Introduction
Different contemplative activities (ContActs) such as meditation, tai chi and yoga have been receiving a growing amount of attention in the scientific field (Chiesa & Serretti, 2010,
Wayne et al. 2014). The growing body of research into this field seems promising with studies into these practices reporting (among other effects) enhancements in cognitive functioning (Lee and Ernst, 2012; Gothe and Mcauley, 2015; Wayne et al., 2014 & Gallant, 2016). However, most of these studies have taken a holistic approach, investigating the effects of a contemplative activity as a whole. Next to this, there has also been some controversy with some studies reporting negative effects on cognitive performance as well (Chiesa, Calati & Serretti, 2010). This begs the question that if these practices indeed enhance cognition, what the underlying factors would be and why. Gerritsen & Band (2018) present a model in which they state breathing exercises as one of these underlying factors. They state that breathing exercises, a practice common in ContActs, can enhance cognition through stimulation of the vagus nerve. In this study we will test their model by investigating the effects of breathing exercises on cognitive functioning. With this, we hope we can begin to understand why ContActs seem to be effective which in turn would help us in improving these activities to improve cognitive health.
Contemplative Activities
Different types of ContActs all share many similarities, with ContActs being defined by Gerritsen & Band (2018) as activities that involve conscious and attentive exercise aimed at changing one’s mental state, contemplation in meaning comparable to ‘‘praying’’ and ‘‘meditating’’. These activities share that they instruct users to regulate ones thoughts and body to enter a certain relaxed mental state to enhance their physical and mental wellbeing (Allen, Chambers & Knight, 2006; Wayne & Fuerst, 2013). However, as previous studies have been done mostly into specific forms of ContActs, it is important to note the different types of ContActs as these will be referred to later in this paper. When looking at meditation, multiple categorizations have been made to group these techniques (Lippelt et al., 2014; Dahl et al., 2015). Using these we can distinguish between: focused attention, open monitoring, constructive and deconstructive meditation techniques. Focused attention techniques instruct the user to fix attention on a specific focus, such as breathing, while teaching how to let go of distractions. Open monitoring techniques on the other hand instruct the user to spread
techniques take on a different approach, not dictating how to use ones attention but instead focusing more on well-being of one’s self and others. And lastly, deconstructive techniques aim at breaking certain mental habits, changing the way one perceives and handles thoughts and emotions with mindfulness meditation being a well-known form of a deconstructive meditation technique. When looking at tai chi and yoga, we see that these techniques differ even further from meditation by also adding different physical exercises, such as striking poses, complex movements and muscle relaxations.
Cognitive Enhancements
Another way in which ContActs seem to be similar is in their effects on cognition, with multiple studies reporting improvements is cognitive health and enhancements in cognitive abilities (Wayne et al., 2014 & Gallant, 2016). While these effects are strongest during long-term interventions, these effects also seem to arise immediately after sessions when sessions are as short as 10 minutes (Zeidan et al., 2010; Colzato et al., 2012, 2015a,b; Gothe et al., 2013). However, these cognitive enhancements do not seem to be global, with most studies reporting distinct enhancements in cognitive control and attentional control. According to Miyake et al.’s (2000) model of fractioned executive functions, three different executive cognitive functions can be distinguished: mental set shifting (‘‘Shifting’’), information
updating and monitoring (‘‘Updating’’), and inhibition of prepotent responses (‘‘Inhibition’’). When looking at ContActs, cognitive enhancements are often found by improving inhibition instead of shifting or updating with different studies into mindfulness meditation (Gallant, 2016 & Zeidan et al., 2010) and yoga (Gothe and Mcauley, 2015) reporting this enhancement. Therefore, it seems that a main cognitive enhancement of ContActs lies in cognitive control. ContActs also seem to have an enhancing effect on attentional control , with studies into focused attention and open monitoring meditation (Shapiro et al., 2003 & Colzato et al., 2015a), mindfulness meditation (Chiesa and Serretti, 2010 & Eberth and Sedlmeier, 2012) and yoga (Gothe and Mcauley, 2015) reporting enhancements to attentional tasks. The type of enhancement to attentional control seems to differ from practice to practice however, with different practices effecting subcomponents of attention in different ways (Posner and Petersen, 1990; Slagter et al., 2011; Hommel and Colzato, 2017).
Respiratory Vagal Stimulation Model of ContAct
According to Gerritsen & Band’s (2018) Respiratory Vagal Stimulation Model of ContAct, a main manner in which ContActs enhance cognition, is through breathing exercises; a practice
commonly found in Contacts. In their model, they propose how these enhancements might arise from breathing exercises in multiple steps. First they argue that certain breathing exercises stimulate the vagus nerve, a vast cranial nerve complex with many different functions but for which a main function is to serve as a conduit between the brain and lungs and facilitate respiration (Chang et al., 2015). This stimulation of the vagus nerve would subsequently increase trait Heart Rate Variability (trait HRV), which refers to the beat-to-beat variation in heart rate brought on by parasympathetic influences while one is at rest, ignoring other influences on heart rate such as exercise. And lastly, they argue that this increase in trait HRV would induce enhancements in cognition. Earlier studies seem to support this. Firstly, previous research has shown that different types of breathing can increase trait HRV and that these increases can be observed as soon as after only 10 minutes of breathing exercises (Grossman, 1992, ; Tavares et al., 2017). During these exercises different breathing rates and ratios have different effects on trait HRV. Breathing rates slower than normal breathing of 4-6 breaths per minute seem to increase trait HRV the most in most participants although this can vary slightly from person to person (Lin, Tai & Fan, 2014; Song & Lehrer, 2003; Steffen et al., 2017). Next to this, certain measures of trait HRV have been known to reflect the activity of the vagal nerve (Chang et al., 2015) and previous studies have shown that higher trait HRV correlates with enhancements in cognitive self-control measures (Zahn et al., 2016). These findings would indeed predict that breathing exercises could enhance cognition. To gain further insight into this process one can also look at mood and arousal. Previous research has shown that the activity of the vagus nerve produces a state of relaxation (Chang et al., 2015). Next to this, previous research has also shown that HRV can be seen as an objective index of the ability regulated emotional responses through the autonomous nerves system (Appelhans & Luecken, 2006). One would therefor expect that these enhancements in cognitive
performance would be paired with more positive mood and lower levels of arousal.
Current Study
In this study we will start testing Gerritsen & Band’s (2018) model by investigating if breathing exercises enhance cognition. In particular, we will investigate the effects of a breathing exercise with a breathing rate slower than normal breathing of 6 breaths per minute (slow BE), known to increase trait HRV, compared to a breathing exercise of 16 breaths per minute (normal BE), a breathing rate close to normal breathing at rest which should not increase trait HRV. As this type of slow BE stems from ContActs we will test for
cognitive control and attentional control. Next to this, as we expect increases in trait HRV to be paired with more positive mood and lower levels of arousal we will also test for changes in mood and arousal.
Effects on cognitive control.
To test for enhancements in cognitive control we will use a Simon task, known to tax inhibition. First used by Craft and Simon (1970) to investigate the Simon effect which they had discovered one year prior (Hui Lu & Proctor, 1995). The Simon is the observation that an irrelevant location of a stimulus has an effect on the time to react on that stimulus with a lateralized response. When the location of the given stimulus and the location of the required reaction are congruent (for example a shape appearing on the right side of a screen and pressing a button with your right hand) the reaction time is faster than when the location of the stimulus and the associated reaction are incongruent (a shape appearing on the right side but pushing a button with your left hand). Even though the participant is told the location of the stimulus is irrelevant, the spatial information is taken into account involuntarily and must be inhibited during incongruent trials. As we expect the slow BE to enhance inhibition, we hypothesize that the increase in reaction time from congruent to incongruent trials will be lessened after the slow BE more so then after the normal BE. Next to the Simon effect, another effect known to emerge during Simon tasks is the Gratton effect (Gratton et al., 1992). The Gratton effect refers to the finding that the Simon effect is smaller on trials following an incongruent task than it is after a congruent one. Previous research has shown that this effect can, for a large part, be explained by the conflict monitoring hypothesis (Blais, Stefanidi & Brewer, 2014). This hypothesis states that an incongruent trail during the Simon task signals for “a need of control” for the following trial, resulting in increased cognitive control for this following trial. This results in a decrease in reaction in trials following
incongruent trials compared to trials following a congruent one. As we expect the slow BE to increase cognitive control, we hypothesize that the Gratton effect will be stronger following the slow BE compared to the normal BE.
Effects on attentional control.
To test for enhancements in attentional control, we will use an Attentional Blink (AB) task, first used by Raymond, Shapiro and Arnell (1992), to test one’s ability to allocate attentional resources (Dux & Marois, 2009). The AB task is used to measure the effect of the same name which states that participants have difficulties in processing multiple visual stimuli in rapid
succession. During the task, participants are asked to find two targets (T1 and T2) within the set of stimuli. By changing the amount of distractor stimuli that are presented between T1 and T2, the lag interval between the targets can be manipulated. To score performance on the task, the accuracy of detecting T2, given participants perceived T1 (T2|T1) is measured for the different lag intervals. Previous research has shown that participants mostly have trouble accurately reporting T2|T1 when the lag time between the two targets is between 200-500 ms, with the lowest accuracy of T2|T1 given a lag interval of approximately 300 ms (Dux & Marois, 2009). These studies suggest that the attentional blink is due to attentional recourses still being allocated to the first target at the time the second target is presented, as processing of the first target is attentionally demanding. As we expect the slow BE to enhance the
flexibility of attentional resources, we expect the window of the attentional blink to be smaller after the slow BE. We therefor hypothesize that T2|T1 around the 300ms lag point will be increased more after the slow BE than after the normal BE.
Effects on mood and arousal
To test for effects on mood and arousal we will use an Affect Grid (Russel, Weiss & Mendelsohn, 1989) to assess self-reported affect along the dimensions of negative-positive mood and sleepiness-arousal. As we expect the slow BE to increase trait HRV, we expect that participants will recover more from the cognitive tasks by performing the slow BE than by performing the normal BE. We therefor hypothesize mood will be increased more after the slow BE than after the normal BE and that arousal will be decreased more after the slow BE than after the normal BE.
Methods Participants
A total of 57 participants were recruited for this study. All participants were recruited within the Faculty of Social Sciences (FSW) using SONA, an online platform used by the faculty where researchers post their studies and participants can sign up for these studies. Participants who had respiratory problems, heavily impaired vision or a severe neurological condition were excluded. Next to this, participants that were on medication which could influence their cognitive abilities, motor skills or sight were also excluded. Lastly, participants who did not maintain the correct breathing rate during either breathing exercise were excluded from further analysis. To determine this, we chose to look at mean breathing rate. A participant
who correctly performed the breathing exercises would have a mean breathing rate of 16 for the normal BE and a mean breathing rate of 6 for the slow BE respectively. Any periods that the participant would not maintain this breathing rate would result in a deviation from this mean. Therefore, we can use the mean breathing rate as a measurement of how well the participant performed the breathing exercise. In this study, we chose to accept participants with a mean breathing rate within a 20% range from the intended breathing rate. Therefore, participants were excluded if their mean breathing rate fell outside 4.8-7.2 breaths per minute during the slow BE or outside 12.8-19.2 breaths per minute during the normal BE.
Participants were compensated by 6 University credits, needed by students of the first year course, or € 20,00. This study was approved by the Psychology Research Ethics Committee (CEP) of Leiden University.
Design
During this study a 2x2 within-subject design was used. Participants came into the lab two times within 1,5 weeks for two sessions. During each session, one of the two breathing exercises was performed (slow BE and normal BE) and during each session 2 measurements of cognitive abilities were made, both directly before and directly after the BE (pre- and posttest) giving the two factors in our design (slow BE vs. normal BE x pre- vs. post-test). The pre- and posttest both consisted out of a Simon and an AB task. The study was
counterbalanced for both task order (either the Simon or the AB task was performed first during all pre- and post-tests) and BE order (either the slow BE or the normal BE was performed during first session). This design gives us our two independent variables: time (pre- vs. posttest) and breathing exercise (slow BE vs. normal BE). Time is measured ordinal, with the times of the pre- and post-tests as the two measuring points and breathing exercise functions as a nominal variable, with each breathing exercise as one of the two levels of this variable. During the breathing exercises, a pneumograph is used to measure breathing rate and ratio to determine if the exercises are performed correctly. The dependent variables in this study are the performances on the Simon and AB task during the pre- and post-tests. For the Simon task both the Simon and Gratton effects will be determined. For the Simon effect the mean reaction time for congruent and incongruent trials will be measured. For the Gratton effect the mean reaction time will be measured for congruent (C) and incongruent (I) trials, given a previous congruent or incongruent trial giving four different trail orders; C-C, C-I, I-C, I-I. For the AB task, the amount of correctly perceived T1 stimuli and T2|T1 stimuli will be measured. These will also be measured on an interval level. Next, an affect grid is used to
asses both arousal and mood at an interval level both at the beginning and end of each session. These will also be used as dependent variables to check for effects on mood and arousal. Next to these variables, multiple possible mediating variables will also be taken into account. The mean breathing rate and ratio measured by the pneumograph will be used as mediating
variables. Previous experience with ContActs will be measured and used as a between-subject variable. A questionnaire will be used to assess previous experience with ContActs on an ordinal scale by distributing each participant over four groups determining the level of prior experience. Lastly, effects of the two factors counterbalanced for, task order and breathing exercise order, will also be checked by using them as between-subject variables.
Procedure
During the study, participants were received at the FSW research labs. The first time they came in participants were seated next to the researcher and briefed on the study during which they were reminded that the study consisted of two sessions and were informed on the overall structure of each session: cognitive tasks, a breathing exercise and cognitive tasks again. Afterwards they were asked to fill out an online questionnaire (using a link sent to them via email to a Qualtrics questionnaire) to collect demographic data and to determine if they had any prior experience concerning any form of ContAct. Next, they were fitted with the pneumograph and escorted to a small room with a desk and a desktop computer. Here, they first filled out an Affect Grid. Then, participants performed two cognitive tasks; a Simon and an AB task; the pre-test. Thereafter, they performed a breathing exercise, described below. After the breathing exercise participants performed the two cognitive tasks again; the post-test. Lastly, participants filled out the Affect Grid once again to determine any effects the current study might have had on their mood and arousal levels. All these different components are described in detail below. Participants came in two times within a week performing each of the two breathing exercises (slow BE and normal BE). Each session took approximately 60-90 minutes making a total of 120-180 minutes over two sessions with most variation in session length due to the varying length of the AB task, which participants could pace
themselves. Instructions between the two sessions did not differ in any way to ensure that any effects of instructions on measurements were equal during both sessions. After the second session, participants were debriefed and given a more detailed explanation of the goal and content of the current study. A short summary of the underlying literature was given together with a more detailed description of the structure of the experiment describing the reasoning behind the Simon and AB tasks and the two types of breathing exercises. Next to this,
participants were allowed to leave their e-mail address if they wished to be updated on the results of the study .
Breathing exercise.
This breathing exercise intervention was developed for this study to induce a certain breathing pattern. During the intervention, participants are fitted with a pneumograph to measure their breathing rate and ratio. The intervention consists out of two parts: a 3 minute trial period and a 15 minute long exercise. During both parts, participants watch an animation which
represents their own breathing. The animation consists out of a blue circle which first grows (representing inhalation) and then shrinks (representing exhalation). Participants are asked to follow the rhythm of the animation with the rhythm of their breath. During the trial period, the researcher checks the measurements of the pneumograph by hand to make sure the band is on correctly, stable measurements are made and the right respiration rates and ratios are
sustained. If not, the pneumograph is corrected and/or extra instructions on the breathing exercise are given after the trial period before starting the actual exercise. Pneumograph measurements made during the actual exercise are later analyzed using the PhysioData Toolbox, developed by Sjak-Shie (2018) at Leiden University, to determine if the right breathing rate was maintained. Two different types of intervention were used; one for each of the two sessions. During one, the intervention tried to induce a breathing rate of 6 breaths per minute (slow BE) by depicting the animation 6 times per minute, during the other the
intervention tried to induce a breathing rate of 16 breaths per minute (normal BE) by depicting the animation 16 times per minute. During both interventions, respiration ratios remained 1:1.
Materials
Simon task.
During this study we used a house version of the Simon task provided by the department of Psychology of the Social and Behavioural Sciences faculty of Leiden University. During the task, participants are seated in front of a desktop computer and are instructed to place one hand on the Q key of their keyboard and the other on the P key. The user is then presented with 3 blocks of 60 trials which they must complete, preceded 1 set of 4 practice trials. During the different sets of trials, the participant is presented with a fixation point on the middle of the screen which the participant is instructed to fixate on. During each trial, the participant
first has 1 second to focus on the fixation point after which a colored circle will appear on the screen for 1,5 seconds, either on the right or left side of the fixation point. Participants are instructed to press the Q key any time a green circle appears and to press the P key any time a blue circle appears. Accuracy and reaction time are measured for congruent and incongruent trials respectively.
Attentional blink task.
During this study we used a house version of the Attentional Blink task, again provided by the department of Psychology of the Social and Behavioural Sciences faculty of Leiden
University. During the task participants are seated in front of a desktop computer and are instructed that during the task they will see a set of letters flash on the screen. Participants are asked to search for the target stimulus; either the letter ‘B’, ‘G’ or ‘S’ (T1) between the set of distractors. If they can identify the target stimulus they are also asked to check if another target stimulus, the letter ‘X’ (T2), appeared afterward. The user is then presented by 3 blocks of 64 trials, preceded by a block of 3 practice trials. During each trial, an ‘*’ flashes for 160 ms as a fixation point foreshadowing where the set of letters will be flashed on the screen afterwards. All letters are presented for 75 ms with an interstimulus interval of 15 ms between them. After the set of stimuli is shown, participants are presented with two questions: ‘Which target was presented?’ and ‘Did you see letter ‘X’ afterward?’. In half of all trials (96 of the total 192), no T2 is presented. In the other half of the trials the lag distance between T1 and T2 is varied from presenting T2 directly after T1 (lag +1) to presenting it as the 8th stimulus after T1 (lag +8). This results in lag intervals of 90, 180, 270, 360, 450, 540, 630 and 720 ms between the presentations of T1 and T2. In total 12 trials are presented for each of the 8 lag intervals. Both T1 and T2|T1 are measured for each lag interval respectively.
Affect grid.
The Affect Grid used in this study consists of a grid of squares with the level of arousal on one axis and the mood (from negative to positive) on another. Both axis are distributed in a 9 point scale, making the grid of 9 by 9 squares. Participants select a single square to
simultaneously score on both axis, which gives insight on the current emotional state. Different states can be identified in this way. High scores on both axis indicates excitement, high on arousal but low on mood indicates stress, low on arousal but high on mood indicates relaxation and low on both axis indicates depression.
Demographic questionnaire.
A short questionnaire was used to assess the amount of previous experience a participant has with ContActs (if any) and to gather demographic data. The questionnaire was developed and used previously by the department of Psychology of the Social and Behavioural Sciences faculty of Leiden University. The questionnaire consists of 15 items. The first few items asses basic demographics (gender, age, education) and determine if there are strong factors which might influence cognitive capabilities (strong medication, neurological conditions). The rest of the questionnaire assesses if the participant is currently practicing meditation, Tai Chi, yoga or some other form of ContAct or has done so in the past and if so, how frequently. If a participant is currently practicing the frequency of practice is determined using an ordinal 6-point scale. If a participant is not currently practicing but has practiced in the past, an ordinal 5-point scale is used to determine how long ago this practice was performed and the same 6-point ordinal scale is used to assess the frequency of practice at that time. Using the answers on these questions, participants are distributed to one of four levels of prior experience on an ordinal scale: Active practitioners, users with heavy experience, users with some experience and users without experience. Participants who practice at least one type of ContAct at least once a month are categorized as active practitioners. Participants who practice ContActs less than once a month or who have practiced at least one type of ContAct at least once a month less than a year ago are categorized as having heavy experience. Participants who practiced some form of Contact in the past but did so more than a year ago and/or less than once a month are categorized as having some experience. Participants who have never performed any type of ContAct are categorized as having no experience.
Analyses
Data preparation.
Data from the pneumographs is analyzed to check if participants followed instructions and performed the breathing exercises correctly. Using the matlab scripts within PhysioData Toolbox (Sjak-Shie, 2018), we calculate the mean breathing rate and ratio from the raw data the pneumographs provide and determine which participants are excluded from further analysis. Participants with a mean breathing rate more than 20% above or below the intended breathing rate are excluded from further analysis. Next, the demographic questionnaire is analyzed to categorize each participant as an active practitioner, someone who has heavy experience, some experience or none at all. Then, mood and arousal scores are extracted from
the affect grids. Next, data from the cognitive tasks is prepared. For the Simon task, mean reaction time is calculated for all correctly answered congruent and incongruent trails and for each type of trial given a previous trial (C-C, C-I, I-C, I-I). For the AB task, the total absolute amount of detected T1 and T2|T1 is calculated, as well as the relative amounts of detected T1 and T2|T1 per lag interval. Lastly, the data from the Simon and AB tasks is checked for outliers and extremes using the Inter Quartile Range rule (Hoaglin & Iglewicz, 1987).For the Simon task, the mean reaction time on correctly answered congruent and incongruent trials is screened. For the AB task, the total amount of correctly detected T1 and T2|T1 is screened. Any extremes found in this way are removed from further analysis.
Statistical analysis.
Further statistical analysis will be done by use of multiple multi-way repeated measures ANOVA’s. To test for reduction of the Simon effect after the slow BE, a 2*2*2 multi-way way repeated measures ANOVA will be performed on reaction time with factors time (pre- vs. post-test), breathing exercise (slow BE vs. normal BE) and congruency (congruent vs. incongruent). To test for the increase in the Gratton effect after the slow BE, a 2*2*4 multi-way repeated measures ANOVA will be performed, interchanging the factor congruency for task order (C-C vs. C-I vs. I-C vs. I-I). To test if the slow BE reduced the attentional blink, a 2*2*8 and a 2*2*2 multi-way repeated measures ANOVA will be performed on percentage of T2|T1, again keeping the factors of time and breathing exercise but adding the lag interval as a third factor during both ANOVA’s. During the first ANOVA, all eight lag intervals will be taken into account (lag+1 vs. lag+2 vs. … vs. lag+8). For the second ANOVA only two lag intervals will be taken into account; lag +3, a lag interval known to lie well within the
attentional blink window and lag +7, a lag interval known to lie well outside of the attentional blink window. Next, these four multi-way repeated measures ANOVA’s will be performed multiple times again but this time while checking for mediation effects of the measured breathing rate, the measured breathing ratio, arousal, mood, prior experience, cognitive task order and breathing exercise order. Each ANOVA will be performed 7 times again while adding either measured breathing rate, the measured breathing ratio, pretest mood or pretest arousal as a covariate or by adding prior experience, cognitive task order or breathing exercise order as a between-subjects variable.
Lastly, to explore the effects of the different breathing exercises on mood and arousal, two 2*2 multi-way repeated measures ANOVA will be performed on levels of mood and arousal again with factors time (pre- vs. post-test) and breathing exercise (slow BE vs. normal BE). Again, these multi-way repeated measures ANOVA’s will be performed multiple times again while checking for mediation effects of the measured breathing rate, the measured breathing ratio, prior experience, cognitive task order and breathing exercise order. Each ANOVA will be performed 5 times again while adding either measured breathing rate, the measured breathing ratio as a covariate or by adding prior experience, cognitive task order or breathing exercise order as a between-subjects variable.
Results Data screening
Of the 57 recruited participants, 2 were unable to attend the second session and were therefor excluded from further analysis. Analysis of the data registered by the pneumograph indicated that 8 participants did not maintain the right breathing rate during one of their sessions. Two participants were unable to maintain a high enough breathing rate during the normal BE and 6 were unable to keep a slow enough breathing rate during the slow BE. These participants were therefor also excluded from further analysis. Of the remaining 47 participants, the mean age was 22,21 (Min= 18, Max= 52, SD = 4,899), 9 were male and 38 were female, 21 were Dutch speaking and 26 were English speaking and 12 were active practitioners of ContAct, 12 had heavy experience, 10 had some experience and 13 had no experience. Next, when using the Inter Quartile Range rule (Hoaglin and Iglewicz, 1987) to screen for extremes in the data of the Simon task and the AB task no extremes were found. Therefore, no further participants were excluded from data analysis. However, when examining the data from the AB task, we found an extremely low amount of T2|T1 of 32.69% for the +7 lag interval. Further
investigation revealed that during the experiment the T2 stimuli for the lag interval +7 stimuli were only shown for 15 ms instead of 75 ms. This lag interval was there for scrapped from further analysis.
Data analysis
Analysis of Simon effect. Analysis of main factors.
Table 1. shows the mean reaction times on congruent and incongruent trials during the different measurement points of our study. A multi-way repeated measures ANOVA with factors time (pre- vs. post-test), breathing exercise (slow BE vs. normal BE) and congruency (congruent vs. incongruent) found multiple significant results. A main effect for congruency was found (F(1,46) = 130.665, p < .001) with reaction time on incongruent trials being significantly slower than on congruent trials, as well as a main effect for time (F(1,46)= 11.038, p = .002) with reaction times being significantly faster during the posttests than during the pretests. Next to these main effects, an interaction effect was also found between breathing exercise and congruency (F(1,46)= 5.403, p = .025) with the reaction time of incongruent trials being significantly faster during the session where the slow BE was performed. Graph 1. shows the changes in reaction time from pre- to post test for both congruencies.
Analysis of covariates and between-subject variables.
When conducting the multi-way repeated measures ANOVA with factors time, breathing exercise and congruency again while adding prior experience, cognitive task order or breathing exercise order as a between-subjects variable, no additional effects were found. When instead pretest mood, pretest arousal, measured breathing rate or measured breathing ratio were added as covariate, no additional effects were found either.
Analysis of Gratton effect. Analysis of main factors.
Table 2. shows the mean reaction time per trial given a previous trial: C-C, C-I, I-C, I-I. A multi-way repeated measures ANOVA with factors time (pre- vs. post-test), breathing exercise (slow BE vs. normal BE) and trial order (congruent vs.
congruent-incongruent vs. congruent-incongruent-congruent vs. congruent-incongruent-congruent-incongruent) found similar significant results as the previous multi-way repeated measures ANOVA. Again a main effect for time was found (F(1,46)= 10.657, p = .002) with reaction times being significantly lower during the posttests than during the pretests. A main effect for trial order was also found (F(3,44)= 58.798, p< .001). Pairwise comparisons using a Bonferroni correction then showed that all trial orders significantly differed from each other (p< .001) except for the trial orders
incongruent-congruent and incongruent-incongruent. With these comparisons, reaction times could now be ranked as such: C-C < I-C and I-I < C-I.
Analysis of covariates and between-subject variables.
When conducting the multi-way repeated measures ANOVA with factors time, breathing exercise and trial order again while adding prior experience or cognitive task order as a between-subjects variable, or pretest mood, pretest arousal, measured breathing rate or measured breathing ratio as covariate, no additional effects were found. However, when breathing exercise order was added as a between-subjects variable a significant interaction effect was found between breathing exercise, breathing exercise order and trial order
(F(3,43)= 6.152, p = .001). Further investigation of this interaction effect showed that reaction times for trial orders C-I, I-C and I-I are all significantly slower during the session with the normal BE, if this was the first session.
Analysis of Attentional Blink. Analysis of main factors.
A multi-way repeated measures ANOVA with factors time (pre- vs. post-test), breathing exercise (slow BE vs. normal BE) and lag interval (lag +1 vs. … lag +7) found two
significant main effects. A main effect of time was found (F(1,46)= 43.594, p< .001) with the amount of T2|T1 being higher during posttests than during pretests. Next to that, a main effect of lag interval was found (F(6,41)= 13.438, p< .001). Pairwise comparisons using a
Bonferroni correction then showed that during lag +2 and lag +3 the amount of detected T2|T1 was significantly lower than during all other lag intervals (p < .001) , lag +6, lag +7 were significantly higher than all other lag intervals (p = .006 or smaller) and lag +1, lag +4 and lag +5 had an amount of detected T2|T1 in between that differed significantly from the other two groups with lag +1 not differing from lag +4 and lag +5 (p = .007) but lag +4 and lag +5 being significantly different from one another. These two effects are clearly illustrated in graphs 2. Which show the mean amount of T2|T1 in percentages during the pre- and posttests of both sessions during which either the normal BE or slow BE was performed. This multi-way repeated measures ANOVA was then performed again, only taking into account 2 lag intervals; lag +3 and lag +7 and again found two main effects. A main effect for time (F(1,46)= 16.584, p< .001) was found with the amount of T2|T1 being higher during posttests than during pretests and a main effect for lag interval (F(1,46)= 66.066, p< .001) was found with the amount of T2|T1 being higher during lag +7 than lag +3. Next to these main effects two interaction effects were also found. An interaction effect was found between time and breathing exercise (F(1,46)= 5.697, p= .021) with the amount of T2|T1 increasing more from
pre- to posttest during the session with the normal BE then during the session with the slow BE. Next to this, an interaction effect was also found between time and lag interval (F(1,46)= 6.571, p= .014) with the amount of T2|T1 increasing more from pre- to posttest for the +3 lag interval than for the +7 lag interval.
Analysis of covariates and between-subject variables.
When conducting the multi-way repeated measures ANOVA with factors time, breathing exercise and lag interval (7 intervals) while adding prior experience or cognitive task order as a between-subjects variable, or pretest mood, pretest arousal, measured breathing rate or measured breathing ratio as covariate, no additional effects were found. However, when breathing exercise order was added as a between subjects variable a significant interaction effect was found between breathing exercise and breathing exercise order (F(1,45)= 48.632, p < .001). Further investigation of this interaction effect showed that the amount of detected T2|T1 during the first session was significantly lower than the second session if this was the session with the slow BE. When conducting the same multi-way repeated measures
ANOVA’s with factors time, breathing exercise and lag interval while using only the +3 and +7 lag intervals and adding prior experience or cognitive task order as a between-subjects variable, or pretest mood, pretest arousal, measured breathing rate or measured breathing ratio as covariate, no additional effects were found. However, when breathing exercise order was added as a between subjects variable a significant interaction effects was again found between breathing exercise and breathing exercise order (F(1,45)= 29.373, p< .001) with the amount of detected T2|T1 during the first session being significantly lower than the second session if the first session was the session with the slow BE.
Analysis of effects on mood and arousal. Analysis of main factors.
A multi-way repeated measures ANOVA with factors time (pre- vs. post-test) and breathing exercise (slow BE vs. normal BE) did not find any significant effects on mood. Another multi-way repeated measures ANOVA with factors time (pre- vs. post-test) and breathing exercise (slow BE vs. normal BE) did find significant main effects on arousal however. A main effect was found for time (F(1,46)= 52.751, p< .001) with self-reported arousal being significantly lower after the posttest than at the beginning of the sessions. Another main effect
for breathing exercise (F(1,46)= 5.064, p= .029) was also found, with self-reported arousal being lower during the session with the slow BE than during the session with the normal BE.
Analysis of between-subject variables.
When again conducting a multi-way repeated measures ANOVA with factors time and breathing exercise on mood with cognitive task order as a between-subjects variable, a significant interaction effect is found between breathing exercise and cognitive task order (F(1,45)= 9.546, p= .003) with mood being more positive during the session with the slow BE if the AB task is performed first than during the session with the normal BE. When
instead breathing exercise order or prior experience were added as between-subject variables, or measured breathing rate or measured breathing ratio as covariates, no significant effects were found. When conducting a multi-way repeated measures ANOVA with factors time and breathing exercise on arousal with cognitive task order as a between-subjects variable, a significant interaction effect is found between time and cognitive task order (F(1,45)= 4.408, p= .041) with the decrease in arousal form the start to the end of the session being
significantly larger when the AB task is performed first. When instead breathing exercise order or prior experience were added as between-subject variables or measured breathing rate or measured breathing ratio as covariates, no significant effects were found.
Discussion
In this study we set out to investigate if, following Gerritsen & Band’s (2018) Respiratory Vagal Stimulation Model of ContAct, breathing exercises with a slower than normal breathing rate enhance cognitive and attentional control. To do this we used a 2*2 within-subject design in which we tested cognitive abilities both before and after a slow breathing exercise of 6 breaths per minute and after a normal breathing exercise of 16 breaths per minute. During the statistical analysis, Simon, Gratton and AB effects were all measured as expected. A Simon effect was found with reaction times on incongruent trials being significantly slower than on congruent ones. A Gratton effect was found, with cognitive control increasing after
incongruent trials. This resulted in well needed higher attentional control during the I-I trial order, resulting in a faster reaction time on I-I trials compared to C-I trials. But this also resulted in unneeded extra attentional control during the I-C trial order, resulting in a lower reaction time on I-C trials compared to C-C trials. An AB effect was found with the amount of reported T2|T1 being lower around the 200ms - 500ms lag interval. Statistical analysis did not however reveal that the slow BE enhanced cognitive abilities more so than the normal BE.
But, both breathing exercises did seem to have strong enhancing effects on both cognitive and attentional control, with both reaction time on the Simon task going down significantly and percentages of perceived T2|T1 going up significantly from pre- to post tests. This could be due to many things. One possible explanation for this is that not the act of breathing at a certain rate was the factor which enhanced cognition, but rather that focusing on one’s breath was. When comparing our intervention with the different known ContActs, we find many similarities between both sessions and focused attention meditation. Both during the slow BE and the normal BE, participants focused on their breath by following the animation which might have led them to let go of other distractors. This would have led to the same therapeutic effect of focused attention meditation after both breathing exercises, which might explain why increases in performance were present during both sessions. Another possible explanation for this is that test-retest effects were to influential. As the cognitive tasks during both sessions followed each-other relatively quickly and the repeatability of these specific versions of the AB and Simon tasks were not established beforehand, the increases in performance from pre- to post-test might be due to a low test-retest reliability. These effects might have
overshadowed the effects of the intervention. Next to this, a large part of our participants were already heavily experienced with ContActs, with 22 of the 47 participants being active
practitioners or having practiced a ContAct monthly in the past. This high level of experience might have overshadowed the effects of the intervention as well. Lastly, another explanation for the null results may lie with learning effects as some of our findings support this. First, possible learning effects can be detected with the amount of detected T2|T1 being
significantly lower during the pretest of the first session than most other tests; the first time participants would perform the AB task. Next to this, some results seem to point to the fact that effects of the breathing exercise seemed worse during the first session if this was the session with the normal BE and better during the second session if this was the session with the slow BE. This is indicated with reaction times on all trial of the Simon task being
significantly higher during the first session if this was the session with the normal BE and the amount of detected T2|T1 increasing more from the first to second session if the second session was the session with the slow BE. It therefor seems that cognitive enhancements due to the slow BE might be higher than those brought on by the normal BE, but that these effects are inferior to the effect of having experience with the experimental setup. In the end
however, it is difficult to say with any amount of confidence which effects led to our null findings.
However, even if no direct results were found, there are some factors which might indicate that the slow BE did have a stronger enhancing effect on cognition than the normal BE. First, even though these results are not significant, a tendency can be seen that reaction time on the Simon task is lowered more from pre- to posttest during the slow BE than during the normal BE session. Next to this, an effect on arousal was found for the slow BE, with arousal being lower during the session with the slow BE than during the session with the normal BE. And as this relaxation is linked to trait HRV and vagus nerve activity, this might indicate that
cognition was indeed also enhanced. To conclude, it would therefor seem that even though this study cannot conclude our main hypothesis that a slow BE enhances cognition more so then a normal BE, it cannot be dismissed either as the negative results might very well be due to the limitations of this study. Our first limitation were our participants themselves. As many had heavy experience with ContActs already and it is impossible to say how this might have influenced the effects of the intervention. Next to this, the cognitive tasks used in this study weren’t thoroughly tested before being used in this study, leading to a possibly low
repeatability. Lastly, we were also limited with the low amount of sessions. With only two sessions, possible learning effects could already be found within our data. Further research must therefore be conducted to counteract these effects. We propose a new setup in which participants conduct three separate sessions, the first of which would only function as a way to familiarize participants with the experimental setup. The second and third of which would be sessions with a slow and normal BE to be focused on during the analyzed. During this experiment, only participants without prior experience with ContActs would need to be allowed, to eliminate any possible effects of experience. With that, we hope to further investigate the link between breathing and cognitive enhancements.
Appendix Table 1
Mean Reaction times on Simon task trials
Breathing Exercise Trial Congruency Pre-test Post-test
Mean SD Mean SD
6 BPM Congruent 432,73 59,86 417,03 58,13
Incongruent 464,65 66,25 451,52 64,32
16 BPM Congruent 433,90 67,07 424,52 60,59
Table 2
Mean Reaction times on Simon task trials, given a previous trial
Breathing Exercise Trial order Pre-test Post-test
Mean SD Mean SD 6 BPM Congruent-congruent 414,74 58,88 403,91 60,78 Congruent-incongruent 477,70 66,58 461,73 63,53 Incongruent-congruent 449,83 63,73 430,29 58,24 Incongruent-incongruent 451,31 69,12 441,72 68,35 16 BPM Congruent-congruent 417,28 63,45 409,60 55,32 Congruent-incongruent 467,68 75,34 464,84 65,89 Incongruent-congruent 449,75 73,66 438,76 68,05 Incongruent-incongruent 452,59 79,92 440,84 64,01
Figure 1. Mean reaction time Simon trials 410 420 430 440 450 460 470 Pre-test Post-test R ea cti on T im e Test Trials Incongruent Normal BE Incongruent Slow Be Congruent Normal BE Congruent Slow BE
Figure 2. Mean % of T2|T1 during the AB task 0 10 20 30 40 50 60 70 80 +1 +2 +3 +4 +5 +6 +7 % of T 2| T 1 Lag Test Pretest Normal BE Posttest Normal BE Pretest Slow BE Posttest Slow Be
References
Allen, N. B., Chambers, R. & Knight, W. (2006). Mindfulness-based psychotherapies: a review of conceptual foundations, empirical evidence and practical considerations. Australian and New Zealand Journal of Psychiatry, 40, 285-294.
Appelhans, B. M. & Luecken, L. J. (2006). Heart Rate Variability as an Index of Regulated Emotional Responding. Review of General Psychology, 10(3), 229-240.
Blais, C., Stefanidi, A. & Brewer, G. A. (2014). The Gratton effect remains after controlling for contingencies and stimulus repetitions. Frontiers in Psychology, 5:1207.
Chang, R. B., Strochlic, D. E., Umans, B. D., Liberles, S. D., & Williams, E. K. (2015). Vagal sensory neuron subtypes that differentially control breathing. Cell 161, 622–633.
Chiesa, A., Calati R. & Serretti A. (2010). Does mindfulness training improve cognitive abilities? A systematic review of neuropsychological findings. Clinical Psychology Review, 31, 449-464.
Chiesa, A., and Serretti, A. (2010). A systematic review of neurobiological and clinical features of mindfulness meditations. Psychol. Med. 40, 1239–1252.
Colzato, L. S., Ozturk, A., and Hommel, B. (2012). Meditate to create: the impact of focused-attention and open-monitoring training on convergent and divergent thinking. Front. Psychol. 3:116.
Colzato, L. S., Sellaro, R., Samara, I., Baas, M., and Hommel, B. (2015a). Meditation-induced states predict attentional control over time. Conscious. Cogn. 37, 57–62.
Colzato, L. S., Sellaro, R., Samara, I., and Hommel, B. (2015b). Meditation-induced
cognitive-control states regulate response-conflict adaptation: evidence from trial-to-trial adjustments in the Simon task. Conscious. Cogn. 35, 110–114.
Craft, J. L., & Simon, J. R. (1970). Processing symbolic information from a visual display: Interference from an irrelevant directional cue. Journal of Experimental Psychology, 83, 415-420.
Dahl, C. J., Lutz, A., and Davidson, R. J. (2015). Reconstructing and deconstructing the self: cognitive mechanisms in meditation practice. Trends Cogn. Sci. 19,515–523.
Dux, P. E. & Marois, R. (2009). The attentional blink: A review of data and theory. Attention, Perception, & Psychophysics, 71(8), 1683-1700.
Eberth, J., and Sedlmeier, P. (2012). The effects of mindfulness meditation: a metaanalysis. Mindfulness 3, 174–189.
Gallant, S.N. (2016). Mindfulness meditation practice and executive functioning: Breaking down the benefit. Consciousness and Cognition, 40, 116-130.
Gerritsen R.J.S. & Band G.P.H. (2018). Breath of Life: The Respiratory Vagal Stimulation Model of Contemplative Activity. Front. Hum. Neurosci. 12:397.
Gothe, N., Pontifex, M. B., Hillman, C., and McAuley, E. (2013). The acute effects of yoga on executive function. J. Phys. Act. Health 10, 488–495.
Gothe, N. P., and Mcauley, E. (2015). Yoga and cognition: a meta-analysis of chronic and acute effects. Psychosom. Med. 77, 784–797.
Gratton G., Coles M., Donchin E. (1992). Optimizing the use of information: strategic control of activation of responses. J. Exp. Psychol. Gen. 121, 480–506.
Grossman, P. (1992). Respiratory and cardiac rhythms as windows to central and autonomic biobehavioral regulation: Selection of window frames, keeping the panes clean and viewing the neural topography. Biological Psychology, 34, 131-161.
Grossman, P., Niemann, L., Schmidt, S., and Walach, H. (2004). Mindfulnessbased stress reduction and health benefits: a meta-analysis. J. Psychosom. Res. 57, 35–43.
Hoaglin, D.C. and Iglewicz, B. (1987). Fine-Tuning Some Resistant Rules for Outlier Labeling. Journal of the American Statistical Association. 82, 1147-1149.
Hommel, B., and Colzato, L. S. (2017). Meditation and metacontrol. J. Cogn. Enhanc. 1, 115–121.
Hui Lu, C. & Proctor, R. W. (1995). The influence of irrelevant location information on performance: A review of the Simon and spatial Stroop effects. Psychonomic Bulletin & Review 1995, 2(2), 174- 207.
Lee, M. S., and Ernst, E. (2012). Systematic reviews of t’ai chi: an overview. Br. J. Sports Med. 46, 713–718.
Lin, I. M., Tai, L. Y. & Fan, S. Y. (2014). Breathing at a rate of 5.5 breaths per minute with equal inhalation-to-exhalation ratio increases heart rate variability. International Journal of Psychophysiology, 91, 206-211.
Lippelt, D. P., Hommel, B., and Colzato, L. S. (2014). Focused attention, open monitoring and loving kindness meditation: effects on attention, conflict monitoring and reativity—A review. Front. Psychol. 5:1083.
Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A. H. & Howerter, A. (2000). The Unity and Diversity of Executive Functions and Their Contributions to Complex ‘‘Frontal Lobe’’ Tasks: A Latent Variable Analysis. Cognitive psychology, 41, 49-100. Posner, M. I., and Petersen, S. E. (1990). The attention system of the human brain. Annu.
Rev. Neurosci. 13, 25–42.
Raymond, J. E., Shapiro, K. L. & Arnell, K. M. (1992). Temporary suppression of visual processing in an RSVP task: An attentional blink? Journal of Experimental psychology: Human Perception & Performance, 18, 849-860.
Russel, J. A., Weiss, A. & Mendelsohn, G. A. (1989). Affect Grid: A Single-Item Scale of Pleasure and Arousal. Journal of Personality and Social Psychology, 57(3), 493-502
Shapiro, S. L., Walsh, R., and Britton, W. B. (2003). An analysis of recent meditation research and suggestions for future directions. Hum. Psychol. 31, 86–114.
Sjak-Shie, E. E. (2018). PhysioData Toolbox (Version 0.4) [Computer software]. Retrieved from https://PhysioDataToolbox.leidenuniv.nl
Slagter, H. A., Davidson, R. J., and Lutz, A. (2011). Mental training as a tool in the neuroscientific study of brain and cognitive plasticity. Front. Hum. Neurosci. 5:17.
Song, H., & Lehrer, P. M. (2003). The Effects of Specific Respiratory Rates on Heart Rate and Heart Rate Variability. Applied Psychophysiology and Biofeedback, 28(1).
Steffen, P. R., Austin, T., DeBarros, A. & Brown, T. (2017). The Impact of Resonance Frequency Breathing on Measures of Heart Rate Variability, Blood Pressure, and Mood. Frontiers in Public Health, 5, 222.
Tavares, B. S., Vidigal, G. P., Garner, D. M., Raimundo, R. D., de Abreu, L. C. & Valenti, V. E. (2017). Effects of guided breath exercise on complex behaviour of heart rate
dynamics. Clinical Physiological Functional Imaging, 37, 622-629.
Wang, C., Bannuru, R., Ramel, J., Kupelnick, B., Scott, T. & Schmid, C. H. (2010). Tai Chi on psychological well-being: systematic review and meta-analysis. BMC
Complementary and Alternative Medicine, 10:23.
Wayne, P. M., Walsh J. N., Taylor-Piliae R. E., Wells R. E., Papp K. V., Donovan N. J. & Yeh G. Y. (2014). Effect of Tai Chi on Cognitive Performance in Older Adults: Systematic Review and Meta-Analysis. Journal of the American Geriatrics Society, 62, 25-39.
Zahn D., Adams J., Krohn J., Wenzel M., Mann C. G., Gomille L. K., Jacobi-Scherbening V. & Kubiak T. (2016). Heart rate variability and self-control—A meta-analysis.
Biological Psychology, 115, 9-26.
Zeidan F., Johnson S. K., Diamond B. J., David Z. & Goolkasian P. (2010). Mindfulness meditation improves cognition: Evidence of brief mental training. Consciousness and Cognition, 19, 597-605.
