Effects of mindfulness breathing meditation on stress and cognitive functions: a heart rate variability and eye-tracking study

Abstract

Abstract

Mindfulness meditation has been suggested to protect mental health and enhance cognitive functions. However, the effectiveness of mindfulness breathing meditation for the general population in Malaysia has not been well documented. This study investigated the effects of mindfulness breathing meditation on perceived stress levels and cognitive functions (i.e., visual attention, inhibitory control, and cognitive flexibility) using heart rate variability (HRV) and eye-tracking methods. Fifty Malaysian adults aged between 18 and 34 years participated in this study. Participants were randomly allocated to either mindfulness breathing meditation or active control (music therapy) groups. Participants underwent a four-week intervention consisting of 30 min of online classes once a week and 10 min of daily meditation self-practice or music listening. The results showed significant improvement in cognitive flexibility and reduction in perceived stress levels after the intervention in the mindfulness breathing meditation group compared to the active control group. Participants in the mindfulness breathing meditation group self-reported this intervention as highly acceptable and effective in promoting stress reduction, emotional regulation, and attentional control. However, no significant improvement was found in attention, inhibitory control, and HRV after the intervention, suggesting that the mindfulness breathing meditation had no effect on these variables. While short-term engagement with mindfulness breathing meditation practice is achievable, the three-month follow-up using the Mindfulness Adherence Questionnaire showed that sustaining long-term adherence remains a challenge.

Protocol registration

The approved Stage 1 protocol can be found under “Files”, named “SR_HLY_Manuscript_OSF”, via this link: 10.17605/OSF.IO/P67EG.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-23727-z.

Introduction

In recent years, there has been an increase in research on the effect of mindfulness meditation-based intervention on mental health^1–7. Meditation is an umbrella term that involves a variety of mental practices^8,9. Meditation can be used to define the state of consciousness or as the techniques used in mental training^10–12. Meditation is easy to learn and can be learned by most people regardless of age, gender, ethnicity, educational level, and medical or psychological conditions^13–18. Mindfulness can be defined as a state of being conscious and aware of the present moment¹⁹. It also implies attention, awareness, and remembering to bring back our attention to the present in a receptive way²⁰. From a psychological point of view, mindfulness is a cognitive process that brings awareness of one’s inner state, surroundings, and openness to new experiences or information^21,22. However, other researchers stated that words could not easily describe mindfulness because it is an abstruse, subtle, and non-verbal experience²³. The word “Pranayama” has the meaning of breath regulation practice, with “prana” meaning breath or vital energy and “yama” meaning control^24,25. Mindfulness breathing meditation can be more accessible to beginners as it provides a clear and easily accessible object of attention, i.e., breath, that may be easier for some individuals to establish a meditation practice with a single and straightforward focus. Previous studies have reported the beneficial effects of different types of mindfulness breathing meditation, i.e., 4-4-4-4 breathing meditation, Kapalabhati Pranayama (skull shining breath), Bhastrika Pranayama (bellows breath), Nadishodhan Pranayama (alternate nostril breath), and Bhramari Pranayama (bee breath) on promoting mental health and enhancing cognitive functions^25–29. Although a recent study examined the effects of combining meditation techniques, i.e., Mahapran dhvani (buzzing bee sound meditation) and leśyā dhyāna (colour meditation) on attention³⁰; the effects of combining these five types of mindfulness breathing meditation on perceived stress level and cognitive functions have yet to be studied. A combination of these mindfulness breathing meditation techniques could potentially offer a more comprehensive set of benefits compared to just single technique (see the outline of past studies’ mindfulness meditation intervention, which includes a variety of meditation techniques^31,32. This approach would also enhance the intervention appeal, increase participant engagement, and reduce the risk of boredom, particularly among novice participants. The similarities and differences between the five mindfulness breathing meditation techniques are detailed in Table 1.

Table 1.

Similarities and differences between the five mindfulness breathing meditation techniques.

Types of meditation	Types of breathing	Involvement of body gesture	Sitting posture	Sound
4-4-4-4 breathing meditation	Deep breathing	–	Any meditative posture	–
Kapalabhati pranayama	Rapid breathing	Involves the contraction and relaxation on belly-button area	Any meditative posture	Hissing sound
Bhastrika pranayama	Rapid breathing	Involves arm movement	Vajrasana pose	Hissing sound
Nadishodhan pranayama	Deep breathing	Involves hand gestures	Any meditative posture	-
Bhramari pranayama	Deep breathing	Involves hand gestures	Any meditative posture	Humming sound

Vajrasana pose is a kneeling yoga posture which is commonly used for meditation and mindfulness breathing meditation practices. To perform the Vajrasana pose, place the knees on the floor with knees close together, sit back on the heels, keep the spine straight, and rest the hands on the thighs.

In Malaysia, although many studies^33–44 investigated the effect of mindfulness-based interventions on psychological and cognitive functioning, only a few studies focused on the effects of mindfulness breathing meditation^45–47. For example, a five-minute mindful breathing was useful in distress reduction⁴⁵, but not in pain reduction⁴⁶ in the palliative care setting; whereas a longer session of 30-minute mindful breathing was found effective in fatigue reduction among cancer patients⁴⁷. However, the effectiveness of mindfulness breathing meditation for the general population in Malaysia has not been well-documented. While previous studies have explored the effect of mindfulness meditation on individual aspects such as stress and cognitive functions, the current study proposes to explore these aspects (i.e., HRV, perceived stress levels, visual attention, inhibitory control, and cognitive flexibility) simultaneously using mindfulness breathing meditation intervention and the eye tracking method. Such a holistic approach may offer a comprehensive understanding of the mechanisms for cognitive enhancement.

One of the most highlighted beneficial effects of mindfulness meditation is to reduce stress levels^48,49. Previous studies found that more frequent mindfulness and/or meditation practices were associated with lower perceived stress levels^50,51. Another study suggested that this relationship may be mediated by emotional intelligence⁵², as it is a good predictor for stress^53,54. A moderate-to-high level of stress could decrease the volume of the hippocampus (functions in regulating learning and memory) and prefrontal cortex (functions in regulating actions, thoughts, and emotions) but increase the volume of the amygdala (functions in mediating emotional learning and behaviour)⁵⁵. In contrast, mindfulness-based studies mitigated perceived stress levels, and this was associated with an increase in grey matter density in the hippocampus⁵⁶ and a reduction of grey matter density in the amygdala⁵⁷. Thus, these changes in brain mechanisms support that mindfulness meditation might be a potential intervention for reducing stress levels⁵⁸. As mindfulness meditation training facilitates individuals to mindfully cope with obstacles or even daily life experiences, this may likely improve resilience and adaptation to stress.

The 4-4-4-4 breathing meditation showed beneficial impacts on stress relief²⁶, emotion regulation, reduction in anxiety, depression, post-traumatic stress disorder⁵⁹, improves cardio-inhibitory response of the vagus nerve, activates the parasympathetic system, and regulates the autonomic nervous system⁶⁰. Moreover, a single 18-minute mindfulness breathing meditation intervention using Kapalabhati Pranayama could improve the accuracy rate in the six-letter cancellation task⁶¹. Meanwhile, another study found no improvement in the visual attention aspect but with an increase in the error rate after a three-month Kapalabhati Pranayama intervention using digit letter substitution and six-letter cancellation tasks⁶². However, the absence of a control group in this study limits the ability to draw definitive conclusions. Bhastrika Pranayama was found to have beneficial impacts on perceived stress, depression, anxiety, self-regulation, and emotion regulation^63–65. A functional magnetic resonance imaging study²⁵ found that a four-week Bhastrika Pranayama intervention led to significantly reduced anxiety and negative emotions. These improvements were associated with increased connectivity between the right anterior insula and the ventrolateral prefrontal cortex, two regions that critically involved in emotional regulation, interoceptive awareness, and top-down control of affective responses. Additionally, Nadishodhan Pranayama was effective in protecting mental health²⁷ and relieving stress^66–68. Stress can raise the acidic level in the body (e.g., by increasing the acid production in the stomach and causing acid reflux), whereby deep breathing through alternate nostrils (Nadishodhan) can reduce stress and inflammation by reducing the acidic content in our blood^69–71. Bhramari Pranayama was also effective in reducing stress^28,72–74 and enhancing inhibitory control^72,75, visual and auditory processing with improved reaction time after a six-month intervention⁷⁶.

Recent studies showed that mindfulness meditation interventions could improve HRV among the healthy population^77,78 and individuals with mental or physical illness^31,79. HRV measures the intervals between each heartbeat, and it is also a marker of the autonomic nervous system, which functions in regulating the involuntary physiological processes (e.g., heart rate, respiration, and blood pressure) with two main branches, i.e., parasympathetic (rest-and-digest) and sympathetic (fight-or-flight) nervous systems^80,81. In general, having a higher HRV (e.g., between 50 and 100 milliseconds) is better than having a lower HRV (e.g., below 50 milliseconds), as a higher HRV suggests that the autonomic nervous system is more resilient and adaptable in responding to various stressors effectively⁸². Neuroimaging studies showed that HRV may link to cortical regions such as the ventromedial prefrontal cortex, which involves stressful event appraisal^83,84. Based on neurological evidence, HRV is likely to be impacted by stress; thus, it is often used as an objective measurement of stress⁸⁵. Mindfulness-based stress reduction program was found to improve HRV at the post-intervention through the measurement of standard deviation of all normal-to-normal intervals (SDNN)^31,79, which is the antistress indicator. However, a recent meta-analysis⁸⁶ reported inconclusive findings due to the low methodological quality, and some other studies found no association between HRV and mindfulness meditation intervention^87–90.

Experienced mindfulness meditators have better saccadic eye movement control compared to less experienced meditators^91,92. The underlying mechanism could be that mindfulness meditation training focuses on the breath and the present moment; therefore, it strengthens the ability to direct our attention consciously. This enhanced attentional control may indirectly link to the improvement of saccades, enabling us to shift our gaze more precisely. Eye movements serve as a significant indicator of human interaction with the environment, revealing how people process extensive information from the surroundings⁹³, and provide valuable insights into cognitive functions⁹⁴. During mindfulness breathing meditation, the eyes can be kept open or closed, depending on the individual’s preference; however, the focus usually redirects the attention away from external stimuli and turns the attention inward or towards the inner self. A past study found that the meditation and control groups did not differ in the saccadic eye movement control using pro- and antisaccades tasks after a 7-day intervention⁹⁵, or mixed results⁹¹ when comparing meditators to non-meditators, while no differences were observed in the prosaccade task, the meditators showed lower intra-individual variability in measures such as velocity gain and spatial error (attention and visuospatial processing) during the antisaccade task. Possibly, the short-term intervention⁹¹ (i.e., one week) was insufficient to induce measurable effects. Additionally, tasks requiring high levels of inhibitory control may not effectively detect the early-stage benefits of mindfulness meditation⁹⁶.

With most empirical studies reporting either positive or null effects of mindfulness meditation intervention, a growing body of evidence asserts likely detrimental outcomes (e.g., unpleasant⁹⁷, unwanted⁹⁸, and unintended consequences⁹⁹ with some common adverse experiences such as affective difficulties, delusions, and derealization due to meditation practice¹⁰⁰. Studies also reported the estimated prevalence of adverse experiences of meditation practices, as follows: 3.7%, 22.2%, and 25.0% from experimental, observational, and survey studies, respectively^97,98,101. Recent research in the United States¹⁰² found that 71% of the participants reported lifetime exposure to mindfulness meditation, and 50% of the participants reported at least one adverse experience as measured by the 11-item Meditation-Related Adverse Effects Scale¹⁰³, such as anxiety, emotional instability, and difficulty in sleeping, highlighting the importance of monitoring for potential adverse experience in meditation-based interventions. Additionally, an international cross-sectional study¹⁰⁴ demonstrated that undesirable meditation-related events were predominant among meditators, and these adverse experiences were severe enough to warrant further scientific inquiry. A previous study showed that only a limited clinical guideline concerning meditation-related risks existed¹⁰⁵. While those are in their infancy, only a few organisations and regulatory agencies, such as the American Psychiatric Association^106–108 and the National Institutes of Health¹⁰⁹, had issued statements regarding the reporting of meditation-related adverse effects. Therefore, the current study measured the potential adverse experiences of intervention using the Negative Effects Questionnaire (NEQ).

The current study aims to compare the HRV, perceived stress levels, and cognitive functions (i.e., visual attention, inhibitory control, and cognitive flexibility) between the mindfulness breathing meditation and active control groups with a four-week intervention. It was hypothesized that the mindfulness breathing meditation group would show (1) higher HRV, (2) reduction in self-reported perceived stress level, and (3) higher accuracy rate and faster reaction time on cognitive functions, using antisaccade task (inhibitory control), mix-saccade task (cognitive flexibility) and Flanker Task (visual attention and inhibitory control) as compared to the active control group at the post-intervention.

Methods

Deviations from preregistration

We deviated from preregistration in two aspects. Firstly, three participants were excluded from the saccadic task analysis due to technical difficulties in detecting pupil and corneal reflections. Secondly, although the preregistered design described a within-subjects factor with two levels (pre- and post-intervention) for HRV, but in the actual implementation for HRV, data were collected at four time points (Pre 1, Pre 2, Post 1, and Post 2), with Pre 1 and Post 1 having HRV measured before the cognitive tasks, whereas Pre 2 and Post 2 having HRV measured after the cognitive tasks. By measuring HRV before and after the cognitive tasks, such a change in design could capture both baseline and task-related changes more accurately, by distinguishing whether changes are attributable to the intervention itself or to transient fluctuations induced by task engagement. Therefore, the four-timepoint design approach would provide a more accurate representation of the intervention’s effects by capturing both resting autonomic function and dynamic responses to cognitive demands, which may otherwise be ignored in a simpler two-timepoint design.

Participants

Eighty-four participants (57 females, 25 males, and 2 gender preferred not to say) between the ages of 18 and 48 (M_age = 22.38, SD = 4.54) signed up. Thirty-one participants were excluded during pre-screen because they were not Malaysian. Three participants dropped out during the post-intervention task (one from the mindfulness breathing meditation group and two from the active control group) due to travelling. Thus, leaving a final sample size of 50 participants, aged range from 18 to 34 (M_age = 21.90, SD = 2.96) with mindfulness breathing meditation (n = 25, M_age = 22.64, SD = 3.74) and active control (n = 25, M_age = 21.16, SD = 1.57) groups. See Table 2 for the participants’ demographic characteristics.

Table 2.

Demographic characteristics (n = 50).

Characteristics	Mindfulness breathing meditation		Active control		All
	n	%	n	%	n	%
Gender
Male	7	28	8	32	15	30
Female	18	72	17	68	35	70
Race
Chinese	16	64	16	64	32	64
Malay	7	28	3	12	10	20
Indian	2	8	6	24	8	16
Religion
Buddhism	12	48	11	44	23	46
Islam	6	24	3	12	9	18
Hinduism	1	4	4	16	5	10
Christianity	2	8	4	16	6	12
No religion	4	16	3	12	7	14
Meditation experience
No experience	15	60	21	84	36	72
Have experience but no current practice	5	20	3	12	8	16
Have experience but irregular practice	3	12	1	4	4	8
Have experience but somewhat regular practice	1	4	0	0	1	2
Have experience and regular practice	1	4	0	0	1	2

All participants were recruited using convenient sampling, through the SONA research panel (a research participation recruitment system for psychology students), poster, email, social media platform, and word-of-mouth. A minimum sample size of 28 participants was required, based on the power analysis using the G*Power 3.1.9.7 software¹¹⁰, with an effect size (f) of.25^1,111, an alpha value of 0.05, and a power of 0.80. To ensure randomisation, this study utilised a stratified random allocation method. We expected a 16.25% to 29.2% attrition rate^112,113, as lack of discipline, effort, engagement, and intrinsic motivation might play a role¹¹⁴. As such, seven more participants were needed.

Ethics information

This project obtained ethical approval from the Faculty of Science Ethics Committee at the University of Nottingham Malaysia (Ethics application number: HLY180523). All methods were performed in accordance with the relevant guidelines and regulations. Participation in this experiment was entirely voluntary, and the participants were free to withdraw at any time without giving any reason. The informed consent of the participants was received electronically. All participants who fully completed this research were compensated with four course credits (for the University of Nottingham Malaysia psychology students) or a chance to win a lucky draw of RM50 cash vouchers.

Design

A randomised controlled trial with a pre-and post-test mixed design and with a four-week intervention was conducted. There were two conditions, i.e., a mindfulness breathing meditation and an active control (music therapy) group. Participants’ HRV and cognitive performance were measured, and the perceived stress levels, acceptance towards the intervention, quantity and quality of meditation practices, and adverse experiences of the intervention were self-reported by the participants. Data collection and analysis were carried out in a blinded manner regarding the experimental conditions, to reduce bias and ensure that the experimenter was unaware of participants’ group allocation. Participants were pre-assigned to their respective groups by an independent researcher not involved in data collection. Participants were assigned an ID to remain anonymous throughout the experiment and during data analysis. This blinding was designed to prevent differential treatment of participants and reduce the risk of expectancy or confirmation biases affecting experimental outcomes.

Measures

Perceived stress level

The 10-item Perceived Stress Scale (PSS-10)¹¹⁵ was used to measure the perception of stress. Participants were required to respond based on the presence of the feeling during the previous month and indicate their level of agreement for each item using a range from 0 (Never) to 4 (Very Often). See Appendix 3 for the questionnaire. Four out of ten items were scored reversely (Q4, Q5, Q7, and Q8). The total scores were calculated. The lowest score was zero, while the highest score was 40. Scores ranging from zero to 13, 14 to 26, and 27 to 40 were considered low, moderate, and high perceived stress levels, respectively. PSS-10 had good reliability with Cronbach’s alpha of 0.87¹¹⁶, and adequate convergent validity for stressful life events (number: r =.13, p <.001; impact: r =.23, p <.001)¹¹⁷.

Heart rate variability

The HRV was measured during the participants’ resting state for five minutes (before and after the cognitive tasks at pre- and post-intervention), using the Brain Vision electrocardiogram (https://www.brainproducts.com/solutions/bip2aux/). It was recorded from three electrodes, with one placed on the participant’s left earlobe and the other two placed on the participant’s left and right wrists. The heart rate signal was band-pass filtered (8–20 Hz), to reduce the impact of high-frequency noise and the baseline fluctuation of the cardiac signal¹¹⁸. The automatic detection of cardiac beats was conducted using Brainstorm¹¹⁹, followed by manual correction of potentially missing peaks or erroneous.

SDNN is one of the time-domain indices of HRV, which quantifies the amount of variability in measurements of the interbeat interval measured in milliseconds (ms). The conventional standard short-term recording for SDNN is five minutes^120,121. SDNN reflects the overall autonomic nervous system activity, with higher values indicating a greater variability and adaptability of the cardiovascular system. SDNN is contributed by both sympathetic and parasympathetic nervous system activities; thus, making it a valuable indicator of cardiac autonomic regulation. The normal range for SDNN value could be from 32ms to 93ms based on 27 past studies⁹⁵.

Saccadic paradigm

In the prosaccade tasks, participants were required to look as fast and exact as possible at a suddenly appearing peripheral saccade stimulus, (i.e., a green square, rgb = [0, 128, 0]) on a white background, whereas in the antisaccade task, participants were required to look away from the saccade stimulus (i.e., a red square, rgb = [255, 0, 0]) to the opposite direction as fast as possible. The mix-saccade task involved both pro- and antisaccades in a randomized order, and participants were required to respond according to the colour of the square. There were three saccade blocks in total, with prosaccades presented as the first block, followed by antisaccades and mix-saccades in this order. Each block was presented with four practice trials and followed by 40 actual trials. There were no practical trials for mix-saccades. The entire saccadic task took approximately eight minutes to complete.

At the beginning of a trial, a fixation cross, with a size of 3 × 3 mm, appeared in the centre of the screen for either 700 ms, 900 ms, 1100 ms, 1300 ms, or 1500 ms before the onset of a saccade stimulus. These various fixation intervals were used to avoid the predictability of the upcoming stimuli. The saccade stimulus (i.e., a square, with a size of 10 × 10 mm) was presented either to the left or right 85 mm horizontally from the centre of the screen for 1000ms. See Fig. 1 for an illustration of the saccadic task and Fig. 2 for prosaccades, antisaccades, and mix-saccades timelines.

Fig. 1.

Illustration of saccadic task.

Fig. 2.

Timeline for pro, anti, and mix-saccade tasks.

The saccadic paradigm was programmed using Matlab (R2023a), and the participants’ saccadic eye movements were tracked and recorded using SR Research EyeLink portable duo eye tracker. A binocular tracking was used to detect the position of both eyes with a 1000 Hz sampling rate, 22° per second velocity threshold, and 75% illumination level. The nine-point calibration type was performed to ensure and validate that the tracking accuracy is within 1° visual angle¹²². The stimuli were presented on a 21” monitor. Participants placed their heads on a chinrest to ensure a constant viewing distance of 75 cm. The accuracy of saccades and saccade latency were measured as saccadic performance.

Visual attention and inhibitory control

The Flanker Task was used to measure selective attention and inhibitory function¹²³. There were two types of trials in this task, i.e., congruent and incongruent. For congruent trials, a target arrow was flanked at each side by two non-target arrows pointing in the same direction. For incongruent trials, the target arrow was flanked by non-target arrows, pointing in the opposite direction. Participants were shown five arrows in each trial and required to determine the direction of the central arrow. If the central arrow was pointing to the left, the left shift key should be pressed, whereas if the central arrow was pointing to the right, the right shift key should be pressed. The participants were instructed to ignore the surrounding arrows and respond as fast and accurately as possible to the central arrow. There were eight practice trials (feedback and rate of reaction time were provided), followed by 80 actual trials (no feedback was given). The entire Flanker Task took approximately three minutes to complete.

Acceptability towards mindfulness breathing meditation intervention

A modified open-ended acceptability questionnaire¹²⁴ was used to measure the participants’ acceptance of the intervention. The modified version measured four aspects: (1) impact of the intervention on the study, work, and/or home life, (2) the perceived difficulty level of completing the intervention with justification, (3) an overall usefulness rating (1 = not useful to 10 = very useful), and (4) rationale for the rating. Participants in the mindfulness breathing meditation group answered this questionnaire during the post-intervention.

Quantity and quality of meditation practices, and three-month follow-up assessment

The Mindfulness Adherence Questionnaire (MAQ) was used to measure the participants’ quantity and quality of meditation practices¹²⁵. MAQ is a self-report questionnaire with 12 items. The first two items measured the quantity (frequency and average duration) of participants’ formal practices. The other 10 items measured the quality of the participants’ formal and informal practices. Participants were required to indicate their level of agreement for each item using a range from 0 (Never) to 6 (Always). A higher score indicated a higher level of meditation practice quality. MAQ has good reliability and validity with Cronbach’s alpha values of 0.67 to 0.87 for the formal subscale and 0.91 to 0.93 for the informal subscale¹²⁵. The mindfulness breathing meditation participants answered the MAQ once a week for four weeks throughout the intervention and three months after the post-intervention as a follow-up assessment.

Adverse events questionnaire

The 20-item Negative Effects Questionnaire (NEQ) was used to measure the adverse effects of the intervention^126,127. This self-report questionnaire was administered to monitor and report intervention’s incidences and impacts. It consisted of three parts, (1) a close-ended question asking the participants if they experienced the stated incidents and effects, (2) self-rating scales where participants were required to indicate their level of agreement using a range from 0 (Not at all) to 4 (Extremely), for how negatively the incidents and effects affected them, and (3) close-ended question asking participants to attribute the negative effect to either “intervention” or “other circumstances”. At the end of the questionnaire, there was an open-ended question for participants to describe in their own words whether there were any other negative incidents or effects, and what characterized them. Items 9, 11, 14, 15, 16, 17, and 19 were removed due to being irrelevant to this intervention (e.g., “I became afraid that other people would find out about my treatment”); thus, the total number of items used in this study was 13. NEQ had excellent internal consistency with Cronbach’s alpha values of.95¹²². Participants answered this questionnaire during the post-intervention.

Procedure

Interested participants filled in a screening questionnaire, and those who met the inclusion criteria (i.e., Malaysian, aged between 18 and 55 years old, non-current and regular mindfulness breathing meditation practitioner, and without any diagnosed physical and mental disorders) were included in this study. After informed consent, participants were randomly assigned to either a mindfulness breathing meditation or an active control group. A briefing section was arranged for both groups separately to briefly explain the procedures of this study. The pre- and post-measurements were conducted in a laboratory setting, and the intervention was conducted online. During the pre-test task, participants’ HRV were measured for five minutes during the resting state, they answered the demographic information and PSS-10, followed by the cognitive tasks (i.e., Pro, Anti, and Mix-saccade tasks and Flanker Task), and then their HRV were measured again for five minutes. The four-week intervention was carried out after the pre-test.

All participants in the mindfulness breathing meditation group underwent a four-week intervention in a group format with approximately half an hour of online meditation class per week via Microsoft Teams and 10 min of daily self-practice (link: https://youtu.be/iS4czj88A0M) at their home. Each class consisted of a combination of five types of mindfulness breathing meditation techniques (i.e., 4-4-4-4 breathing meditation, Kapalabhati Pranayama, Bhastrika Pranayama, Nadishodhan Pranayama, and Bhramari Pranayama). The participants were required to fill in the MAQ once a week after each meditation class. This intervention was designed and delivered by a Pranayama instructor with more than ten years of meditation experience (Credential ID: #f9472f915af84499).

All participants in the active control group underwent a four-week intervention in a group format with approximately half an hour of online music therapy per week via Microsoft Teams and 10 min of daily music listening at their homes. This intervention was structured to match the music therapy part of the Health Education Program¹²⁸. This intervention was delivered by a music expert with more than 18 years of experience. The choice of music was mainly happy or calm classical music, and it was controlled for the tempo, valence, and arousal^128,129. A weekly reminder and log sheets were sent to the participants to ensure they performed their tasks accordingly.

The post-test task was the same as the pre-test task for both groups, except for answering the demographic information. In addition, participants in the mindfulness breathing meditation group were required to answer the NEQ and acceptability questionnaire. After that, all participants were debriefed and thanked for their participation. The mindfulness breathing meditation participants were required to fill in the MAQ as the follow-up questionnaire after three months from the post-intervention. See Fig. 3 for the procedural flow chart.

Fig. 3.

Procedural Flow Chart.

Data analysis

To analyse the main effect and interaction of mindfulness breathing meditation and active control groups and to determine if there were significant differences between the two groups, the 2 (stage: pre- and post-intervention, as the within-subjects factor) × 2 (group: mindfulness breathing meditation versus active control, as the between-subjects factor) mixed ANOVAs on perceived stress level, visual attention, inhibitory control, and cognitive flexibility, and 4 (stage: Pre 1, Pre 2, Post 1, Post 2) × 2 (group: mindfulness breathing meditation versus active control, as the between-subjects factor) mixed ANOVA on HRV were conducted. For additional analyses, separate one-way ANOVAs were performed to evaluate the (1) frequency and (2) average duration of a mindfulness breathing meditation practice session; and a 5 (week: week one to four and week 13, i.e., three-month follow-up) × 2 (formality: formal and informal practices) mixed ANOVA was performed to evaluate the quality of mindfulness breathing meditation practices, as measured on the MAQ across the intervention. Bonferroni corrections were applied to adjust for multiple pairwise comparisons. To be able to interpret the null results, Bayesian analysis was conducted using the statistical software JASP (0.19.2)¹³⁰.

The HRV was indexed by SDNN. The pro-, anti-, and mix-saccadic performances were indexed by (1) average accuracy of left and right eyes [i.e., (total number of accurate trials/total number of trials) × 100%] and (2) average saccade latency of left and right eyes (automatically recorded, in ms). The Flanker Task performances were indexed by (1) accuracy rate [i.e., (total number of correct trials/total number of trials) × 100%], (2) reaction time (automatically recorded, in ms), and (3) Flanker Effect, i.e., reaction time difference, in which incongruent minus congruent trials (automatically recorded, in ms). The outliers, missing values, and dropouts were discarded during data analysis. All data for saccadic tasks with invalid trials (e.g., invalid latency or latency shorter than 80ms¹³¹ were removed as outliers.

One participant from the mindfulness breathing meditation group and two participants from the active control group were excluded from the saccadic task analysis due to technical difficulties in detecting pupil and corneal reflections. For mindfulness adherence, one participant did not fill out the MAQ in week 1, and three participants did not fill out the MAQ in week 3. For the post-intervention questionnaire, one participant from the mindfulness breathing meditation group did not fill out the PSS-10, MAQ, NEQ, and acceptability questionnaire. Lastly, three participants did not fill out the MAQ during the three-month follow-up. These participants were excluded from the respective analysis due to missing values.

Results

Preregistered analyses

The descriptive statistics for HRV, PSS, Saccadic, and Flanker Tasks were presented in Table 3.

Table 3.

Descriptive statistics for HRV, PSS, saccadic and flanker tasks.

Tasks	Mindfulness breathing meditation		Active control
	M	SD	M	SD
HRV
SDNN (ms)
Pre 1	59.09	18.51	53.67	20.01
Pre 2	61.64	28.45	51.44	15.29
Post 1	58.39	23.18	46.22	18.32
Post 2	75.41	90.85	47.32	13.10
PSS
Pre-intervention	21.25	6.16	20.76	7.37
Post-intervention	12.25	6.06	19.04	7.25
Prosaccade task
Pre-intervention
Accuraccy rate (%)	98.05	2.21	97.84	4.43
Reaction time (ms)	177.17	21.22	187.85	34.31
Post-intervention
Accuraccy rate (%)	98.13	3.42	98.13	3.42
Reaction time (ms)	171.50	32.06	174.27	30.45
Antisaccade task
Pre-intervention
Accuraccy rate (%)	78.71	17.90	82.22	12.25
Reaction time (ms)	239.65	42.47	265.62	57.64
Post-intervention
Accuraccy rate (%)	87.27	12.48	84.40	12.35
Reaction time (ms)	232.75	51.11	242.13	37.02
Mix-saccade task
Pre-intervention
Accuraccy rate (%)	74.21	12.57	75.10	10.11
Reaction time (ms)	243.55	56.79	264.03	57.03
Post-intervention
Accuraccy rate (%)	82.15	12.05	73.06	9.61
Reaction time (ms)	223.86	53.29	243.06	31.04
Flanker task
Pre-intervention
Accuracy rate (%)	96.00	5.25	89.30	22.66
Reaction time (ms)	418.53	45.94	436.14	120.29
Flanker effect (ms)	41.68	17.61	57.25	100.78
Post-intervention
Accuraccy rate (%)	94.80	4.56	89.30	11.51
Reaction time (ms)	391.86	50.41	454.09	98.82
Flanker effect (ms)	31.92	18.67	35.60	99.49

Heart rate variability

The HRV was measured four times (stage), i.e., Pre 1 (during the pre-intervention, before the cognitive tasks), Pre 2 (during the pre-intervention, after the cognitive tasks), Post 1 (during the post-intervention, before the cognitive tasks), and Post 2 (during the post-intervention, after the cognitive tasks).

The main effect of group was significant, F(1, 43) = 5.22, p =.03, η_p² = 0.11, BF₁₀ = 1.17, with the mindfulness breathing meditation group having a higher SDNN value (M = 63.63, SE = 4.18) compared to the active control group (M = 49.66, SE = 4.47). However, the main effect of stage was not significant, F(3, 129) = 0.44, p =.72, η_p² = 0.01, BF₁₀ = 0.05, with Pre 1 (M = 56.38, SE = 2.87), Pre 2 (M = 56.54, SE = 3.48), Post 1 (M = 52.30, SE = 3.15), and Post 2 (M = 61.37, SE = 10.02). There was no significant interaction between stage and group, F(3, 129) = 0.78, p =.51, ηp² = 0.02, BF₁₀ = 0.15. Bonferroni corrected multiple comparisons did not reveal any significant differences of the groups on HRV, with all p >.0042 (0.05/12).

Perceived stress level

The main effect of stage was significant, F(1, 47) = 19.90, p <.001, η_p² = 0.30, BF₁₀ = 400.27. There was a higher perceived stress level in the pre-intervention (M = 21.01, SE = 0.97) compared to the post-intervention (M = 15.65, SE = 0.96). The main effect of group was also significant, F(1, 47) = 4.35, p =.04, η_p² = 0.09, BF₁₀ = 1.27. There was a lower perceived stress level in the mindfulness breathing meditation group (M = 16.75, SE = 1.08) compared to the active control group (M = 19.90, SE = 1.06). There was a significant interaction between stage and group, F(1, 47) = 9.18, p =.004, ηp² = 0.16, BF₁₀ = 13.89. The perceived stress level in the mindfulness breathing meditation group reduced significantly at the post-intervention (M_diff = 9.00, SE = 1.72, p <.001) but not in the active control group (M_diff = 1.72, SE = 1.68, p =.31). See Fig. 4 for the bar chart.

Fig. 4.

Mean value of perceived stress level for pre- and post-intervention in mindfulness breathing meditation and active control groups. Error bars represent the standard error of the mean. *** p <.001.

Saccadic task

For prosaccades’ accuracy rates, the main effect of stage was not significant, F(1, 45) = 0.19, p =.66, η_p² = 0.004, BF₁₀ = 0.24, with pre-intervention (M = 97.95, SE = 0.51) and post-intervention (M = 98.18, SE = 0.41). The main effect of group was also not significant, F(1, 45) = 0.01, p =.94, η_p² = 0.000, BF₁₀ = 0.32, as the accuracy rate for mindfulness breathing meditation (M = 98.09, SE = 0.53) and active control (M = 98.04, SE = 0.54) were about the same. There was no significant interaction between stage and group, F(1, 45) = 0.08, p =.78, η_p² = 0.002, BF₁₀ = 0.28.

For prosaccades’ reaction times, the main effect of stage was not significant, F(1, 45) = 3.21, p =.08, η_p² = 0.07, BF₁₀ = 0.95, with pre-intervention (M = 182.51, SE = 4.14) and post-intervention (M = 172.88, SE = 4.57). The main effect of group was also not significant, F(1, 45) = 0.96, p =.33, η_p² = 0.02, BF₁₀ = 0.41, with mindfulness breathing meditation (M = 174.33, SE = 4.80) and active control (M = 181.06, SE = 4.90). There was no significant interaction between stage and group, F(1, 45) = 0.54, p =.47, η_p² = 0.01, BF₁₀ = 0.36.

For antisaccades’ accuracy rates, the main effect of stage was significant, F(1, 45) = 5.88, p =.02, η_p² = 0.12, BF₁₀ = 2.94, with pre-intervention (M = 80.46, SE = 2.25) and post-intervention (M = 85.83, SE = 1.81). The main effect of group was not significant, F(1, 45) = 0.01, p =.93, η_p² = 0.000, BF₁₀ = 0.32, as the accuracy rate for mindfulness breathing meditation (M = 82.99, SE = 2.40) and active control (M = 83.31, SE = 2.45) were about the same. There was no significant interaction between stage and group, F(1, 45) = 2.07, p =.16, η_p² = 0.04, BF₁₀ = 0.67.

For antisaccades’ reaction times, the main effect of stage was not significant, F(1, 45) = 3.55, p =.07, η_p² = 0.07, BF₁₀ = 1.06, with pre-intervention (M = 252.63, SE = 7.36) and post-intervention (M = 237.44, SE = 6.53). The main effect of group was also not significant, F(1, 45) = 2.43, p =.13, η_p² = 0.05, BF₁₀ = 0.77, with mindfulness breathing meditation (M = 236.20, SE = 7.94) and active control (M = 253.87, SE = 8.11). There was no significant interaction between stage and group, F(1, 45) = 1.06, p =.31, η_p² = 0.02, BF₁₀ = 0.44.

For mix-saccades’ accuracy rates, the main effect of stage was not significant, F(1, 45) = 2.37, p =.13, η_p² = 0.05, BF₁₀ = 0.60, with pre-intervention (M = 74.65, SE = 1.67) and post-intervention (M = 77.61, SE = 1.59). The main effect of group was also not significant, F(1, 45) = 2.41, p =.13, η_p² = 0.05, BF₁₀ = 0.71, with mindfulness breathing meditation (M = 78.18, SE = 1.85) and active control (M = 74.08, SE = 1.89). However, there was a significant interaction between stage and group, F(1, 45) = 6.77, p =.01, η_p² = 0.13, BF₁₀ = 4.97. The accuracy rates in the mindfulness breathing meditation group increased significantly at the post intervention (M_diff = 7.94, SE = 2.68, p =.01) but not in the active control group (M_diff = 2.04, SE = 2.74, p =.46). See Fig. 5 for the bar chart.

Fig. 5.

Mean value of accuracy rate for mix-saccade task during pre- and post-intervention in mindfulness breathing meditation and active control groups. Error bars represent the standard error of the mean. ** p <.01.

For mix-saccades’ reaction times, the main effect of stage was significant, F(1, 45) = 7.27, p =.01, η_p² = 0.14, BF₁₀ = 5.32, with pre-intervention (M = 253.79, SE = 8.30) and post-intervention (M = 233.46, SE = 6.40). The main effect of group was not significant, F(1, 45) = 2.42, p =.13, η_p² = 0.05, BF₁₀ = 0.91, with mindfulness breathing meditation (M = 233.71, SE = 8.93) and active control (M = 253.54, SE = 9.12). There was no significant interaction between stage and group, F(1, 45) = 0.01, p =.93, η_p² = 0.000, BF₁₀ = 0.27.

Flanker task

For accuracy rates, the main effect of group was significant, F(1, 48) = 5.22, p =.03, η_p² = 0.01, BF₁₀ = 1.55. There was a higher accuracy rate in the mindfulness breathing meditation group (M = 95.40, SE = 1.89) compared to the active control group (M = 89.30, SE = 1.89). However, the main effect of stage was not significant, F(1, 48) = 0.05, p =.82, η_p² = 0.001, BF₁₀ = 0.21, with pre-intervention (M = 92.65, SE = 2.33) and post-intervention (M = 92.05, SE = 1.24). There was no significant interaction between stage and group, F(1, 48) = 0.05, p =.82, η_p² = 0.001, BF₁₀ = 0.28.

For reaction times, the main effect of stage was not significant, F(1, 48) = 0.08, p =.77, η_p² = 0.002, BF₁₀ = 0.24, with pre-intervention (M = 427.34, SE = 12.88) and post-intervention (M = 422.98, SE = 13.39). The main effect of group was also not significant, F(1, 48) = 3.46, p =.07, η_p² = 0.07, BF₁₀ = 1.16, with mindfulness breathing meditation (M = 405.19, SE = 15.19) and active control (M = 445.12, SE = 15.19). There was no significant interaction between stage and group, F(1, 48) = 2.18, p =.15, η_p² = 0.04, BF₁₀ = 0.66.

Non-preregistered analyses

Flanker task

For the Flanker Effect, the main effect of stage was not significant, F(1, 48) = 1.84, p =.18, η_p² = 0.04, BF₁₀ = 0.50, with pre-intervention (M = 49.47, SE = 10.23) and post-intervention (M = 33.76, SE = 10.12). The main effect of group was also not significant, F(1, 48) = 0.33, p =.57, η_p² = 0.01, BF₁₀ = 0.33, with mindfulness breathing meditation (M = 36.80, SE = 11.83) and active control (M = 46.42, SE = 11.83). There was no significant interaction between stage and group, F(1, 48) = 0.26, p =.61, η_p² = 0.01, BF₁₀ = 0.32.

Acceptance towards the intervention

In terms of perceived value and intervention impact, the most commonly reported themes included stress reduction, calming and relaxing effects, improved focus, concentration, and attention. 63% of the participants reported that this mindfulness breathing meditation practice had helped in dealing with stress, and they were able to pay more attention and stay focused while studying and doing revision for exams. Participants stated that these beneficial impacts contributed to improved concentration when completing tasks both at work and at home. Other commonly reported themes were improved emotional regulation and sleep quality. Participants reported that this meditation practice improved their mental health and general emotional awareness, for example, managing emotions better and enhancing sleep quality. Seven participants stated that practising mindfulness breathing meditation could clear their minds and help to kickstart the day by feeling productive. With the help of meditation practices, they could switch to calm mode easily when facing challenges or frustrations; thus, improving their problem-solving skill. One participant mentioned that the overall impact of this meditation practice was minimal; however, it provided a sense of structure and facilitated the allocation of time for relaxation.

In terms of difficulty level to complete this mindfulness breathing meditation intervention, 75% of the participants stated that it was easy to complete, whereas 25% of the participants had a mixed opinion. Participants found that this intervention was easy to complete due to several key factors. The guided meditation video offered clear and easy-to-follow steps, ensuring accessibility for participants of varying meditation experience levels. The duration of this meditation practice was also ideal (i.e., 10-minute daily and 30-minute once a week), allowing them to integrate their practice into their daily routine. As breathing is the main component in this intervention, it was not constrained by environment or time; thus, providing flexibility in practising the mindfulness breathing meditation. One participant shared that the perceived benefits of the meditation practice served as an intrinsic motivation, leading to the automatic engagement and consistency in incorporating this mindfulness breathing meditation into their routine. However, some participants encountered several challenges that made this intervention somewhat difficult to complete. Two participants reported that time felt slow during mindfulness breathing meditation, making it harder for them to stay engaged and maintain focus. This could potentially affect the effectiveness of the mindfulness breathing meditation practice.

Participants provided an average overall rating of 84.2% (SD = 1.08) regarding the usefulness of the mindfulness breathing meditation intervention. When asked to justify their rating, participants stated that it was a great experience in learning new ways to meditate as a relaxing activity. Participants mentioned that it helped to manage psychological distress and effectively reduced stress levels, contributing to an overall sense of well-being. The calming effect of the meditation practice allowed them to feel more at ease, and the ability of mindfulness breathing meditation to promote mindfulness and grounding techniques contributed to the emotional regulation. Four participants rated 10 out of 10 for this intervention due to its simplicity and significant positive effects. For example, it was efficient in terms of time commitment and did not require any additional equipment, making it highly accessible and practical for regular practice. See Table 4 for the common themes and sample quotes on the overall acceptability and perceived value of the mindfulness breathing meditation intervention.

Table 4.

Themes and sample quotes from the acceptability questionnaire towards the mindfulness breathing meditation intervention.

Theme	Quotes
Stress reduction	“This meditation practices tremendously helped in dealing with stress.”
	“I feel less stress and nervous before and during presentation.”
	“It reduces my stress level for both study and work.”
Improvement in attention and concentration	“I am able to pay more attention and concentration while studying and doing revision.”
	“It improves my concentration when completing tasks.”
	“Learning the breathing techniques also helped me to focus on something more relaxing.”
Emotional and mental well-being	“It really helps me to feel calmer in stressful situations and overall improved on my mental health.”
	“I feel more relax and calm after practicing the breathing meditation.”
	“It can instantly calm down my anxious feeling by doing for just 10 min only.”
Improvement in sleep quality	“It helped me to sleep better and improved sleep quality, I liked to meditate before bed. I also liked the meditation moves in the video provided. Thus was my first attempt to meditating and it has given me a positive experience.”
	“It helps me to sleep better, relax and reduce stress.”
	“Easy to do and useful to improve sleep quality and stress levels.”
Clearly structured guided video	“The guidance video is very easy to understand, and I could do it at my free time.”
	“The instructions in the guided video were clear and easy to follow. I can simply do it anywhere even without the yoga mat (not much restrictions).”
	“The steps and postures were easy to follow, and the guided video was clearly presented.”

Quantity and quality of meditation practices and three-month follow-up

Descriptive statistics for quantity and quality of mindfulness breathing meditation practices are presented in Table 5.

Table 5.

Descriptive statistics for quantity and quality of mindfulness breathing meditation practices.

Week	N	Quantity of practice				Quality of practice
		Frequency		Duration		Formal		Informal
		M	SD	M	SD	M	SD	M	SD
1	24	4.29	2.27	17.75	12.37	70.66	11.16	62.15	15.92
2	25	5.00	1.91	13.32	12.28	69.00	13.61	67.44	13.57
3	22	5.32	1.64	13.14	10.66	75.95	13.17	76.01	13.68
4	24	5.08	1.44	13.75	12.95	74.48	13.02	72.69	15.51
13	22	1.45	1.57	8.14	7.07	64.39	42.86	68.06	15.40

A one-way ANOVA was performed to evaluate the frequency of the mindfulness breathing meditation practices as measured on the MAQ across the four-week intervention and Week 13 (three-month follow-up), and to check if there was a significant difference between the first and last week of the meditation practices. There was a significant main effect of practice weeks, F(1, 4) = 17.57, p <.001, η_p² = 0.39, BF₁₀ = 1.00. There were significant differences between the frequency of practices in Week 1 to Week 4 compared to Week 13 (all p <.001), with Bonferroni-corrected α of 0.005 (0.05/10), indicating a decline in participants’ adherence to the intervention over time, but no significant differences between Week 1 to Week 4 (all p >.05).

A one-way ANOVA was performed to evaluate the average duration of a mindfulness breathing meditation practice session as measured on the MAQ across the four-week intervention and Week 13. There was no significant main effect of practice duration, F(1, 4) = 2.07, p =.09, η_p² = 0.07, BF₁₀ = 0.53. There were no significant differences between all weeks (all p >.05), after Bonferroni-corrected α of 0.005 (0.05/10).

A 5 (week: week 1 to 4 and week 13) × 2 (formality: formal practices and informal practices) mixed ANOVA was performed to evaluate the quality of mindfulness breathing meditation on formal and informal practices as measured on the MAQ across the four-week intervention and follow-up. There was no significant interaction between week and formality, F(1, 4) = 1.50, p =.21, η_p² = 0.05, BF₁₀ = 0.26. However, there were significant differences on informal practices between Week 1 and Week 3 (p =.002), Week 1 and Week 4 (p =.02), and Week 2 and Week 3 (p =.05). The main effect of formality was not significant, F(1, 4) = 1.02, p =.32, η_p² = 0.01, BF₁₀ = 0.24, with formal practice (M = 70.90, SE = 2.02) and informal practice (M = 69.27, SE = 1.38) scoring about the same. The main effect of week was also not significant, F(1, 4) = 1.65, p =.17, η_p² = 0.06, BF₁₀ = 0.43. Bonferroni-corrected multiple comparisons did not reveal any significant differences between the week and formality of practices (all p >.05).

Adverse effects of meditation intervention

In the NEQ, participants reported a mean of 1.72 (SD = 2.36) negative effects, ranging from zero to nine. The most commonly reported adverse experience was having a problem with sleep (33.3% of the total sample). Of those reporting this adverse experience, only one participant stated that it was related to this intervention. The other commonly reported adverse experiences included stress and unpleasant memories resurfacing (20.8%), anxiety (16.7%), worried and suicidal thoughts (12.5%), unpleasant feelings and helplessness (8.3%), and hopelessness (4.2%). Of those reporting these adverse experiences, none of the participants stated that it was related to the intervention.

Discussion

The current findings did not show a significant interaction effect in HRV between the mindfulness breathing meditation and active control groups. The significant main effect of group, with the mindfulness breathing meditation group having higher SDNN values compared to the active control group at the Post 2 stage, might be attributed to a random effect. The breath-focused nature of mindfulness breathing meditation may facilitate parasympathetic dominance by reducing physiological arousal and supporting the downregulation of stress-related sympathetic activation¹³². However, the absence of significant main effects of stage or a stage-group interaction suggests that the intervention was ineffective. Alternatively, the duration of the intervention may not be long enough to show within-subject changes in SDNN across the pre- and post-intervention. Future studies should explore a longer intervention duration or individualised mindfulness training to better understand the autonomic effects of mindfulness-based meditation practices.

The prosaccade task showed neither significant interaction in accuracy nor reaction times in both the mindfulness breathing meditation and active control groups. This finding is not surprising, as prosaccades are primarily automatic and subcortically driven responses mediated by the brain structures such as the superior colliculus, a layered structure located in the midbrain that processes visual stimuli and guides attentional orientation which requires minimal top-down cognitive control¹³³. Given that mindfulness-based interventions predominantly enhance higher-order cognitive functions rather than automatic sensorimotor pathways¹³⁴, the lack of improvement in the prosaccade performance aligns with prior findings and theoretical expectations^91,122. In contrast, the antisaccade task, which requires inhibitory control to suppress a reflexive saccade and voluntarily look in the opposite direction, showed a significant main effect of stage, with improved accuracy rate from pre- to post-intervention. However, the absence of a significant group effect or interaction indicates that while inhibitory control improved over time, mindfulness breathing meditation did not show more improvement compared to the active control group. This result is consistent with previous studies that short-term meditation interventions did not show significant improvements in inhibitory control as measured using antisaccade tasks^91,122. The most highlighted findings were from the mix-saccade task, which integrates reflexive response and inhibitory control, thereby placing a higher demand on cognitive flexibility and task-switching ability. There was a significant interaction effect observed between stage and group, with the mindfulness breathing meditation group showing a significant improvement in accuracy rates at the post-intervention compared to the active control group. The improvement in mix-saccade task accuracy aligns with previous findings, suggesting that mindful breathing enhances amygdala and prefrontal cortex activity and connectivity, leading to improvements in cognitive flexibility and task-switching abilities^135,136. Given that the mix-saccade task requires rapid switching between reflexive responses and inhibitory control, these results support that mindfulness breathing meditation could enhance higher-order cognitive functions by promoting metacognitive awareness and attentional shifting^137–139. The improvement in reaction times from pre- to post-intervention, although not specific to the mindfulness breathing meditation group, indicates that both groups showed improvement in processing speed efficiency. However, the absence of a significant group effect for reaction time suggests that while the accuracy rate was significantly enhanced in the mindfulness breathing meditation group, the improvement of the speed of processing stimuli was similar in both groups. The observation that the mindfulness breathing meditation intervention only induced an effect on mix-saccade performance but not on pro- and antisaccades suggests that the effect of the mindfulness breathing meditation intervention on cognitive functions may be task-dependent, or specific cognitions may be affected differently, with greater effects observed in the tasks requiring adaptive cognitive control.

For the Flanker Task, no significant interaction effect was found between the mindfulness breathing meditation and active control groups. Although the mindfulness breathing meditation group scored a significantly higher accuracy rate compared to the active control group, it is important to note that accuracy is typically not the most informative metric in the Flanker Task that involves normal healthy participants due to this natural ceiling effect. Instead, the reaction times (especially the Flanker Effect) would provide a more sensitive measure of inhibitory control^123,140. Although both groups scored high accuracy rates, there were no significant changes in the reaction times or Flanker Effect magnitude following the intervention, suggesting that the mindfulness breathing meditation intervention did not meaningfully enhance participants’ efficiency in resolving cognitive interference. The lack of reaction time improvement indicates that while mindfulness meditation practices are theorised to support the top-down attentional control^96,138, a four-week mindfulness breathing meditation intervention may not produce improvement in attention and inhibitory control in these high-functioning and healthy populations.

The current finding showed a significant interaction between the mindfulness breathing meditation and active control groups on perceived stress levels, which aligned with the past findings^141–143. Participants in the mindfulness breathing meditation group reported a significant reduction in the perceived stress levels after the intervention, whereas the active control group experienced only a slight but non-significant decrease in the perceived stress levels. These results suggest that mindfulness breathing meditation may be a more effective stress-reduction strategy than relaxation through music therapy. This aligns with previous research demonstrating that mindfulness-based interventions, particularly those incorporating breath-focused techniques, can enhance emotional regulation and reduce subjective stress perceptions by promoting greater awareness and control over physiological and psychological stress responses^132,144. The finding also showed that relaxation through music therapy alone may not provide the same level of active engagement that mindfulness breathing meditation offers, although previous studies have shown that listening to happy classical music could induce a positive mood and reduce perceived stress levels^145,146. Previous studies have found that active self-regulation strategies, such as mindfulness meditation, tend to have stronger and longer-lasting effects on stress reduction compared to relaxation techniques¹⁴⁷. Given that both mindfulness breathing meditation and music listening are effective for stress reduction, future studies could integrate them, such as using calming or alpha wave music during mindfulness breathing meditation sessions and encouraging breath synchrony with the music’s rhythm to combine mindfulness with the calming effect of music. However, effectiveness may vary based on individual preferences, music choice, tempo alignment, and structured integration. Future research should investigate optimal integration methods and evaluate their efficacy across diverse populations to enhance generalizability.

Discussion for exploratory analyses

Regarding the acceptance towards this online mindfulness breathing meditation intervention, the current findings showed that this intervention was well-received and structured, with participants reporting psychological and cognitive function improvement, particularly in stress reduction, emotional regulation, attention, and sleep quality. Additionally, most participants found this intervention easy to complete, citing clear guidance, time efficiency, and flexibility as the key facilitators of engagement. However, a minority of the participants experienced difficulties in maintaining focus and perceptions of time moving slowly during meditation, which hindered their practices. These issues align with the previous research, indicating that novice meditation practitioners often struggle with attentional control and mind-wandering, particularly during the initial stages of the mindfulness meditation practice⁹⁶. Overall, this mindfulness breathing meditation intervention received a high acceptability rating, with participants emphasizing its simplicity, accessibility, and positive effects on psychological well-being.

The adherence to mindfulness breathing meditation practices was examined in terms of both quantity and quality of the meditation practices. For the quantity of the mindfulness breathing meditation practices, a progressive increase in practice frequency peaked in Week 3 but followed by a slight decline in Week 4. These findings align with the previous studies, indicating that adherence to structured mindfulness programs tends to improve as the participants develop familiarity with the practice and begin experiencing its benefits^148,149. Regarding the practice duration, participants initially engaged in a longer daily meditation practice during Week 1, followed by a decrease from Week 2 onward. This pattern suggests that participants may have started the intervention with a strong commitment but adjusted their practice duration to a more sustainable level after the first week. A similar pattern has been observed in other mindfulness-based interventions, where initial engagement is characterized by longer practice sessions, followed by stabilization at a shorter yet consistent duration¹⁵⁰. The decline after Week 1 may reflect realistic time constraints or adaptation to the intervention, where participants would choose shorter and more manageable mindfulness breathing meditation practices but still experience its beneficial effects. However, the practice frequency and duration dropped significantly during the three-month follow-up, suggesting challenges in maintaining long-term adherence and indicating that while some participants continued their mindfulness breathing meditation practice, they engaged in significantly shorter practices. This substantial reduction also highlights the challenge of maintaining the mindfulness breathing meditation practice beyond a structured intervention period, showing that adherence to mindfulness breathing meditation may decline once formal program support is ended. For the quality of the mindfulness breathing meditation practices, the engagement with formal mindfulness breathing meditation practice remained relatively stable throughout the entire intervention, with only minor fluctuations. During the three-month follow-up, the formal mindfulness breathing meditation practice engagement dropped, which suggests that while some participants continued practicing mindfulness, others might struggle to maintain the structured practices without the external support (e.g., guided weekly sessions, reminders, and motivation) of the intervention. The informal mindfulness breathing meditation practice showed a progressive increase in engagement but a slight decrease in Week 4. This pattern may suggest that participants gradually incorporated mindfulness into their daily life, reinforcing the practical and accessible nature of informal mindfulness techniques. Unlike formal practice, informal mindfulness does not require dedicated time or structured sessions, allowing participants to apply mindfulness in real-life situations, such as during work, study, or daily tasks. The informal mindfulness engagement remained relatively high during the three-month follow-up compared to the formal mindfulness engagement. This may suggest that participants retained their ability to integrate mindfulness techniques into daily activities even after the structured mindfulness breathing meditation intervention concluded.

While mindfulness-based interventions are generally associated with positive outcomes, emerging research has highlighted the need to assess potential adverse effects^{97,98,101,105}. In the current study, a minority of participants reported sleep-related issues, with one explicitly attributing the issue to the intervention. This finding aligns with previous research indicating that while mindfulness meditation is often used as a tool to improve sleep quality¹⁵¹, it may paradoxically lead to increased arousal and sleep disruption in certain individuals due to heightened cognitive and emotional awareness before bedtime¹⁵². Neurophysiological research suggested that mindfulness breathing meditation could modulate the locus coeruleus-noradrenergic system (a brainstem system involved in attention and arousal regulation)¹⁵³. This coupling of respiration and attentional engagement may result in heightened alertness, in which, while beneficial for daytime cognitive functioning, it may be counterproductive if practice is done close to bedtime¹⁵³. Given the individual variability in sleep responses, it is important for future interventions to provide clearer guidance on optimal practice timing based on individual post-practice responses. In this study, participants were advised to observe and reflect on their post-practice mental states (e.g., alertness, relaxation, or drowsiness) for adjusting their practice timing accordingly. For instance, if a participant experiences post-practice alertness or energetic, practicing mindfulness breathing meditation earlier in the day may be more appropriate, whereas feelings of calmness and drowsiness may suggest suitability for pre-sleep practice. This adaptive approach reflects the need to tailor interventions to individual physiological responses instead of applying a standardized practice timing schedule.

Limitations, future directions, and conclusion

Several methodological and conceptual limitations of this study should be acknowledged. The differences in intervention duration, frequency, and instructor expertise may influence outcomes, and the brief four-week intervention duration may have limited the extent of cognitive and physiological changes observed. Future studies could consider extending the intervention to eight weeks with bi- or thrice-weekly sessions, as this duration and frequency have been commonly associated with more robust effects in mindfulness-based programs¹¹². Second, although this study did include a long-term three-month follow-up, it only focused on the adherence and quality of the mindfulness breathing meditation practice, but not the cognitive and physiological measures, thereby restricting conclusions about the sustainability of observed benefits. Without longitudinal data, it remains unclear whether improvements, particularly in cognitive flexibility, persist over time or require ongoing practice to be maintained. Third, HRV is a highly sensitive metric, and without an elicited stress response, its utility in assessing the impact of mindfulness breathing meditation may be limited. Although the participants’ HRV were assessed before and after the cognitive tasks, the “demanding or stressful” effects caused by the cognitive tasks may not be comparable with a standard stress test (e.g., Trier Social Stress Test)¹⁵⁴.

Despite the limitations, this study offers several notable strengths. Importantly, this study incorporated both subjective measurements, and objective physiological and behavioral indices of stress and cognitive function. This multimodal approach could enhance the internal validity of the findings by reducing common method bias and providing a more comprehensive assessment of the intervention effects. Future studies are encouraged to include follow-up assessments on cognitive and physiological measurements and include a stress test¹⁵⁵ to increase the sensitivity of the HRV measurement.

In conclusion, this four-week online mindfulness breathing meditation intervention shows a significant reduction in perceived stress levels and enhanced cognitive flexibility. Participants in the mindfulness breathing meditation group self-reported this intervention as highly acceptable and effective in promoting mental well-being, and no negative effects were directly attributed to the mindfulness breathing meditation intervention. However, no significant improvement was found in attention, inhibitory control, and HRV after the intervention. Overall, the mixed results suggest that the effect of mindfulness breathing meditation could be task dependent, subject to duration of practice and individuality. Also, the challenge of maintaining long-term adherence suggests a need for future research to explore strategies for promoting sustained engagements, such as integrating motivational supports, habit-forming techniques, or adaptive digital platforms. Longitudinal studies are also warranted to examine whether the observed benefits, though already minimal, persist over time and how continued practice might influence cognitive and emotional outcomes.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

HLY was involved in the recruitment of participants, guided the mindfulness breathing meditation intervention and music therapy, data analysis, and manuscript writing. CPL, TKW, MdV, and WHK were involved in the supervision role. All authors conceptualised the study design and were involved in revision.

Data availability

The data can be found under “Files”, named “SR_HLY_Raw Data”, via this link: 10.17605/OSF.IO/P67EG.

Code availability

The code for HRV analysis can be found under “Files”, named “SR_HLY_Rscript”, via this link: 10.17605/OSF.IO/P67EG.

Declarations

Competing interests

The authors declare no competing interests.