ms on both sides. EEG signal was bandpass filtered at 110-140Hz and an adapted version of the automatic trial rejection procedure allowed for artifact removal, by calculating the variance in within-subject z-score cut-off (de Vries, van Driel & Olivers, 2017). Following trial rejection, independent component analysis (ICA) was performed to exclude eye movements. No baseline correction was applied, which will be discussed further in the ‘Limitations’ section below. 6 electrode channels (‘ACC0’, ‘ACC1’, ‘ACC2’, ‘TP9’, ‘FT9’, ‘FT10’) were removed from the analysis, as they were not included in the standard montage (‘biosemi64’) required for our MNE analysis pipeline. These electrodes are unlikely to affect our decoding analysis, as they are located in highly lateralized regions.
Results
Decoding
To perform the decoding analysis, more formally known as multivariate pattern analysis (MVPA), we trained a linear classifier, specifically a backwards decoding model on the EEG data. EEG data was downsampled at 128Hz. For all decoding analysis we followed a 10-fold cross-validation, in which 90% of the data was split into a training-set and 10% into a testing-set. This procedure was repeated until all trials were included in the testing-set once. Data was trained and tested on independently per subject. All analysis was performed on the split single trial data, using broadband EEG.
We conducted a 3-way classification, collapsing across all conditions (Feel/Think/Rest) and furthermore performed 3 binary classifications (Feel/Think; Think/Rest; Rest/Feel). Our main interest lies in the classification of Feel/Think states, as we hypothesize there to be more distinct neural patterns and thus higher decoding accuracy of those states following the intensive mindfulness retreat.
Non-parametric cluster-based permutation testing was carried out to correct for multiple comparisons, and to assess significance of decoding performance determined with an alpha below 0.05. This analysis was chosen as it controls for family-wise error (Maris & Oostenveld, 2007). Decoding plots were visualized using custom-written scripts in Jupyter Notebook (www.jupyter.org).
In the pre-retreat session, we expected lower decoding accuracy in all comparisons (binary and collapsed across all conditions) relative to the post-retreat session. Decoding of the binary feel/think comparison is predicted to show the highest classification accuracy of all binary comparisons post-retreat. The think/rest classification is hypothesized to change following the retreat, as prior the Vipassanā retreat, subjects are more prone to be in the narrative state of self
while at rest, whereas following the retreat subject are expected to continue in a more meditative, experiential state of self.
In Figure 2 we show pre- and post-retreat differences in decoding accuracy, collapsed across all conditions (3-way comparison) and 3 binary comparisons (feel/think; feel/rest; think/rest). Above-chance decoding was found in all comparisons, both at pre- and post-retreat testing points; significant decoding was found in all but one comparison (i.e. think versus rest pre-retreat).
The uppermost panel (a and e) demonstrates the classifier accuracy across all conditions. A longer time-period of significant clusters can be observed at baseline (pre-retreat), however after conducting a two tailed t-test, it was revealed that there are no significant differences (t(3080)=-0.02; p=0.97) in decoding accuracy at pre- and post-retreat testing times (see Table 1). The classifier was able to significantly decode above-chance (i.e. above 33% for all conditions). In the second panel (b and f) the binary feel/think comparisons are presented. We hypothesized there to be an increase in significant decoding accuracy following the retreat. As shown in Table 1, contrary to our expectations, there is a significant decrease (t(3080)=2.87; p=0.004); hence the classifier was able to discriminate more accuratley at the pre-retreat testing point. This indicates that subjects did not improve in accessing the narrative (think) and experiential (feel) self. In the third panel (c and g) the binary feel/rest comparisons are shown. We expected there to be little effect, however pre- and post-retreat show significant differences in decoding accuracy, and both plots indicate significant decoding clusters. There is a highly significant increase in decoding accuracy post-retreat (see Table 1, t(3080)=-7.88; p=4,61E-15). The final panel (d and h) presents the binary think/rest comparisons. As expected, there is more variability in the think/rest state pre-and post-retreat. A significant decrease in decoding accuracy at post-retreat testing point (see Table 1, t(3080)=13.76; p=7,09E-42), relative to pre-retreat time is shown. From these findings it appears that the resting mind is in fact most affected by the meditation retreat, as the binary
comparisons with a rest condition revealed the most significant differences at pre- versus post-retreat testing.
Time (s) Time (s) Time (s) Time (s) (e) (a)
POST
PRE
(f) (h) (d) (g) (c) Time (s) Time (s) (b) Time (s) Time (s)Classification Pre-Retreat mean Post-Retreat mean p-value
All Conditions 0.54 0.54 0.97
Feel vs Think 0.56 0.55 0.004*
Rest vs Feel 0.54 0.56 4,61E-15**
Think vs Rest 0.57 0.54 7,09E-42**
Fig. 2 Decoding accuracy across 1s trials at pre- and post-retreat testing points.
Graphs on the left denote pre-retreat results, while graphs on the right denote post-retreat results. The dotted horizontal line presents chance level (i.e. 0.33 in 3-way classification and 0.5 in binary comparisons); all plots show decoding accuracy above-chance. Periods of significance (sig.; p<0.05) after cluster-based permutation testing are indicated with a blue line. X-axis represents time in s, with 0s being the trial onset. Y-axis represents decoding accuracy as area under the curve (AUC).
(a) Decoding accuracy collapsed across all conditions with significant clusters from 0-0.77s. (b) Decoding accuracy in binary comparison of feel/think states with sig. clusters from 0-0.26s. (c) Decoding accuracy of feel/rest states with sig. clusters from 0-0.1s. (d) Decoding accuracy of think/rest states without any sig. decoding. (e) Decoding accuracy collapsed across all conditions with sig. clusters from 0-0.5s. (f) Decoding accuracy of feel/think states with sig. clusters from 0-0.25s. (g) Decoding accuracy of feel/rest states with sig. clusters from 0-0.5s. (h) Decoding accuracy of think/rest states with sig. clusters from 0-0.27s.
*Note. In some of the plots significant clusteres were found prior to onset (i.e. before 0.00s), however, since no baseline correction was possible (due to splitting trials into 1s epochs) these are not illustrated.
Note. *Significant mean changes at the 0.05 level (two-tailed) **Significant mean changes at the 0.001 level (two-tailed)
The observed highly significant p-values are likely the result of a large dataset (n=3080), as the pre- and Table 1
Questionnaires
Differences between the pre- and post-retreat testing points of the experienced meditators were examined using a two-tailed t-test. The same procedure was applied for the RRQ and for the MAIA. Results of 20 participants were included in the analysis of both questionnaires.
Rumination Reflection Questionnaire (RRQ)
In line with our hypotheses, a two-tailed t-test revealed significant changes across time on the score for trait ‘rumination’ (t(19)=2.86, p=0.01), however no significant changes were observed for the trait ‘reflection’ (t(19)=-1.88, p=0.08). Scores on rumination decreased post-retreat, compared to baseline, while scores on reflection increased over time. These results indicate that meditation decreased ruminative thoughts, thus perhaps decreasing engagement in the narrative self-focus. Figure 3 shows a significant decrease of 0.46 points on the Rumination scale (Mpre=3.05, Mpost=2.59) and a non-significant increase of 0.15 points on the Reflection scale (Mpre=4.03, Mpost=4.18).
Rumination Reflection 3.05 4.03 2.59 4.18 Pre R R Q S co re s
Fig. 3 Scores on Rumination and Reflection in the RRQ at baseline and following the meditation retreat
(Vipassanā). X-axis represents the two traits (rumination and reflection) and y-axis is the scoring on the traits. Dark blue indicates pre-retreat scores and light blue indicates post-retreat scores. *Changes are significant as revealed by the two-tailed t-test (p<0.05).
Multidimensional Assessment of Interoceptive Awareness (MAIA)
Table 2 shows the changes in mean scores at baseline and post-retreat on the 8 dimensions of the MAIA scale. Significant changes across time were found in 6 (Noticing, Attention Regulation, Emotional Awareness, Self-Regulation, Body Listening, Trusting) dimensions. The most significant changes were observed on the traits Noticing with an increased score of 0.68 points (t(19)=-4.18, p=0.0005) and Body Listening with an increase of 0.67 points (t(19)=-4.25, p=0.0004). In the MAIA, Noticing is characterized by awareness of bodily sensation, whether they are comfortable, neutral or uncomfortable. Body Listening is described as actively listening to the body for insight (Mehling et al., 2012).
These results confirm our hypotheses on the effect of meditation on increasing interoceptive awareness, as indicated by increased scores on the MAIA.
Table 2
MAIA scores pre- and post-retreat
Trait Pre-Retreat mean Post-Retreat mean p-value
Noticing 3.5 4.18 0.0005** Not Distracting 2.18 2.57 0.15 Not Worrying 2.95 3 0.77 Attention Regulation 3.47 3.9 0.01* Emotional Awareness 3.66 3.96 0.03* Self-Regulation 3.45 3.79 0.02* Body Listening 3 3.67 0.0004** Trusting 3.52 3.93 0.02*
Note. *Significant mean changes at the level 0.05 (two-tailed t-test) **Significant mean changes at the level 0.001 (two-tailed t-test)
Discussion
Growing evidence is supportive of meditation being able to induce functional and structural neuroplastic changes. Our study had the aim of contributing further to this claim by
investigating the potential impact of a 6-day long Vipassanā retreat on neural representations when accessing the narrative (think) or experiential (feel) mode of self-reference. Unfortunately, our results were inconsistent with previous neuroimaging findings (Farb et al., 2007; Lutz et al., 2016), as following the intensive mindfulness meditation, we only observed more significantly distinct neural patterns in the think/rest comparison. All other classifications resulted in decreased decoding accuracy (feel/think, rest/feel) or non-significant differences between pre- and post-retreat decoding (collapsed across all conditions).
Nevertheless, applying a decoding-based technique holds promising for future empirical research in this field. Particularly in the task of discriminating between mental states, decoding has the ability to detect spatial patterns that could have been neglected in the common univariate analysis applied on EEG data. In the current study, decoding allowed us to gain insight into changes in representational states across a continuum and across spatial regions, rather than only being able to report higher or lower neural activation in specific brain regions.
Additionally, we obtained subjective measurements through two self-report tools (RRQ and MAIA), giving us insight into whether neuroimaging findings relate with the subjective experience. Below I will briefly summarize the findings, followed by limitations and suggestions for improvement, subsequently leading to a conclusion.
Summary of Results Questionnaires
Our self-report results were supportive of our hypothesis; ruminative scores decreased significantly post-retreat, while reflective scores increased. The decrease in rumination would therefore indicate a decreased engagement in the narrative self. Increased self-reflection (Nb. Non-significant increase) potentially suggests increased meta-awareness and therefore being able to access the different self-states more easily.
Scores on the MAIA are also in line with our hypothesis, revealing higher interoceptive awareness post-retreat.
Decoding
The classifier was able to discriminate between conditions at above-chance level for all analyses. Periods of significant decoding were demonstrated in all comparisons, with one exception being the baseline think/rest comparison. Decoding accuracy showed the least variability and most significant clusters in the pre-retreat 3-way classification. This can be explained by the classifier receiving the most information (i.e. from all three conditions). However, it is suspicious that the decoder was able to classify way above chance (>0.33) but does not reveal significant clusters throughout the entire time period. This is suggestive of an error in the analysis pipeline.
Unfortunately, our main hypothesis that think/feel states would show significantly increased decoding accuracy following the retreat as in previous research (Farb et al., 2007; Lutz et al., 2016), was not confirmed. In fact, a significant decrease (as reported in ‘Results’) following the Vipassanā retreat was found. However, perhaps subjects were able to volitionally access these states at ease prior to the retreat due to their general meditation experience. In future studies when including an active control group of novices, more valid conclusions can be drawn.
Decoding accuracy for the feel versus rest classification increased following the retreat. This is in fact contrary to our expectations, as we hypothesized the resting mind after meditating to be more experiential and would therefore decrease discrimination between the experiential and rest condition. Yet, these findings could also be explained by meditators being more able to access these different states, and their mind returning to a more default narrative resting mind, which they had pre-retreat during the task when required to be in the rest condition.
Regarding the think versus rest states, we expected no significant decoding pre-retreat, as we hypothesized subjects to be more engaged in the narrative self during rest, and thus would not
show significant differences compared to the think (narrative) state. This is also described by Farb et al., (2007) as a ‘default bias towards a narrative focus’. However, after conducting a two-tailed t-test, we conflictingly found a significant decrease in decoding accuracy post-retreat, indicating a more intertwined (correlated) thinking and resting mind.
Thus, the last two findings are contradictory, as we would have expected the resting mind to become more meditative following the retreat. Our findings are therefore contrary to our hypotheses, as we would have expected an increased decoding accuracy of think/rest comparisons and a decreased decoding accuracy of feel/rest classifications post-retreat.
It is highly important to note, that the differences, even though appearing highly significant due to the reported p-values, are in fact minor between pre- and post-retreat, as can be inferred from their means in Table 1. Decoding accuracies from pre- to post-retreat differed by 1-2%. The above results should therefore not be overinterpreted (see Lin, Lucas & Shmueli, 2013).
Limitations
Several limitations regarding this study will now be discussed. First and foremost, due to Covid-19, we were not able to collect further data to include a control group. Hence, the improved ability of meditators to engage in the two distinct modes of self-relating processes could simply be the result of practice, as they underwent the FTR paradigm twice with only a 6-day intermission. Thus, the confounding factor of a training effect cannot be excluded. Testing controls who do not take part in the meditation retreat, but are nevertheless tested on these two time points, would allow us to draw more accurate conclusions on whether the meditation retreat indeed is the reason for improved ability to access mental states and therefore for increased decoding accuracy.
Furthermore, we encountered various obstacles in the process of the decoding analysis. Our raw data length was 12s long per trial, however most decoding studies use epochs of 1-2s long. Thus, for more efficient decoding, we decided to split the 12s trials into 12 segments of 1s,
to increase trial number and thereby permit 10-fold cross validation. This however led to a failure to implement baseline correction.
Future Direction
Taking the results and limitations of this study into account, I will now propose several ways in which this experiment could be improved in future research.
Firstly, EEG decoding is applied to time-series data and commonly used for the analysis of event-related potentials (ERPs) (Grootswagers et al., 2017). However, the timing of decoding accuracy was expected to be highly variable in our experiment and irrelevant for our research question. Therefore, a procedure that could be utilized in future studies is decoding with common spatial pattern (CSP) filters. CSP filters decompose EEG data into spatial patterns and improve signal-to-noise ratio. Thus far, CSP filters have been a widely used approach in brain computer interfaces to discriminate mental states. Instead of simply decoding along the time-course, spatial filters allow for improved spatial resolution and thereby higher accuracy in classification of mental states (Blankertz, Tomioka, Lemm, Kawanabe & Müller, 2008).
Secondly, as touched upon in the ‘Limitation’ section, this study is absent of a control group. Conducting a cross-validation study would give us more insight into the within- and between-subject differences. By adding a control group of novices and experienced meditators, and an intervention group also consisting of novices and experienced meditators, this would eradicate some of the confounding variables. An active intervention for the control group would be suitable, as otherwise the effect of factors unrelated to meditation experienced during the retreat, such as the diet or the daily interviews led by the meditation teachers, cannot be excluded. For control subjects this could perhaps include other relaxation exercises and education on stress management and ensuring that a similar diet is followed (see Tang et al., 2015).
Thirdly, there are various aspects which could increase classifier performance and therefore decoding accuracy. As early as in the pre-processing stage, there are various decisions to
be made which can affect the results. Averaging across trials increases signal-to-noise ratio, with an average of up to 7 trials being optimal (Chan et al., 2011). Moreover, a different cross-validation design (Nb. our study utilized 10-fold cross-cross-validation) could be chosen, such as the leave-one-out cross-validation which, according to Grootswagers et al., (2017) is the best. Furthermore, in our case, baseline correction could be included by implementing it prior to epoch-splitting. All these discussed steps could potentially improve decoding accuracy. Another point to implement in a future decoding pipeline, is splitting the epochs into 0.1s segments rather than 1s trials, as this would allow for more visually comparable results and furthermore increase the power.
Lastly, regarding the self-report questionnaires, the significant changes observed on both the RRQ and the MAIA could be unconsciously biased by the participants, as they might be more likely to conform to meditation-related expected changes when answering the questionnaire after the retreat. Therefore, in future research, it would be valuable to perform a correlation analysis between scores on the RRQ/MAIA and decoding accuracy in individual subjects, to understand whether e.g. increased interoceptive awareness scores on the MAIA indeed leads to increased decoding accuracy in that individual subject. Additionally, a correlation analysis could be conducted with less and more experienced meditators, as findings could differ depending on the subjects’ level of expertise; perhaps an optimum level of meditation experience could be revealed, at which subjects show the most flexibility in accessing the narrative and experiential self.
Conclusion
By employing a backwards decoding model, we explored the potential effect of a meditation retreat on neural activity patterns specifically in relation to self-referential processing.
Contrary to our expectations, we did not find more distinct neural patterns in the think versus feel states, however, we acquired insight into how the resting mind may be influenced by meditation and thereby affect self-referencing. The pre- and post-retreat differences in the feel
versus rest states, are suggestive of meditation allowing the mind to switch more easily between a default resting thinking mind and the more meditative experiential, and thus could indirectly imply an improved ability to switch between the feel versus think state, as hypothesized.
Overall, the capacity to engage in the experiential focus at will, as is indexed by a higher decoding accuracy between feel versus rest states post-retreat, has implications for several mental health disorders such as depression, due to increasing present-moment awareness. This could result in being able to decrease worrying thoughts related to past and future events. Thus, meditation has the potential to improve well-being and cultivate happiness through increased interoceptive awareness (Davidson et al., 2004), by being able to volitionally engage in the experiential self more readily. Finally, although some of our findings are promising, we did not compare them to an an active control group, and therefore, cannot conclude whether the changes in neural activity patterns are in fact a result of the meditation retreat itself.
References
Bishop, S., Lau, M., Shapiro, S., Carlson, L., Anderson, N., & Carmody, J. et al. (2004). Mindfulness: A Proposed Operational Definition. Clinical Psychology: Science And Practice, 11(3), 230-241. doi: 10.1093/clipsy.bph077
Blankertz, B., Tomioka, R., Lemm, S., Kawanabe, M., & Muller, K. (2008). Optimizing Spatial filters for Robust EEG Single-Trial Analysis. IEEE Signal Processing Magazine, 25(1), 41-56. doi: 10.1109/msp.2008.4408441
Bodhi, B. (2011). What does mindfulness really mean? A canonical perspective. Contemporary Buddhism, 12(1), 19-39. doi: 10.1080/14639947.2011.564813
Cauchoix, M., Barragan-Jason, G., Serre, T., & Barbeau, E. J. (2014). The neural dynamics of face detection in the wild revealed by MVPA. Journal of Neuroscience, 34, 846–854.
Chambers, R., Gullone, E., & Allen, N. B. (2009). Mindful emotion regulation: An integrative
review. Clinical psychology review, 29(6), 560–572.
https://doi.org/10.1016/j.cpr.2009.06.005
Dahl, C. J., Lutz, A., & Davidson, R. J. (2015). Reconstructing and deconstructing the self: cognitive mechanisms in meditation practice. Trends in cognitive sciences, 19(9), 515– 523. https://doi.org/10.1016/j.tics.2015.07.001
Davidson, R. J., & Lutz, A. (2008). Buddha's Brain: Neuroplasticity and Meditation. IEEE signal processing magazine, 25(1), 176–174. https://doi.org/10.1109/msp.2008.4431873
Davidson, R. (2004). Well–being and affective style: neural substrates and biobehavioural correlates. Philosophical Transactions Of The Royal Society Of London. Series B: Biological Sciences, 359(1449), 1395-1411. doi: 10.1098/rstb.2004.1510
Desbordes, G., Gard, T., Hoge, E. A., Hölzel, B. K., Kerr, C., Lazar, S. W., Olendzki, A., & Vago, D. R. (2014). Moving beyond Mindfulness: Defining Equanimity as an Outcome
Measure in Meditation and Contemplative Research. Mindfulness, 2014(January), 356– 372. https://doi.org/10.1007/s12671-013-0269-8
Dor-Ziderman, Y., Berkovich-Ohana, A., Glicksohn, J., & Goldstein, A. (2013). Mindfulness-induced selflessness: a MEG neurophenomenological study. Frontiers In Human Neuroscience, 7. doi: 10.3389/fnhum.2013.00582
Dunne, J. (2011). Toward an understanding of non-dual mindfulness. Contemporary Buddhism, 12(1), 71-88. doi: 10.1080/14639947.2011.564820
Emavardhana, T., & Tori, C. (1997). Changes in Self-Concept, Ego Defense Mechanisms, and Religiosity Following Seven-Day Vipassanā Meditation Retreats. Journal For The Scientific Study Of Religion, 36(2), 194. doi: 10.2307/1387552
Farb, N. A., Segal, Z. V., Mayberg, H., Bean, J., McKeon, D., Fatima, Z., & Anderson, A. K. (2007). Attending to the present: mindfulness meditation reveals distinct neural modes of self-reference. Social cognitive and affective neuroscience, 2(4), 313–322. https://doi.org/10.1093/scan/nsm030
Farb, N. A., Segal, Z. V., & Anderson, A. K. (2013). Mindfulness meditation training alters cortical representations of interoceptive attention. Social cognitive and affective neuroscience, 8(1), 15–26. https://doi.org/10.1093/scan/nss066
Gallagher, S. (2000). Philosophical conceptions of the self: implications for cognitive science. Trends In Cognitive Sciences, 4(1), 14-21. doi: 10.1016/s1364-6613(99)01417-5
Gibson, J. (2019). Mindfulness, Interoception, and the Body: A Contemporary Perspective. Frontiers In Psychology, 10. doi: 10.3389/fpsyg.2019.02012
Goldin, P., Ramel, W., & Gross, J. (2009). Mindfulness Meditation Training and Self-Referential Processing in Social Anxiety Disorder: Behavioral and Neural Effects. Journal of cognitive psychotherapy, 23(3), 242–257. https://doi.org/10.1891/0889-8391.23.3.242
Goyal, M., Singh, S., Sibinga, E. M., Gould, N. F., Rowland-Seymour, A., Sharma, R., Berger, Z., Sleicher, D., Maron, D. D., Shihab, H. M., Ranasinghe, P. D., Linn, S., Saha, S., Bass, E. B., & Haythornthwaite, J. A. (2014). Meditation programs for psychological stress and well-being: a systematic review and meta-analysis. JAMA internal medicine, 174(3), 357–368. https://doi.org/10.1001/jamainternmed.2013.13018
Grootswagers, T., Wardle, S., & Carlson, T. (2017). Decoding Dynamic Brain Patterns from Evoked Responses: A Tutorial on Multivariate Pattern Analysis Applied to Time Series Neuroimaging Data. Journal Of Cognitive Neuroscience, 29(4), 677-697. doi: 10.1162/jocn_a_01068
Gunderson G. The science is in, and meditation may be the next big business
opportunity. Forbes. 2016 Jun 28; Retrieved
from http://www.forbes.com/sites/garrettgunderson/2016/06/28/the-science-is-in-and-meditation-may-be-the-next-big-business-opportunity/#2f729c7123c0.
Gusnard, D., Akbudak, E., Shulman, G., & Raichle, M. (2001). Medial prefrontal cortex and self-referential mental activity: Relation to a default mode of brain function. Proceedings Of The National Academy Of Sciences, 98(7), 4259-4264. doi: 10.1073/pnas.071043098 Harvey, P. (1990). An introduction to Buddhism: Teaching, history and practices. Cambridge:
Cambridge University Press.
Herbert, C., Pauli, P., & Herbert, B. M. (2011). Self-reference modulates the processing of emotional stimuli in the absence of explicit self-referential appraisal instructions. Social cognitive and affective neuroscience, 6(5), 653–661. https://doi.org/10.1093/scan/nsq082 Herwig, U., Kaffenberger, T., Jäncke, L., & Brühl, A. (2010). Self-related awareness and emotion
regulation. Neuroimage, 50(2), 734-741. doi: 10.1016/j.neuroimage.2009.12.089
Hölzel, B. K., Lazar, S. W., Gard, T., Schuman-Olivier, Z., Vago, D. R., & Ott, U. (2011). How does mindfulness meditation work? Proposing mechanisms of action from a conceptual and neural perspective. Perspectives on psychological science, 6(6), 537-559.
Jain, S., Shapiro, S. L., Swanick, S., Roesch, S. C., Mills, P. J., Bell, I., & Schwartz, G. E. (2007). A randomized controlled trial of mindfulness meditation versus relaxation training: effects on distress, positive states of mind, rumination, and distraction. Annals of behavioral medicine : a publication of the Society of Behavioral Medicine, 33(1), 11–21. https://doi.org/10.1207/s15324796abm3301_2
Jha, A. P., Krompinger, J., & Baime, M. J. (2007). Mindfulness training modifies subsystems of attention. Cognitive, affective & behavioral neuroscience, 7(2), 109–119. https://doi.org/10.3758/cabn.7.2.109
Kabat-Zinn J. (1990). Full catastrophe living: Using the wisdom of your body and mind to face stress, pain, and illness. Dell Publishing; New York.
Kabat-Zinn, J. (2011). Some reflections on the origins of MBSR, skillful means, and the trouble
with maps, Contemporary Buddhism, 12:1,
281-306, DOI: 10.1080/14639947.2011.564844
Killingsworth, M., & Gilbert, D. (2010). A Wandering Mind Is an Unhappy Mind. Science, 330(6006), 932-932. doi: 10.1126/science.1192439
Kilpatrick, L. A., Suyenobu, B. Y., Smith, S. R., Bueller, J. A., Goodman, T., Creswell, J. D., Tillisch, K., Mayer, E. A., & Naliboff, B. D. (2011). Impact of Mindfulness-Based Stress Reduction training on intrinsic brain connectivity. NeuroImage, 56(1), 290–298. https://doi.org/10.1016/j.neuroimage.2011.02.034
Kross, E., Davidson, M., Weber, J., and Ochsner, K. (2009). Coping with emotions past: the neural bases of regulating affect associated with negative autobiographical memories. Biol. Psychiatry 65, 361–366. doi:10.1016/j.biopsych. 2008.10.019
Lin, M., Lucas, H. C., Jr., & Shmueli, G. (2013). Too big to fail: Large samples and the p-value
problem. Information Systems Research, 24(4), 906–
Lindahl, J. R., & Britton, W. B. (2019). “I have this feeling of not really being here:” Buddhist meditation and changes in sense of self. Journal of Consciousness Studies, 26(7–8), 157–183.
Lippelt, D. P., Hommel, B., & Colzato, L. S. (2014). Focused attention, open monitoring and loving kindness meditation: effects on attention, conflict monitoring, and creativity - A review. Frontiers in psychology, 5, 1083. https://doi.org/10.3389/fpsyg.2014.01083
Lutz, A., Slagter, H. A., Dunne, J. D., & Davidson, R. J. (2008). Attention regulation and monitoring in meditation. Trends in cognitive sciences, 12(4), 163–169. https://doi.org/10.1016/j.tics.2008.01.005
Lutz, A., Jha, A. P., Dunne, J. D., & Saron, C. D. (2015). Investigating the phenomenological matrix of mindfulness-related practices from a neurocognitive perspective. The American psychologist, 70(7), 632–658. https://doi.org/10.1037/a0039585
Lutz, J., Brühl, A., Scheerer, H., Jäncke, L., & Herwig, U. (2016). Neural correlates of mindful self-awareness in mindfulness meditators and meditation-naïve subjects revisited. Biological Psychology, 119, 21-30. doi: 10.1016/j.biopsycho.2016.06.010 Maris, E., & Oostenveld, R. (2007). Nonparametric statistical testing of EEG- and
MEG-data. Journal of neuroscience methods, 164(1), 177–190. https://doi.org/10.1016/j.jneumeth.2007.03.024
Mehling, W., Price, C., Daubenmier, J., Bartmess, E., Acree, M., Gopisetty, V., & Stewart, A. (2012). P05.61. The multidimensional assessment of interoceptive awareness (MAIA). BMC Complementary And Alternative Medicine, 12(S1). doi: 10.1186/1472-6882-12-s1-p421
Müller, K.-R., Tangermann, M., Dornhege, G., Krauledat, M., Curio, G., & Blankertz, B. (2008). Machine learning for real-time single-trial EEG-analysis: From brain–computer interfacing to mental state monitoring. Journal of Neuroscience Methods, 167, 82–90.
Nejad, A. B., Fossati, P., & Lemogne, C. (2013). Self-referential processing, rumination, and cortical midline structures in major depression. Frontiers in human neuroscience, 7, 666. https://doi.org/10.3389/fnhum.2013.00666
Tang, Y. Y., Hölzel, B. K., & Posner, M. I. (2015). The neuroscience of mindfulness meditation. Nature reviews. Neuroscience, 16(4), 213–225. https://doi.org/10.1038/nrn3916
Trapnell, P. D., & Campbell, J. D. (1999). Private self-consciousness and the five-factor model of personality: Distinguishing rumination from reflection. Journal of Personality and Social Psychology, 76(2), 284–304. https://doi.org/10.1037/0022-3514.76.2.284
Vago, D., & Zeidan, F. (2016). The brain on silent: mind wandering, mindful awareness, and states of mental tranquility. Annals Of The New York Academy Of Sciences, 1373(1), 96-113. doi: 10.1111/nyas.13171
Van Dam, N. T., van Vugt, M. K., Vago, D. R., Schmalzl, L., Saron, C. D., Olendzki, A., Meissner, T., Lazar, S. W., Kerr, C. E., Gorchov, J., Fox, K., Field, B. A., Britton, W. B., Brefczynski-Lewis, J. A., & Meyer, D. E. (2018). Mind the Hype: A Critical Evaluation and Prescriptive Agenda for Research on Mindfulness and Meditation. Perspectives on psychological science : a journal of the Association for Psychological Science, 13(1), 36–61. https://doi.org/10.1177/1745691617709589
de Vries, I. E., van Driel, J., & Olivers, C. N. (2017). Posterior α EEG Dynamics Dissociate Current from Future Goals in Working Memory-Guided Visual Search. The Journal of neuroscience : the official journal of the Society for Neuroscience, 37(6), 1591–1603. https://doi.org/10.1523/JNEUROSCI.2945-16.2016
Appendix
A. Exact instructions FTR paradigm
The practice block consisted of one round with each stimulus being presented once. The conditions were presented as visual stimuli, each presented with a white letter on grey background to indicate the current condition. “T” marked the think, “F” the feel and “R” the rest condition. A white fixation cross was shown for 3s prior to each condition, additionally a sound indicated a change in condition. The exact instructions given to the participants for the different conditions, each lasting 12s, during the task were as follows:
Feel→ “Feel into yourself, simply be aware of body sensations and/or emotions in this moment without trying to change them.”
Think → “Think about yourself, reflect who you are, what you do, like, etc.” Rest → “Do nothing specific, just await the next instruction”.
