Compassion is a valuable, trainable skill which can bring significant benefits to oneself and others. One method for developing compassion towards others is Tonglen, a Tibetan Buddhist meditation which involves taking in suffering from others and sending them well-being. The aim of this study was to investigate the psychophysiological outcomes of Tonglen meditation in healthcare workers, a population who have frequent contact with the suffering of others.
Method
Sixty participants were randomly assigned to listen to a 15-min audio of either guided Tonglen meditation, or a story in the control condition, and completed assessments before and after.
Results
The results showed that, compared to the control condition, Tonglen significantly increased heart rate variability, compassion state, and affective responses to suffering.
Conclusions
This is the first study to show the acute effectiveness of Tonglen in healthcare workers. Results suggest a potential general applicability of this active compassion meditation to populations experiencing empathic distress and burnout.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
The importance of compassion has been widely recognized across various social domains—including culture, religion, healthcare, education, and justice—and is also receiving increasing research attention. Compassion can be defined as “the feeling that arises in witnessing another’s suffering and that motivates a subsequent desire to help” (Goetz et al., 2010, p. 351). Specifically, in an integrative review, Strauss et al. (2016) conceptualized compassion as a cognitive, affective, and behavioral process encompassing five elements: recognizing suffering; understanding its universality; feeling empathy for the person in distress (emotional resonance); tolerating uncomfortable feelings (emotional resilience); and being motivated to alleviate the suffering (behavioral intention).
Several Compassion-Based Interventions (CBIs) have been developed to cultivate compassion in diverse contexts (Kirby et al., 2017a), including Cognitively Based Compassion Training (CBCT; Ash et al., 2021) and Compassion Cultivation Training (CCT; Goldin & Jazaieri, 2017). Overall, evidence has shown that these interventions are effective in improving empathy and compassion towards oneself and others (Brito-Pons et al., 2018) and decreasing anxiety and depressive symptoms (Kirby et al., 2017a).
Moreover, brief periods of compassion practice or CBIs have shown measurable electrophysiological changes. Electroencephalogram (EEG) studies have shown increases in alpha or gamma spectral power associated with different meditative states and with better psychological outcomes (Lomas et al., 2015; Lutz et al., 2004; Schoenberg et al., 2018). Additionally, a recent meta-analysis revealed a significant positive association between compassion and the vagally mediated heart rate variability (vmHRV), a measure of the parasympathetic nervous system’s contribution to heart rate regulation (Di Bello et al., 2020). Compelling evidence supports the role of the vagus nerve in care-giving ability, as well as the feelings of safety and connectedness in social contexts (Petrocchi & Cheli, 2019). An enhanced vmHRV is increasingly recognized as a critical outcome measure in compassion training studies and is associated with better mental and physical health (Kim et al., 2020; Kirby et al., 2017b).
CBIs often incorporate meditation techniques to cultivate compassion (Brito-Pons et al., 2018; Jazaieri et al., 2013), including Tonglen (Goldin & Jazaieri, 2017). Tonglen is a core practice in Lojong (“mind training”) teachings, which originated in tenth-century CE Tibet (Chödron, 1994). These teachings aim to transform our habitual self-centered focus into a more compassionate, altruistic state (Jinpa, 2011). In Tibetan, Tong means “giving or sending” and len translates to “receiving or taking”; Tonglen means “giving and receiving” (Drolma, 2019). This relational practice involves the visualization of compassionately taking in others’ suffering, transforming it, then sending them happiness and well-being (Drolma, 2019). Traditionally, Tonglen meditation combines visualization with breathing techniques: practitioners visualize inhaling others’ suffering as dark clouds, then visualize exhaling their own happiness as white clouds (Drolma, 2019).
Despite some CBIs incorporating Tonglen meditation, research specifically examining this practice remains limited. Current literature contains few studies, most of which are small quantitative pilot studies or qualitative investigations (Gilbert et al., 2023; McKnight, 2012; Pagliaro et al., 2016; Pardy, 2016). This lack may be attributed to the conceptualization of Tonglen as a complex meditation practice, requiring foundational skills such as sustained attention, mindfulness, high mental imagery ability, and loving-kindness (Drolma, 2019; Gilbert et al., 2023). However, some authors argue that Tonglen’s originality and potential benefits may justify its inclusion in the early stages of meditation training, such as within the 8-week CCT program (Goldin & Jazaieri, 2017), or even in shorter, novice practitioner programs (McKnight, 2012).
Preliminary studies highlight Tonglen’s potential benefits. For example, Pardy (2016) conducted a grounded theory study to explore the subjective experiences of experienced meditators after 28 days of Tonglen practice. Participants reported various benefits, including a deeper sense of interconnection with others and, for some, profound insights into the nature of the self. Additionally, two experimental studies (Gilbert et al., 2023; McKnight, 2012) found that Tonglen training significantly increased compassion, positive affect, social safeness, and overall well-being. Traditionally, Tonglen is designed to promote compassion for others and cultivate the strength and courage to confront suffering and remain present with it. These skills are particularly important for professionals working in environments where they regularly witness suffering.
Healthcare workers (HCWs), including nurses, physicians, medical assistants, physical therapists, clinical social workers, and clinical psychologists, are highly affected by burnout (Schaufeli & Greenglass, 2001; Shechter & Norful, 2022; Torrente et al., 2021). HCWs are frequently exposed to human suffering and work in high-pressure environments characterized by time constraints and personnel shortages (Jaiswal et al., 2024). These factors often contribute to burnout, which can lead to reduced job satisfaction (Myhren et al., 2013), diminished quality of professional-patient interactions, and increased incidence of unprofessional behavior and attitudes (Dyrbye et al., 2010). One significant contributor to burnout among HCWs is empathic distress, an emotional struggle experienced when empathizing with others’ suffering (Klimecki & Singer, 2012). This can lead to increased emotional exhaustion and overwhelm, reducing motivation to assist those in need (Klimecki & Singer, 2012).
Recent studies have shown that practicing compassion can effectively counteract the effects of empathic distress (Klimecki et al., 2013, 2014). For instance, Klimecki et al. (2013, 2014) discovered that when participants were confronted with videos of human suffering, they initially exhibited empathic responses accompanied by negative affect. However, after undergoing compassion training, participants experienced a notable increase in positive affective experiences, even when witnessing distress in others. Even though participants continued experiencing negative affect in response to human suffering, this was mitigated by positive affective experiences, specifically those related to feelings of love and affiliation. These results suggest that deliberately cultivating compassion can serve as a coping strategy when faced with the distress of others, thereby reducing burnout.
Furthermore, recent studies have investigated the impact of CBIs in HCWs, indicating increases in self-compassion, mindfulness, and well-being (Beaumont et al., 2016; Jaiswal et al., 2024; Scarlet et al., 2017; Sinclair et al., 2016). Specifically, Scarlet et al. (2017) investigated the effects of CCT on burnout-related outcomes, finding notable reductions in workplace interpersonal conflicts and significant improvements in self-reported job satisfaction. The authors suggest that CBI may promote mental health resilience in HCWs, enhance patient care, and help prevent burnout.
Considering these findings, compassion practices have demonstrated benefits in contexts characterized by frequent exposure to suffering, such as healthcare. Moreover, according to Goetz and Simon-Thomas (2017), compassion research should not only focus on increasing temporary positive emotions after training, but also on developing more robust resilience and better abilities to cope with suffering. In this regard, Tonglen stands out as particularly useful, as it cultivates inner strength and courage while promoting empathetic connection with others’ suffering (Chödron, 1994). To our knowledge, there are no published studies that have specifically used Tonglen meditation with HCW.
The primary aim of the present study was to assess the impact of a single 15-min Tonglen meditation on compassion measurements and responses to the suffering of others in a HCW population. In this randomized controlled study, measurements included self-reported questionnaires, behavioral data, and electrophysiological responses before and after the guided Tonglen meditation or an active control condition. We expected that, compared to an active control condition, Tonglen meditation would increase state compassion scores together with willingness to help, empathy, and positive affect in response to the suffering of others. We also expected increases in vmHRV, as well as in alpha and gamma EEG spectral power.
Method
Participants
Participants were healthcare workers (nurses, physicians, medical assistants, physical therapists, clinical social workers, and clinical psychologists) who voluntarily participated in the study. Participants were recruited through social media and publication dissemination in hospitals, healthcare centers, and institutions associated with the University. The Ethics Committee of the University of Valencia approved the study (registration number: 2016902), which was also pre-registered in Open Science Framework (https://osf.io/x7yh4). As preregistered, the sample size was specified as n = 30 for each group via a priori power analysis, counting for a total sample size of n = 60. All activities conducted in studies involving human participants adhered to the principles outlined in the 1964 Helsinki Declaration and its subsequent revisions or adhered to equivalent ethical standards.
The study inclusion criteria were being over 18 years old, Spanish proficiency, and currently being a healthcare professional. Exclusion criteria included having a cardiac or neurological medical condition or device that prevented electrode use for recording physiological responses, and having a moderately severe or severe score for depression on the Patient Health Questionnaire-9 (PHQ-9) scale. After providing informed consent, 159 individuals completed the online screening, of whom 60 fulfilled the inclusion and exclusion criteria and agreed to participate. Participants were randomized into the Tonglen group (n = 30) or control group (n = 30) before attending the laboratory session. The final sample consisted of 46 cisgender women and 14 cisgender male participants, aged 24–63 years (M = 38.63; SD = 9.76). Of the 60 participants, 49 had some degree of previous experience with meditation (even a single meditation session), and 33 reported a regular weekly practice of meditation. A diagram of the participant flow through the study is shown in Fig. 1. No significant differences were identified between groups (Table 1, all p > 0.05). One Tonglen group participant was excluded from questionnaire analysis due to technical data problems, two control group participants were excluded from ECG analysis (technical data issues), and three from the control group were excluded from EEG analysis because of low signal quality.
Fig. 1
Flow chart for participants in the study. This includes enrollment, allocation, and analysis of the participants. Reasons for exclusion are mentioned
Table 1
Summary of demographic information and baseline trait questionnaires scores for both groups
Variable
Tonglen group (n = 30)
Control group (n = 30)
Age
40.87 (10.31)
36.4 (8.87)
Gender
Cisgender women
23
23
Cisgender men
7
7
Meditation experience
Yes
26
23
No
4
7
Current meditation practice
Yes
19
14
No
7
9
Baseline questionnaires
SOCS-O
84.03 (7.03)
82.87 (6.36)
FCS
19.10 (5.71)
17.47 (6.29)
CMAS
58.41 (13.18)
54.00 (10.61)
No significant between-group differences
Procedure
Participants were instructed to abstain from smoking, consuming caffeine, or engaging in exercise for at least 2 h prior to the experiment. After participants provided signed informed consent, they completed baseline questionnaires. Participants were then seated in a comfortable chair, and the ECG and EEG electrodes were attached. After checking the quality of the electrophysiological signals, task instructions for a baseline resting state were given. Participants completed a 5-min eyes-closed resting state while EEG and ECG data were recorded. EEG and ECG data were continuously recorded during the whole session. Subsequently, participants performed the modified socio-affective video task; responses were recorded in iMotions software (baseline measurement). Following this, the participants listened to a 15-min audio recording with a guided-imagery Tonglen meditation for the experimental group, or a story they were asked to visualize for the control group (additional data are given in Online Resource 1). This active control condition was chosen as a neutral and commonly used active comparison in the field. Participants in the experimental group were instructed to practice Tonglen with a mental image of a loved one in suffering. The guided meditation included three phases, each 5 min in length: first, a focused attention meditation observing the breath; second, the Tonglen meditation itself, with the imagery of the loved one and the exchange of dark and light smoke done with the breathing; and lastly, another phase of focused attention meditation observing the body and/or the respiration while resting in a compassionate state. This guided Tonglen script was derived from the practice used in the CCT program. While participants of both groups listened to the audio, the electrophysiological activity continued to be recorded, and the middle 5 min served as a time window for the analysis (the Tonglen phase in the experimental group). Afterwards, participants repeated the compassion state questionnaire and the SoVT (post-measurement) and then the electrodes were removed. The complete session lasted about 90 min.
Measures
The outcomes of the study included a socio-affective video task, several questionnaires administered in the laboratory via Qualtrics to measure state and trait compassion, and electrophysiological measures such as (vmHRV) in ECG and spectral power in EEG.
Socio-Affective Video Task
A modified version of the Socio-affective Video Task (SoVT) was used to assess social affective responses to short videos (Klimecki et al., 2013, 2014). As in the original version of the SoVT, distinct video sets were used for each measurement time point to avoid habituation. Short videos of 10–18 s were presented without sound and showed raw material drawn from news segments or extracts from documentaries. These depicted men, women, and children. Each video set was composed of nine high-emotion (HE) videos which depicted people who were suffering and nine low-emotion (LE) videos which depicted people performing everyday activities. This resulted in a total of 18 videos per measurement point. Following the original version of the task, video sets were shown in three blocks of three HE or three LE videos. In this adaptation of the original task, participants were asked to rate how much they had experienced empathy, sadness, anger, positive affect, and willingness to help after viewing each block. The sequence of these items was presented in a randomized order. The scale for each question was from 0 (not at all) to 10 (very much). Participants were instructed to observe the videos and pay attention to their affective responses. To assure that all participants had the same understanding, they were informed before the experimental task that the empathy rating would refer to the degree to which they shared the emotion of the people depicted; sadness would refer to their sensitivity to the suffering of the people depicted; and positive affect would refer to their tenderness and care towards the people depicted. After answering all the questions related to one block of videos, a null event was presented (a display with the text “Observe”) in order to prepare them to be ready for the following block. The presentations of visual stimuli and the participant questionnaires were carried out using iMotions v.10 software (iMotions A/S, 2022). The total task duration for the SoVT was around 10 min, and the SoVT was administered at baseline and post-audio task (Tonglen or control).
Questionnaires
Sussex-Oxford Compassion for Others Scale
Sussex-Oxford Compassion for Others Scale (SOCS; Gu et al., 2020) measures trait compassion for others, consisting of 20 items on a 5-point Likert scale: 1 (not at all true) to 5 (always true). Participants completed this scale only at the beginning of the session (baseline). Internal consistency in our sample was found to be good (α = 0.87, ω = 0.87).
Fears of Compassion Scale
Fears of Compassion Scale (FCS; Gilbert et al., 2011) assesses the experience of resistance to engaging in compassion and being fearful of expressing compassion for others. It consists of 10 items on a 5-point Likert scale: 0 (don’t agree at all) to 4 (completely agree). Participants completed this scale only at baseline. Good internal consistency was demonstrated in our sample (α = 0.81, ω = 0.80).
Compassion Motivation and Action Scales
Compassion Motivation and Action Scales (CMAS; Steindl et al., 2021) measures motivation and action as core components of compassion. The total scale consists of 12 items on a 7-point Likert scale: 1 (strongly disagree) to 7 (strongly agree). Participants completed this scale only at baseline. Our findings indicate good internal consistency for the scale (α = 0.88, ω = 0.89).
Compassion for Others State
This scale was adapted from the State Self-Compassion Scale Short Form (SSCS-S; Neff et al., 2021) to measure the state of compassion towards others in the present moment. The version used in this study consisted of 5 items on a 5-point Likert scale: 1 (not at all true for me) to 5 (very true for me). The five items were selected and developed based on the five-element definition of compassion by Strauss et al. (2016). These items assessed: recognizing suffering, understanding the universality of suffering, emotional resonance, toleration of uncomfortable feelings, and motivation to act to alleviate suffering. Participants completed this scale at baseline and after the audio recording for both the control group and the guided Tonglen meditation group (post). Additional information is given in Online Resource 2. The measure showed good internal consistency in the present sample (α = 0.77, ω = 0.79).
Compassion Practice Quality Scale
Compassion Practice Quality Scale (CPQS; Andreu et al., 2022; Navarrete et al., 2021) contains 12 items rated on a scale ranging from 0 to 100 to assess practice quality after a compassion meditation. The scale, originally in Spanish, has a two-factor structure comprising mental imagery and somatic perception. Scores are derived by calculating the mean of the items. Higher scores indicate greater quality of compassion practice. Tonglen group participants completed this scale after the guided meditation (post). Good internal consistency was found for the mental imagery (α = 0.87, ω = 0.88) and somatic perception (α = 0.72, ω = 0.73) subscales, as well as for the total scale (α = 0.85, ω = 0.85).
Electrophysiological Data
HRV was computed from ECG, recorded with a Biopac MP160, using Ag/AgCl electrodes placed under the clavicles and on the first left lower rib, at a sampling rate of 200 Hz. Recordings from ECG were collected in iMotions software v.10 (iMotions A/S, 2022), which was also used for the analysis. HR data were first checked manually offline for artifacts (electrode noise, movement, and extraordinary peaks), and then were subjected to a HRV analysis with iMotions. Data were filtered with a bandpass filter with a low cutoff frequency fixed at 5 Hz, and a high cutoff frequency fixed at 15 Hz. The root mean square of successive differences (RMSSD) was calculated as a primary index of vmHRV.
EEG was recorded using a Brain Products actiCHamp 32-channel gel-based active electrodes setup at a 500-Hz sampling rate (Brain Products GmbH, Gilching, Germany). The electrodes were mounted on an actiCAP elastic cap, with a layout according to the International 10–20 system, and impedances were kept below 25 kΩ. Recordings from EEG were collected in iMotions software v.10 (iMotions A/S, 2022), which was also used for the analysis. For the spectral power analysis, iMotions software performed the following procedure. First, raw EEG data were filtered for low and high frequencies (retaining 0.1–100 Hz), using a zero phase-lag Butterworth bandpass filter. Then, the power noise was removed via a notch filter using a zero phase-lag Butterworth filter (filtering the 50-Hz frequency). The signal was re-referenced based on the CZ electrode, and artifact rejection was performed with a 120-μV cutoff. Data points with an absolute value superior to this threshold were considered as outliers and removed from further analysis. As a quality check, the percentage of data removed was computed, and participants with more than 20% of data removal were excluded from the analysis due to low quality signal. The computation of the power spectral density (PSD) was estimated using fast Fourier transform (FFT), where the filtered signal was chunked into 0.5-s time series (with an overlap of 0.25 s). If data points for a given 0.5-s window were invalid (considered an artifact), the time series was removed from computation. Then, power band computation was performed by averaging PSD estimates across frequencies of interest: delta (1–3 Hz); theta (4–7 Hz); alpha (8–12 Hz); beta (13–29 Hz); and gamma (30–45 Hz). The values were transformed into decibels (dB). The computation of mean spectral power values for each available channel was performed by averaging PSD estimates (after averaging within powerbands of interest) over the entire stimulus duration, then log transformed into dB. For this analysis, five a priori regions of interest (ROIs) were used: central electrodes; frontal electrodes; occipital electrodes; temporal electrodes; and parietal electrodes. The average spectral power of each ROI was computed by averaging PSD estimates (after averaging within powerbands of interest) across corresponding electrodes. The values were then transformed into dB after averaging.
Data Analyses
Descriptive statistics were obtained, and distribution plots were screened in all groups and times. Minor deviations from normality were observed in some variables. However, the RMSSD HRV showed a more pronounced positive skewness, leading to bias in central tendency estimates. In addition, we tested for potential baseline differences across groups using t-tests when assumptions were met (mainly normality and homoscedasticity) and if not, robust or non-parametric alternatives were implemented.
To test if Tonglen meditation impacted self-reports and EEG patterns, 2 (Tonglen or control) × 2 (pre and post) mixed ANOVAs were implemented with α = 0.05. Assumptions were checked (Levene’s test for equal variances, normality with Q-Q plots and Shapiro–Wilk test, and sphericity not needed since only two time are present). Effect size was estimated with partial eta-squared (ƞ2p), with the corresponding assumed effect size used in the pre-registered power analysis as ƞ2p = 0.13 (equivalent of d = 0.39). Post hoc comparisons were performed with the Bonferroni procedure, along with marginal means with 95% confidence intervals. The 2 × 2 ANOVAs were performed for each of the affective responses in the SoVT; each of the items of the compassion state questionnaire; HRV data; and the spectral power of each ROI.
Regarding RMSSD, a non-parametric approach was implemented using Wilcoxon signed-rank tests (paired samples) for each group across times. More concretely, a separate test for each group, with a correction for multiple testing, was performed, making them analogous for the simple effects of an ANOVA to describe interactions. This approach was preferred to the transformation of the variable (to allow easy interpretations) or alternatives such as generalized linear models, multilevel models, or Bayesian approaches (to prioritize parsimonious findings as this is an exploratory study). Specifically, a one-tailed test for the Tonglen group (pre < post, to increase power) and a two-tailed test for controls, to reflect our hypotheses, were used. Multiple comparison correction was made with Bonferroni correction for two comparisons (given two tests were implemented), resulting in α = 0.02 for the Tonglen group (as a one-tailed test) and 0.01 for the control group (as a two-tailed test). This way, we can establish a non-parametric equivalent of a simple effect for ANOVA, but based on medians, providing a more precise and valid estimate of differences. Effect size was estimated with rank-biserial correlations, with the corresponding assumed effect size used in the pre-registered power analysis as r = 0.19 (equivalent of d = 0.39). In addition, the Hodges-Lehmann estimates of median differences were used with confidence intervals.
All analyses were computed using IBM SPSS Statistics (version 26) software, except for the non-parametric tests, which were computed in JASP (JASP Team, 2024).
Results
Self-Report questionnaires
Preliminary analyses found no significant group differences at baseline on socio-demographic characteristics: gender (χ2(1) = 0.00, p = 1.00); age (t(58) = 1.79, p = 0.08); marital status (χ2(4) = 1.18, p = 0.88); current employment situation (χ2(5) = 3.67, p = 0.59); and previous experience with meditation (χ2(1) = 1.00, p = 0.32). Also, within the participants who reported previous meditation experience (n = 26 for Tonglen group and n = 23 for control group), there was no difference between groups in terms of current regular meditation practice (χ2(1) = 0.83, p = 0.36). Finally, there were no significant group differences in self-reported compassion for others measured by the SOCS-O (t(57) = 0.67, p = 0.51); fears of expressing compassion to others measured by the FCS (t(57) = 1.05, p = 0.30); and compassion motivation and action for others measured by the CMAS (t(57) = 1.42, p = 0.16). Baseline means and standard deviations for trait questionnaires and demographic data by groups are reported in Table 1.
With regard to state compassion for others, from baseline to post, there was a significantly greater increase in the Tonglen versus control group on two of the items (Fig. 2). The items with a significant time × group interaction were recognizing suffering, F(1, 57) = 7.41, p = 0.01, ƞ2p = 0.12; and motivation to act to alleviate suffering, F(1, 57) = 4.48, p = 0.04, ƞ2p = 0.07. There were significant effects of time for all participants, but no time × group interaction for understanding the universality of suffering, emotional resonance, and tolerance of uncomfortable feelings.
Fig. 2
Items of Compassion for Others State questionnaire. Self-reported recognizing suffering (A) and motivation to alleviate suffering (B) significantly increased after Tonglen meditation (purple), compared to the control condition (orange). Error bars indicate the standard deviation of the mean. *p < 0.05
For the quality of the compassion practice in the case of the Tonglen group, results showed high scores on the subscale of mental imagery (M = 77.51, range = 20.29–100, SD = 17.42) indicating that participants reported a high ability to create, inspect, sustain, and transform a vivid mental imagery during the Tonglen meditation and low difficulties with imagination. On the somatic perception subscale, the mean score of 66.44 (range = 27.20–96, SD = 18.52) indicates a good level of somatic perception of feelings of warmth and connection during the Tonglen meditation. A total mean score of 72.67 (range = 33.23–95.77, SD = 15.02) was obtained after the practice.
Baseline and post means and standard deviations, as well as differences in change scores for state questionnaires by group, are reported in Table 2.
Table 2
Descriptive statistics and differences in scores for state compassion for both groups and compassion practice quality scale
Tonglen group
Control group
Between-group differences
Baseline
Post
Baseline
Post
Time × group interaction
M
SD
M
SD
M
SD
M
SD
F
η2p
Compassion for Others State
Recognizing suffering
4.00
0.76
4.48
0.57
4.40
0.77
4.37
0.70
7.41*
0.12
Universality of suffering
4.31
0.89
4.62
0.62
4.37
0.76
4.47
0.68
1.20
0.02
Emotional resonance
3.55
1.09
4.00
0.76
3.30
1.06
3.67
1.06
0.10
0.00
Toleration of feelings
3.69
1.17
4.14
1.06
3.90
0.99
4.13
0.82
0.44
0.01
Motivation to alleviate
4.17
0.85
4.45
0.78
4.40
0.72
4.30
0.95
4.48*
0.07
CPQS
Mental imagery
77.51
17.42
Somatic perception
66.44
18.52
Total
72.67
15.02
*p < 0.05. Baseline comparisons were not significant (all with p > 0.05)
Socio-Affective Video Task
For affective responses in the SoVT regarding the videos of suffering (HE), the RM-ANOVA revealed group by time interactions suggesting that, compared to control, Tonglen resulted in significantly greater increases in empathy, positive affect, sadness, and willingness to help (Fig. 3). The significant time × group interactions were empathy, F(1, 58) = 6.02, p = 0.02, ƞ2p = 0.09; sadness, F(1, 58) = 4.71, p = 0.03, ƞ2p = 0.08; positive affect, F(1, 58) = 8.02, p = 0.01, ƞ2p = 0.12; and willingness to help, F(1, 58) = 4.27, p = 0.04, ƞ2p = 0.07. The only rating with a significant effect of time (with p < 0.05) but not time × group interaction was anger, with significant increases in both groups from baseline to post. Baseline and post means and standard deviations, as well as differences in change scores for SoVT affective ratings by group, are reported in Table 3.
Fig. 3
Affective ratings in the socio-affective video task. Self-reported empathy (A), sadness (B), positive affect (C), and willingness to help (D) significantly increased after Tonglen meditation (purple), compared to the control condition (orange). Error bars indicate the standard deviation of the mean. *p < 0.05; **p < 0.01
Table 3
Descriptive statistics and differences in scores for SoVT affective ratings for both groups
Tonglen group
Control group
Between-group differences
Baseline
Post
Baseline
Post
Time × group interaction
M
SD
M
SD
M
SD
M
SD
F
η2p
SoVT affective ratings for HE videos
Empathy
7.21
1.63
7.47
1.75
7.89
1.11
7.59
1.37
6.02*
0.09
Sadness
6.74
2.25
6.98
2.13
7.54
1.53
6.96
1.99
4.71*
0.08
Anger
2.08
1.95
3.16
2.50
2.76
2.53
3.59
2.69
0.26
0.00
Positive affect
6.18
2.39
7.11
1.98
6.67
1.98
6.67
2.17
8.02**
0.12
Willingness to help
7.08
1.99
7.49
2.03
7.98
1.51
7.76
1.81
4.27*
0.07
*p < 0.05; **p < 0.01. HE, high-emotion videos
For affective responses in the SoVT regarding the neutral videos (LE) as a control measure, the results showed no significant time × group interaction (all with p > 0.05) but a significant effect of time in all the ratings, indicating a decrease in the affective responses in time, but no differences between the Tonglen and control groups. This result suggests that the observed effects in the affective responses are specific to the context of suffering.
Electrophysiological Responses
Baseline and post means and standard deviations, as well as differences in change scores for electrophysiological data by group, are reported in Table S1 (Online Resource 3). No significant differences were found between baseline scores (all with p > 0.05).
For vmHRV, there was a significant increase in RMSSD across time for the Tonglen group (W = 132, Z = − 2.07, p = 0.02, rrb = − 0.43, SErrb = 0.21) with around 4 points as the median increase (HL estimate = − 4.05, 95%CI = [− ∞; − 0.72]. However, the control group showed a non-significant change in RMSSD across time (W = 276, Z = 1.66, p = 0.95, rrb = 0.36, SErrb = 0.21, HL estimate = 2.78, 95%CI = [− 0.74; 7.15]), revealing a significant increase in the Tonglen versus control group from baseline to post (Fig. 4). These analyses are conservative due to baseline distributions being skewed towards lower values. Thus, potential differences in the population could be more pronounced. Baseline comparisons were not significant (W = 318.00, p = 0.08, rrb = − 0.27, SErrb = 0.15).
Fig. 4
Effects of conditions in heart rate variability. The physiological measure RMSSD (vmHRV) significantly increased after Tonglen meditation, compared to the control condition. Plots describe raw data points across time, boxplots with medians, and density distributions
For spectral power EEG data, there was no evidence of time × group interactions in any of the frequencies and the ROIs used (all with p > 0.05). We did observe a significant effect of time (i.e., decreases) across both groups in central delta, F(1, 55) = 8.93, p = 0.01, ƞ2p = 0.14; occipital beta, F(1, 55) = 5.41, p = 0.02, ƞ2p = 0.09; and temporal alpha, F(1, 55) = 5.36, p = 0.02, ƞ2p = 0.09.
Discussion
In this study, we examined the acute effects of a single Tonglen meditation session in healthcare workers via self-report, behavioral, electroencephalographic, and cardiac measures. Compared to an active control group, healthcare workers who practiced Tonglen meditation showed increases in two dimensions of self-reported compassion for others (recognizing suffering and the motivation to alleviate suffering) and increased vmHRV. Additionally, in response to the suffering of others in the context of an experimental task, healthcare workers who practiced Tonglen meditation showed significantly increased empathy, sadness, positive affect, and willingness to help compared to the active control group. There were no between-group differences on any of the EEG measures in prespecified brain regions of interest.
To assess self-reported changes following Tonglen meditation, we used the five elements of compassion proposed by Strauss et al. (2016): recognizing suffering, understanding the universality of human suffering, feeling for the person suffering, tolerating uncomfortable feelings, and motivation to act to alleviate suffering. In this conceptualization, compassion is seen as an awareness of someone’s suffering, being moved by it (emotionally/cognitively), and feeling motivated to help (Strauss et al., 2016). Several common definitions of compassion emphasize that compassion also involves being able to tolerate uncomfortable feelings that arise in oneself as a result of seeing suffering (including frustration or anger), and also recognizing a commonality with the sufferer or understanding the universality of suffering (Strauss et al., 2016). In our results, from baseline to post Tonglen meditation, there was an increase in recognizing the suffering of the other, and in motivation to act to alleviate suffering, but not in understanding the universality of suffering, emotional resonance, and toleration of uncomfortable feelings. Given the “active” nature of Tonglen meditation, an increase in both recognizing or being aware of the other suffering and also the motivation to act to alleviate suffering is expected, since this is directly practiced in Tonglen: first, a recognition of suffering and a visualization of the other’s suffering, then the taking-and-giving phase to alleviate the suffering. Considering the procedure of Tonglen meditation, we did not expect any change in the dimensions of understanding the universality of suffering and emotional resonance, because these dimensions are not directly trained by this particular meditation technique. One dimension that we expected to increase, but did not change, was the toleration of uncomfortable feelings. In theory, Tonglen should also train this ability when taking in the suffering while inhaling. As participants practiced only for a single session of Tonglen meditation, it is likely that more practice would be needed to strengthen this dimension of compassion.
Results for quality of compassion practice showed high scores on mental imagery, suggesting that participants did not have difficulties with imaginative skills, reporting high abilities to create, inspect, sustain, and transform a vivid mental imagery during Tonglen meditation. Additionally, a moderate-to-high level of somatic perception was observed, indicating that participants were able to perceive feelings of warmth and connection with the other during the meditation. Our results indicate that people without much experience in meditation, such as novice meditators, are able to practice Tonglen without having highly challenging experiences. Moreover, a single Tonglen meditation produces similar or higher levels of compassion practice quality as compared to those achieved after 8 weeks of practice with the CCT program (Andreu et al., 2022). Although the short-term effects of one laboratory session and the long-term effects of a compassion intervention are not directly comparable, these results nevertheless offer promising insights for future studies.
Regarding the socio-affective video task, our results showed an increased affective response towards videos of suffering in the Tonglen group, compared to active controls. This effect was observed for empathy (the degree to which participants shared the emotion of the people depicted in the videos), sadness (sensitivity to the suffering of depicted people), positive affect (tenderness and care towards the depicted people) and willingness to help; all effects were of medium magnitude. Our results align with previous studies showing an increase in positive affect after a psychoeducation and compassion training, although this increase in previous studies was observed for both suffering and neutral videos (Klimecki et al., 2013, 2014). This difference may be due to the fact that, in our study, we only used a single practice session and not a training program, which may be necessary to also observe effects in the neutral videos. In the modified version of the task that we used, we added ratings for sadness and anger (as a distinction within negative affect), together with willingness to help. This modification allowed us to find that Tonglen meditation not only increased positive affect and empathy, but also sadness (sensitivity to suffering) and willingness to help, while not increasing anger, compared to controls. This result is aligned with the understanding of compassion as a blend of sadness and love, where sadness is a distinct emotion from compassion, but sadness can be considered an antecedent to others’ compassion, and its expression is expected to elicit compassion for others (Goetz et al., 2010). The results from the behavioral task affective responses are also in line with those obtained in the self-report questionnaires, showing that one acute effect of Tonglen meditation is an increase in the motivation to help or alleviate suffering, highlighting this as a key dimension influenced by Tonglen.
Regarding the electrophysiological results, we found an increase in vmHRV in the Tonglen group compared to active control, with a medium-to-large effect size and no between-group differences in the EEG spectral power analysis. The increase in vmHRV after practicing Tonglen meditation was expected, showing an increase in a consistently suggested physiological marker of vagal parasympathetic activation and a commonly observed physiological marker in compassion practices (Di Bello et al., 2020; Kirby et al., 2017b). Tonglen is not considered a relaxing practice in itself, but our results show that, from a physiological perspective, the effect of Tonglen meditation is similar to other compassion practices in this regard (Di Bello et al., 2020). We observed a slight decrease in vmHRV in the active control group, suggesting a decrease in parasympathetic activation after listening to the story, which may indicate that listening to and visualizing a story for 15 min may even be uncomfortable or stressful, or it may activate the sympathetic nervous system. About the null effects in the EEG findings, we initially expected an increase in alpha and gamma power after Tonglen meditation, given previous studies showing an increase in these frequencies associated with a meditative state (Lomas et al., 2015; Lutz et al., 2004; Schoenberg et al., 2018). It is important to acknowledge that these results were obtained after longer training programs or in populations of expert meditators, which involve very different research designs than a single meditation session. The absence of effects from Tonglen meditation in EEG frequencies suggests that longer training programs with more practice sessions may be needed to impact electrical brain activity. Furthermore, as we used an active control group who were asked to listen to an audio and imagine the story, we aimed to control for several non-specific factors, including mental imagery. Recently, an increasing number of studies in contemplative research have shown an absence of results when using an active control group (van Dam et al., 2018), suggesting that controlling for non-specific factors is key in the field. Also, we did not find many studies focused on compassion meditation using EEG (most of the neuroscientific studies used fMRI), which makes it difficult to compare or interpret our null results in this study. As strongly recommended in contemplative research (van Dam et al., 2018), the use of an active control group allowed us to account for non-specific effects. In particular, because we asked the control group to use imaginative skills during the story, it was possible to control for the mental imagery factor, which can be important in electrophysiological results (Gale et al., 1972). Considering this, it is not surprising that no EEG-level effects were found in this study comparing a single session of Tonglen meditation to an active control group.
Limitations and Future Directions
This study has several limitations. First, the study was exploratory in nature, making our findings preliminary and subject to potential biases. Future studies should aim for a confirmatory design and analysis (e.g., Bayesian approaches with informed priors). In addition, we measured the effects of a single 15-min Tonglen meditation session, which may be too short to obtain significant effects on several of the measures used. Future studies should consider extending the research to longer-term interventions with a more regular Tonglen practice instead of a single practice session, to fully assess the potential benefits of Tonglen meditation. Second, the difficulty in recruiting healthcare workers who could participate in this study in the laboratory meant that the number of participants was not high. Future studies should aim for high-powered representative samples to enhance replicability and generalizability of findings. Also, in this study, we recruited a general population of healthcare workers. Future studies should consider focusing on different populations of healthcare workers, for example, palliative care workers. Third, this study did not aim to distinguish effects in participants with or without prior meditation experience; this was not an inclusion criterion, and no analyses were performed to distinguish these effects. The sample size of this study did not allow us to perform a subgroup analysis to determine whether prior meditation experience made a difference in outcomes. Although most participants had some prior meditation experience, only half of the sample reported a regular weekly meditation practice. Since our sample was not recruited in meditation-related centers, but in health and education centers, we could assure that previous experience in meditation was due to recruitment. However, the nature of the study can be considered an element that induced self-bias, since the participants were invited to join a study about compassion, so those interested in participating were probably familiar with compassion and had previously practiced it. Future studies are needed to determine whether prior experience may be an important factor to consider.
To our knowledge, this is the first study to show the acute effectiveness of a single Tonglen meditation in increasing compassion and affective responses to suffering in healthcare workers. The results of this study suggest a potential general applicability of this active compassion meditation to populations at risk of experiencing empathic distress and burnout, such as emergency physicians, caregivers, palliative care professionals, volunteers from NGOs, and therapists in contexts such as trauma and war, among others. Together with other studies performed in novice meditators, it may be suggested that Tonglen meditation is a suitable practice for people without extensive experience in meditation or previous training. Future studies should measure the effects of a longer compassion training program focused on Tonglen meditation to differentiate the effects of a single practice versus a longer training program and to distinguish whether Tonglen meditation is more particularly beneficial or has different effects in people with higher previous meditation experience than in novice practitioners. Furthermore, qualitative information, for example, collected by micro-phenomenological interviews, may provide useful information to interpret quantitative results and to understand the lived experience of participants with a complex practice such as Tonglen meditation.
Declarations
Ethics
The present study was approved by the Ethics Committee of the University of Valencia (registration number: 2016902).
Informed Consent
All study participants provided informed written consent prior to study enrollment.
Conflict of interest
The authors declare no competing interests.
Use of Artificial Intelligence
AI was not used.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
share
DELEN
Deel dit onderdeel of sectie (kopieer de link)
Optie A:
Klik op de rechtermuisknop op de link en selecteer de optie “linkadres kopiëren”
Met BSL Psychologie Totaal blijf je als professional steeds op de hoogte van de nieuwste ontwikkelingen binnen jouw vak. Met het online abonnement heb je toegang tot een groot aantal boeken, protocollen, vaktijdschriften en e-learnings op het gebied van psychologie en psychiatrie. Zo kun je op je gemak en wanneer het jou het beste uitkomt verdiepen in jouw vakgebied.