Can heart rate variability demonstrate the effects and the levels of mindfulness? A repeated-measures study on experienced and novice mindfulness practitioners

Abstract

Background

Heart rate variability (HRV) is a potential biomarker that might demonstrate the effects of mindfulness, but it might be influenced by practice experiences. This study wanted to elucidate the possibility of using HRV metrics to reveal the effects of mindfulness and examine its variation between novice and experienced mindfulness practitioners.

Methods

Forty-six participants (20 experienced practitioners, 26 novices) were enrolled to practice 14-day mindfulness training. HRV data were collected during three phases (20 min baseline, T1; 20 min mindfulness, T2; 20 min post-mindfulness, T3) using Holter monitoring. The linear mixed model was conducted to explore the effects of group and time based on standardized data.

Results

The experienced group had higher full-scale scores of FFMQ both in the pre-test (t = -3.34, df = 44, p = 0.002) and the post-test (t = -2.35, df = 44, p = 0.025). Both groups showed significant changes in HRV indices (e.g., RMSSD, SDNN, LnHF) from T1 to T2 or T3 (p < 0.05). In the experienced group, significant fluctuations (p < 0.05) were observed at T2, followed by recovery at T3, in SD1/SD2, Sample Entropy, normalized High Frequency (HFn), DFA_α1, and DFA_α2. In contrast, the novice participants only showed monotonic changes in SD1/SD2 and DFA_α1.

Conclusions

This study revealed significant HRV changes during mindfulness practice, with distinct patterns observed between novice and experienced practitioners.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12906-025-04972-1.

Introduction

In the past four decades, mindfulness has garnered increasing attention in both the cultural and scientific spheres. Given the lack of a universal technical definition of mindfulness [1], many researchers adopted Kabat-Zinn’s description [2], which characterized mindfulness as a means of paying attention to the present experience with openness and acceptance [3]. Previous studies have indicated that mindfulness is widely associated with both mental and physical health, including depression and anxiety [4], sleep quality [5], and blood pressure [6], among others.

Measuring the effects of mindfulness interventions is fundamental to understanding their mechanisms on well-being. In most of studies, self-report questionnaires are primary tools for assessing mindfulness [7]. Popular scales include the Five Facet Mindfulness Questionnaire (FFMQ) [8], the Mindful Attention Awareness Scale (MAAS) [9], the Kentucky Inventory of Mindfulness Skills (KIMS) [10], etc. Although these scales have demonstrated good internal reliability and validity, there is a growing concern for exploring alternative objective measures as augmentations regarding common-method variance (CMV) that may compromise the validity of self-report results [11]. Previous research indicated that self-reported mindfulness scales exhibited high sentiment and semantic similarities with other questionnaires related to mental health and emotion regulation, suggesting the presence of potential CMV during a survey [12]. Consequently, there is a need for objective measures to complement self-report questionnaires to measure the effects of mindfulness practice [12, 13].

Extensive research has explored various physiological biomarkers as potential measures of mindfulness, including neurological activities, brain volumes, autonomic measurements (e.g., blood pressure, heart rate variability, and pulmonary function), immune markers, inflammatory markers, endocrine markers, and telomere length [14, 15]. Among these, cardiac activities, which are recorded via electrocardiography (ECG) have emerged as a promising tool to complement existing self-report questionnaires. ECG offers several advantages, including portability, affordability, and the ability to provide real-time feedback. Moreover, ECG is closely associated with both mental and physical well-being, making it a valuable addition to mindfulness research [16, 17].

Heart rate variability (HRV), which refers to the continuous fluctuation in heart rate (HR), has emerged as a significant biomarker derived from ECG recordings. HRV is recognized as a potential indicator of the activities of the automatic nervous system (ANS) [18]. The ANS plays a vital role in both health and disease [19], as well as being a major component of the emotion response [20]. Typically, the ANS is subdivided into two distinct components, namely the sympathetic nervous system (SNS) and the parasympathetic nervous system (PNS). In healthy individuals, these two systems work in harmony to maintain regulatory balance in physiological autonomic function [21].

Can HRV demonstrate the effects of mindfulness interventions sensitively? Theoretically, HRV serves as a viable biomarker for assessing the impact of mindfulness interventions. It is a physiological indicator of cardiac vagal control, which reflects an individual’s self-regulatory capacity [22]. Research suggests that mindfulness training enhances activity in executive brain areas, such as the prefrontal cortex [23–25]. These regions are known for their role in ANS regulation, emotion regulation, and attention management. This provides support for using HRV as a biomarker in mindfulness-based interventions (MBIs) [26]. Therefore, changings in HRV observed during or after MBIs can indicate improvements in self-regulation and provide objective validation for reported benefits of MBIs [26].

Based on experimental evidence, previous studies have indicated the potential of using HRV as a measure of mindfulness practice outcomes [26]. For instance, some studies have demonstrated significant changes in HRV indices (e.g., the root mean square of successive NN interval differences, RMSSD; low frequency, LF) following a period of mindfulness practice, compared to pre-intervention measurements [27–29]. Additionally, certain findings have shown significant increases in HR or RMSSD from rest to mindfulness practice [30–33]. However, it is worth noting that some studies have not found significant within-group effects of mindfulness interventions on HRV metrics [34, 35]. Furthermore, two meta-analysis studies [36, 37] examined the effects of mindfulness-based training on HRV compared to control or active interventions. They found insufficient evidence to support consistent effects. Although the existing results are somewhat inconsistent, it is important to highlight that the majority of the inconsistent outcomes are characterized by a lack of change rather than demonstrating a change in the opposite direction. This suggests that some variables may have an impact on the observed effects, such as the duration of HRV recording [38], the specific HRV indices used [18], and so forth. These potential factors may contribute to the variability of results and potentially conceal positive findings.

Mindfulness experience is a valuable factor to consider in this context. Distinct differences exist between novice and experienced mindfulness practitioners. Firstly, experienced individuals typically exhibit higher levels of trait mindfulness, which has been positively associated with HRV [39–41]. Secondly, participants with more experience tend to possess advanced mindfulness skills, enhancing their state mindfulness during practice. Repeated practice has been shown to improve attention monitoring skills, thereby stimulating activity and connectivity in various related brain regions [42, 43], which in turn can lead to changes in HRV [32]. Thirdly, novice practitioners often face a greater cognitive load when engaging in mindfulness practices. This is primarily due to their efforts to maintain focus on a specific target and their struggle against mind wandering [25]. Consequently, HRV changes associated with this cognitive load may differ from those observed in experienced practitioners. Therefore, the extent of mindfulness experience is a pivotal variable when exploring HRV as a potential biomarker for mindfulness proficiency. HRV response patterns may vary significantly between novice and experienced practitioners, indicating the possibility that HRV changes reflect the general levels of mindfulness.

Furthermore, monitoring changes in HRV from baseline through states of reactivity and recovery during mindfulness interventions offers valuable insights. Assessing HRV alterations during mindfulness practice can shed light on the direct and immediate effects of such interventions. Equally important is a comprehensive analysis of a wide spectrum of HRV indices, which provides a more holistic understanding of these effects [32]. For instance, in the study of Bortolla et al., they used a range of HRV parameters to investigate the ANS adaptation to mindfulness practice [44]. Their findings revealed significant increases in HR, RMSSD, and the standard deviation of NN intervals (SDNN) from baseline to mindfulness states, followed by notable decreases in HR and SDNN but not in RMSSD after the practice. These results highlighted varying patterns of change among different HRV metrics.

In summary, this study aimed to investigate the utility of HRV as an objective measure for demonstrating the effects of mindfulness practice and the levels of mindfulness between novice and experienced practitioners. To achieve this, we utilized a comprehensive range of HRV metrics to examine the variations in HRV during mindfulness practice across individuals with or without pre-experience of mindfulness. Our hypotheses were twofold: (1) HRV could effectively capture the impact of mindfulness practice, and (2) the manifestation of these HRV changes was different among novice and experienced mindfulness practitioners.

Methods

Participants

A total of 52 healthy adults were recruited through convenience sampling via advertisements on several Internet platforms. The inclusion criteria included: (1) an age range of 18 to 65 years; (2) no intake of psychiatric medication in the past two months and throughout the intervention period; and (3) either a lack of prior experience in mindfulness training or engagement in regular practices, such as mindfulness-based stress reduction (MBSR) [45] or mindfulness-based cognitive therapy (MBCT) [46], at least twice a week for the last two months. Three participants were excluded due to an allergy to electrode patches, non-agreement with the informed consent, and the intake of antidepressant drugs during the experiment, respectively. Additionally, data from three participants, who attended less than 30% of the sessions (10 out of 14), were excluded from the analysis due to their limited availability for the experiment. Consequently, the analysis was conducted with the remaining 46 participants in this study. Among these, 20 individuals met the criteria for the experienced group, having pre-experience of mindfulness trained by professional facilitators of mindfulness and having practiced mindfulness regularly in the preceding two months.

Procedure

Participants received devices and guidelines for the study after providing written informed consent. They were required to complete measurements before and after the intervention. The intervention involved a 2-week online mindfulness program, developed by Lindsay et al. [47]. Throughout the intervention, participants engaged in daily mindfulness practice, following the instructions provided in the audio sessions. These sessions, averaging 22 minutes in duration (ranging from 18 to 26 minutes), included a brief introduction, guided practice, and a period for self-guided practice, as per Lindsay et al. [47]. All participants were instructed to wear a Holter monitor for a total duration of approximately 80 minutes daily, encompassing 30 minutes before and 30 minutes after each mindfulness training session. During pre- and post-intervention periods, they were advised to avoid physical exercise and to continue their usual daily activities in a sitting position. Furthermore, to maintain consistency, participants were asked to practice mindfulness at a similar time each day (e.g., in the morning) and to refrain from consuming stimulants (e.g., coffee, energy drinks) for two hours prior to the sessions. During each daily practice session, participants marked the start and end of their mindfulness practice by pressing a button on the device. Additionally, they manually recorded the specific time of day when the practice occurred for double verification.

Measurements

ECG measurements: A single-channel, 200 Hz resolution portable electrocardiographic device (Holter monitor) was used to monitor ECG signals for all subjects. Participants put on the devices themselves, following the provided instructions.

Questionnaire: The short-form Five Facet Mindfulness Questionnaire (FFMQ) [48] is a 15-item scale developed from the original 39-item FFMQ. It contains five facets including Observing, Describing, Nonjudging, Nonreactivity and Acting with awareness. This is a 5-point Likert scale, ranging from 1 never or very rarely true to 5 very often of always true. Higher scores indicate a higher degree of mindfulness. The short form FFMQ has good reliability in the Chinese population (Cronbach’s α from 0.73 to 0.86) [48].

Data analyses

We cleaned the recorded signals, removing abnormalities based on RR interval histograms, scatter plots, and patterns of ventricular premature beats. Subsequently, the daily RR interval data from 46 valid participants was segmented into 29 discrete points (shown in Fig. 1), spanning three distinct periods: 20 min before the mindfulness training (Baseline, T1), during the mindfulness training (Mindfulness Training, T2), and 20 min post-training (Recovery, T3). This data processing approach ensured maximal utilization of the data and minimized the influence of outliers. Building on this, we computed the HRV parameters across time, frequency, and nonlinear domains using NeuroKit2 [49] implemented in Python 3.9.7. For frequency-domain analysis specifically, the RR intervals were resampled at 4 Hz using cubic spline interpolation to create uniformly spaced time series prior to spectral estimation [50].

Fig. 1.

Signals cleaning process

Respiratory frequencies were extracted from the RR interval sequence. Raw RR intervals are resampled at 4 Hz using cubic spline interpolation to address the non-uniform sampling [50]. The signal was then processed through median filtering using a 5-sample window, followed by state-adaptive bandpass filtering. Given that the typical respiratory rate in humans ranges from 12 to 20 breaths per minute [51], and considering that long-term meditators exhibit lower respiratory rate during rest (12 ± 3 breaths/min, [52]) and meditation (8.3 ± 5.4 breaths/min, [53]), we set a wider frequency range during mindfulness states (0.05–0.35 Hz, corresponding to 3–21 breaths/min) and a narrower band during baseline and recovery periods (0.15–0.35 Hz, corresponding to 9–21 breaths/min). Respiratory cycles were identified using trough-preferential peak detection, capitalizing on the inspiratory-cardio acceleratory phase of respiratory sinus arrhythmia [54, 55]. The final respiratory rate was computed from the inter-trough interval of each 5 min RR-interval series and converted to breaths per minute.

Missing data were handled through the following imputation process. For HRV data, missing values constituted 6.50% of the total dataset (1213/18676), primarily due to invalid Holter signal recordings caused by excessive noise during data acquisition. These issues led to the exclusion of entire cases from valid recordings. Given that the missingness appeared to be related to measurement quality rather than participant characteristics, and potentially associated with observed variables such as time points, the missing data were considered to be Missing at Random (MAR). We imputed the missing values using the mice package [56] in R (version 4.4.3). The norm.predict method, which employs Bayesian linear regression without added residual variance, was used to impute continuous variables. The imputation model included age, Group (experienced or novice), Days (1 ~ 14), and time points (1 ~ 29) as predictors. One imputed dataset was randomly selected from the multiple imputed datasets for subsequent statistical analyses. For FFMQ data, missing values accounted for 3.26% (3/92) of the dataset. These were imputed using the same method described above, with predictors including age and group.

Considering individual differences such as age, gender, and variations in the time periods of mindfulness practice among participants, we employed Z-score normalization for standardizing the data. Within each participant’s dataset, the period of T1 on each day was designated as the baseline. Daily standardization on each day was then performed based on the mean and standard deviation of the baseline. This method effectively neutralized baseline disparities across participants, thereby facilitating a more precise exploration of both within-group and between-group effects of mindfulness. In the process of data analysis, four specific points (P10 ~ 11 and P19 ~ 20) were excluded from the analysis. This exclusion was strategically implemented to mitigate the potential for artifacts resulting from participants transitioning between non-mindfulness and mindfulness sessions and vice versa [57]. All subsequent analyses were then conducted based on these normalized values.

Subsequently, nine HRV indices were computed for further analyses, as detailed in Table 1. Each of these indices represents a typical HRV metric within various analysis domains, as outlined in prior literature reviews [21, 58].

Table 1.

The heart rate variability indices utilized in the present study and their biological interpretations

Analysis Domain	Acronym	Description	Biological interpretation
	RR_mean	Mean of RR intervals	Negative correlated to heart rate
Time-domain	SDNN	Standard deviation of NN intervals	Reflect the regulation of parasympathetically mediated respiratory sinus arrhythmia (RSA) in short-term recordings [17]
	RMSSD	Root mean square of successive NN interval differences	Influenced by the PNS [59]
Frequency-domain	LnHF	Natural logarithm of the power spectrum in the frequency range of 0.15–0.4 Hz	Correlated with RMSSD. Lower LnHF is related to mental problem [59]
	HFn	Normalized HF = HF/(HF + LF)	Modulation index of the ANS's parasympathetic branch affecting the sinoatrial node in the heart [60]
Entropy	SampEn	Estimation of complexity using sample entropy	Reflect the complexity of signal [21]
Poincaré index	SD1/SD2	Ratio of Poincaré plot standard deviation perpendicular the line of identity to standard deviation along the line of identity	Reflect sympatho-vagal balance [21, 61]
Fractal dimensions	DFA_α1	The monofractal detrended fluctuation analysis of the signal, corresponding to short-term correlations	Corresponding to short-term correlations which reflects the baroreceptor reflex [59]
	DFA_α2	The monofractal detrended fluctuation analysis of the signal, corresponding to long-term correlations	Corresponding to long-term correlations which reflect the regulation of the cardiac system [59]

Statistical analysis was performed using Jamovi (Version 2.3) [62]. The normality of the data was assessed by the Shapiro–Wilk normality test. Depending on the data characteristics, different tests were applied: the Student’s t-test was used for normally distributed data with equal variances, Welch’s t-test for normally distributed data without homogeneity of variances, and the Mann–Whitney U-test for non-normallydistributed data [63, 64]. The 14-day standardized data were treated as repeated measurements. A linear mixed model (LMM) was utilized to explore the fixed effects of time, group, and their interaction [65]. The statistical model employed was as follows: $HRV\sim 1+Group+Time+Age+Group\ast Time+\left(1,|,i,d\right)+(1|id:day)$. In this model, HRV represented the specific HRV parameters as the dependent variable, Group (experienced or novice) and Time (T1, T2, T3), along with their interaction, were treated as fixed effects, and Age was included as a covariate. Each participant (id) and day (day) were incorporated with random intercepts. For the analysis of LnHF and HFn, respiratory frequency (Resp_fre) was included as a covariate to account for its potential influence on HRV measures. The statistical model was specified as follows: $HRV\sim 1+Group+Time+Age+Group\ast Time+Resp\_fre+\left(1,|,i,d\right)+(1|id:day)$. Parameter estimation was conducted using the restricted maximum likelihood (REML) method because it produces unbiased estimates of covariance parameters compared to maximum likelihood [65]. Results of post hoc comparisons were adjusted by the Bonferroni multiple testing correction method [66].

Results

Demographic information and trait mindfulness of the participants

The demographic and psychological information of the sample is shown in Table 2. Individuals in the experienced group were older than the novice (t = −5.04, df = 22.52, p < 0.001) and had more practice times than the novice (χ² = 37.29, df = 1, p < 0.001). Novice participants had lower full-scale scores of FFMQ than the experience in both the pre-intervention (t = −3.30, df = 44, p = 0.001) and the post-intervention assessments (t = −2.36, df = 44, p = 0.023).

Table 2.

Demographic and psychological information of the participants

	Novice (n = 26)	Experienced (n = 20)	t/χ2	p
Age, M (SD)	23.65 (3.39)	35.20 (9.80)	−5.04	< 0.001
Gender, n (%)			2.06	0.151
Male	7 (26.92)	2 (10.00)
Female	19 (73.08)	18 (90.00)
Educational level, n (%)			0.73	0.393
Undergraduate	15 (57.69)	11 (55.00)
Postgraduate	11 (42.31)	9 (45.00)
Practice times in the past 2 months, M (SD)	1.46 (4.58)	41.65 (15.92)	37.29	< 0.001
FFMQ, M (SD)
Pre-intervention	45.27 (6.97)	53.55 (9.57)	−3.30	0.001
Post-intervention	51.92 (7.73)	57.65 (8.67)	−2.36	0.023

FFMQ the Five Facet Mindfulness Questionnaire

Fixed effects of time and group on HRV

The fixed effects of time and group on each index of HRV are shown in Table 3. Group showed no fixed effect, whereas time had fixed effects on all indices, and the interaction between group and time showed fixed effects on indices (Table 3). Post-Hoc tests indicated that there were significant differences between T1 and T2 or T3 depending on HRV indices. The original and z-standardized means and standard deviations of each HRV index and respiratory rate across different groups and time periods are presented in Supplemental Tables 1 and 2.

Table 3.

Fixed effects of time and group on each index of heart rate variability

Index	Fgroup (p)	Ftime (p)	Fgroup*time (p)	Post Hoc Test – Time
RR_mean	0.53 (0.469)	194.03 (< 0.001)	51.30 (< 0.001)	T1 vs. T2; T1 vs. T3; T2 vs. T3
RMSSD	0.78 (0.383)	261.02 (< 0.001)	28.12 (< 0.001)	T1 vs. T2; T1 vs. T3; T2 vs. T3
SDNN	1.57 (0.217)	276.68 (< 0.001)	39.19 (< 0.001)	T1 vs. T3; T2 vs. T3
LnHF	0.35 (0.556)	59.38 (< 0.001)	39.03 (< 0.001)	T1 vs. T2; T1 vs. T3; T2 vs. T3
HFn	0.92 (0.344)	39.81 (< 0.001)	44.34 (< 0.001)	T1 vs. T2; T2 vs. T3
SD1/SD2	0.94 (0.338)	94.80 (< 0.001)	114.88 (< 0.001)	T1 vs. T2; T1 vs. T3; T2 vs. T3
SampEn	2.25 (0.142)	27.24 (< 0.001)	58.03 (< 0.001)	T1 vs. T2; T2 vs. T3
DFA_α1	0.01 (0.940)	47.27 (< 0.001)	65.88 (< 0.001)	T1 vs. T2; T1 vs. T3; T2 vs. T3
DFA_α2	2.70 (0.108)	53.19 (< 0.001)	45.34 (< 0.001)	T1 vs. T2; T2 vs. T3

T1 means the time during the baseline, T2 during mindfulness practice, T3 after the practice. Tx vs. Ty means there are significant differences between Tiem x and Time y. Bold font means significant differences (p < 0.05)

RR_mean Mean of RR intervals, SDNN Standard deviation of NN intervals, RMSSD Root mean square of successive NN interval differences, LnHF Natural logarithm of high frequency, HFn Normalized HF, SampEn Sample entropy, SD1/SD2 Ratio of Poincaré plot standard deviation perpendicular the line of identity to standard deviation along the line of identity, DAF_α1 The monofractal detrended fluctuation analysis of the signal, corresponding to short-term correlations, DFA_α2 The monofractal detrended fluctuation analysis of the signal, corresponding to long-term correlations

HRV Changes across mindfulness practice in each group

Figure 2 illustrated the temporal changes in nine z-standardized HRV indices across three time points for novice and experienced participants. The corresponding statistical results are presented in Supplemental Table 3. In the time-domain indices, the novice group exhibited a consistent increasing trend across RR_Mean (T1-T3: t = −20.65, df = 15,464.66, p < 0.001), RMSSD (T1-T3: t = −21.35, df = 15,462.61, p < 0.001), and SDNN (T1-T3: t = −18.67, df = 15,462.43, p < 0.001). In contrast, the experienced group showed augmentations in RR_Mean (T1-T3: t = −8.02, df = 15,464.66, p < 0.001) and RMSSD (T1-T3: t = −10.09, df = 15,462.61, p < 0.001), while SDNN showed a decrease followed by an increase (T1-T2: t = 6.40, df = 15,462.43, p < 0.001; T2-T3: t = −16.49, df = 15,462.43, p < 0.001).

Fig. 2.

Z-standardized HRV indices at the baseline (Time = 1), during mindfulness (Time = 2), and post-mindfulness (Time = 3) between the experienced (n = 20) and the novice (n = 26) groups

For frequency-domain measures, both groups demonstrated an upward trend in LnHF (Novice T1-T3: t = −13.03, df = 15,460.89, p < 0.001; Experienced T1-T3: t = −3.03, df = 15,462.39, p = 0.036). Additionally, the experienced group showed a rise-then-fall pattern in HFn (T1-T2: t = −9.15, df = 15,498.84, p < 0.001; T2-T3: t = 11.69, df = 15,524.87, p < 0.001), while the novice group showed no significant changes (p > 0.05).

Regarding nonlinear indices, the novice group showed an increase in SD1/SD2 (T1-T3: t = −4.57, df = 15,461.47, p < 0.001) and a decrease in DFA_α1 (T1-T3: t = 4.47, df = 15,460.97, p < 0.001). In comparison, the experienced group exhibited a rise-then-fall pattern in SD1/SD2 (T1-T2: t = −17.49, df = 15,461.47, p < 0.001; T2-T3: t = 15.51, df = 15,461.47, p < 0.001) and SampEn (T1-T2: t = −9.93, df = 15,463.21, p < 0.001; T2-T3: t = 11.26, df = 15,463.21, p < 0.001), and a fall-then-rise pattern in DFA_α1 (T1-T2: t = 12.53, df = 15,460.97, p < 0.001; T2-T3: t = −11.2, df = 15,460.97, p < 0.001) and DFA_α2 (T1-T2: t = 11.98, df = 15,460.61, p < 0.001; T2-T3: t = −11.04, df = 15,460.61, p < 0.001).

Differences between two groups

During mindfulness, the novice group showed lower SD1/SD2 and SampEn than the experienced group (SD1/SD2: t = −4.71, df = 46.79, p < 0.001; SampEn: t = −4.77, df = 47.82, p < 0.001), and had higher DFA_α2 (t = 4.54, df = 55.68, p < 0.001). In addition, the experienced group showed a greater increase in HFn from baseline to the mindfulness phase compared to the novice group, with the result approaching statistical significance (t = −2.89, df = 47.30, p = 0.088).

Discussion

The main aim of this study was to elucidate whether HRV could demonstrate the effects of mindfulness practice and the levels of mindfulness among novice and experienced mindfulness practitioners.

Immediate effects of mindfulness on HRV

This study suggests that HRV can reveal the effects of mindfulness by mirroring ANS activity. Several HRV indices changed during or after mindfulness for both the novice and the experienced groups. Specifically, participants showed lower heart rate, higher RMSSD, LnHF and SDNN after mindfulness practice. Indices like RMSSD, LnHF, and HFn predominantly reflect the level of parasympathetic modulation, and SDNN is indicative of overall cardiac variability [59, 67]. Besides, both groups demonstrated significant changes in SD1/SD2 and DFA_α1. The SD1/SD2 ratio is a measure of the sympatho-vagal balance [59, 68], and DFA_α1 indicates short-term correlations of the RR series [59]. Changes in these indices suggest that mindfulness practice could augment PNS activity via the vagal nerve and bring about a lower short-term correlation, which collectively indicates an enhancement in self-regulation capacity [26]. These observations are in line with previous research demonstrating the efficacy of mindfulness in increasing HRV [e.g., 26, 29, 44, 63].

The changes in HRV among both the experienced and the novice participants suggest an immediate effect of mindfulness practice. This immediate effect may be related to engagement in attentional and emotional regulation processes during the practice [69], which have been associated with activity in the prefrontal cortex and anterior cingulate cortex (ACC) [25]. According to the Central Autonomic Network (CAN) model [70], brain regions such as the ACC, insula, and orbitofrontal cortex form an integrated system that modulates cardiac function via sympathetic and parasympathetic pathways, constituting the so-called brain–heart axis [71]. Although the present study did not include direct assessments of central nervous system activity, it is possible that mindfulness practice influences autonomic regulation through enhanced functional interactions between the central and autonomic nervous systems [29]. Further research incorporating neuroimaging or electrophysiological measures is needed to test this hypothesis.

Different patterns of HRV during mindfulness

Although both groups exhibited changes in HRV, the patterns of change were not entirely consistent. Specifically, for SDNN, the novice group showed a continuous increase, whereas the experienced group displayed a decrease followed by an increase. Previous studies have shown that SDNN is associated with attentional processes [72]. Moreover, temporary short-term suppression of HRV, along with focused attention, plays a key role in effective self-regulation [73]. Therefore, the pattern of changes in SDNN may reflect a short-term suppression of overall autonomic activity due to the initial high-level allocation of attentional resources during mindfulness practice.

Regarding HFn, the novice group did not show significant changes, while the experienced group demonstrated an increase during mindfulness practice followed by a decrease afterward. HFn represents the proportion of HF relative to total power, which reflects the ANS's parasympathetic branch affecting the sinoatrial node in the heart [60]. This pattern suggests increased parasympathetic dominance in the experienced group during mindfulness.

In the nonlinear domain, the experienced group showed significant changes during mindfulness practice, followed by rapid recovery after practice, whereas the novice group exhibited delayed responses. For instance, in SD1/SD2, the experienced group showed an initial increase and subsequent decrease, while the novice group only showed a significant increase after the mindfulness session. SD1/SD2 reflects the balance between sympathetic and parasympathetic activity [21, 61]. These findings suggest that the experienced group has higher autonomic flexibility and self-regulatory capacity [26]. Similarly, both DFA_α1 and DFA_α2 in the experienced group showed a decrease during mindfulness and recovery afterward, indicating a temporary reduction in short- and long-term heartbeat correlations [59], corresponding to the increase of SampEn during the mindfulness phase.

Overall, the experienced participants demonstrated more rapid, flexible, and well-coordinated ANS responses during mindfulness practice. Specifically, they showed greater relative parasympathetic dominance alongside reduced overall ANS activity. These patterns may reflect enhanced attentional control and emotional regulation capacities in experienced individuals compared to novice participants. Furthermore, experienced participants may demonstrate a shift toward top-down attentional control and vagal dominance, potentially mediated by the CAN [23, 71]. This enhanced capability could allow for more fluid engagement and disengagement from the mindful state, as reflected in dynamic changes in HRV during and following the practice. Future research is needed to further examine these mechanisms, particularly by integrating assessments of both central and peripheral nervous system activity.

In summary, this study utilized a range of HRV metrics to investigate the variations in HRV during mindfulness practice and its subsequent recovery period, offering a more comprehensive understanding of mindfulness effects compared to previous research [32, 44]. The results highlighted significant changes in HRV both during and after mindfulness practice and captured the differing patterns of these changes across various experience levels. This illuminates the potential of using HRV as a tool to demonstrate the effects of mindfulness. Furthermore, the observed differential HRV patterns between novice and experienced mindfulness practitioners underscore the feasibility of utilizing HRV as a means to distinguish varying levels of mindfulness proficiency.

Despite the promising results of this study, several limitations need to be considered. First, although this study offers an initial comparison of the HRV patterns between individuals with and without previous mindfulness experience, more comprehensive studies are warranted. Future research should include experienced mindfulness practitioners with varied durations of practice experiences (e.g., short-term vs. long-term practitioners) to deepen the understanding of HRV changes across different levels of mindfulness experience. Second, our sample predominantly consisted of female participants (37 females, 9 males), and age had notable differences between novice and experienced groups. Previous studies have indicated the influence of sex and age in autonomic responses [59]. Although we attempted to mitigate the impact of these variables by standardizing the data and including age as a covariate, future research will benefit from a sample that is more balanced in terms of age. This will enhance the generalizability and applicability of the results. Third, the respiratory rate, which influences HRV, was not directly measured during the experiment. Although we attempted to derive it from RR interval data, direct recording (e.g., using a respiratory belt sensor) would have provided more accurate measurements.

In conclusion, this study investigated the changes of HRV indices among individuals with and without previous mindfulness practice experience, both during and following the practice sessions. The significant changes in these indices indicate the viability of HRV as a biomarker for the effects of mindfulness. Furthermore, the distinct patterns of HRV changes between novice and experienced practitioners during mindfulness practice suggest that HRV can also serve as an indicator of varying levels of mindfulness.

Authors’ contributions

YW and YX designed this study, contributed to the conduct of this trial. YX preprocessed the HRV data. YW conducted the statistical analysis and interpreted the results. YW and YX wrote the initial manuscript. WC and JZ critically revised the manuscript. HC and SC contributed to the conceptualization of the study and supervised the whole process. All authors approved the final version of the manuscript.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

The datasets used during the current study available in the following link: https://osf.io/ksnz5/?view_only=c1234963064246529b9c839f851505a2.

Declarations

Ethics approval and consent to participate

Informed consent was obtained from all individual participants included in the study, and all procedures performed in studies involving human participants were in accordance with the ethical standards of the medical ethics committee of the Department of Psychological and Behavioral Sciences at Zhejiang University (Ethic-approved number: [2022]074).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.