Skip to main content
NCBI home page
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2026 Feb 3;42(1):e70145. doi: 10.1002/smi.70145

Self‐Soothing Touch Reduces Momentary Stress, Fatigue, and Loneliness Comparable to Brief Meditation: A Randomised Controlled Trial

Franziska Maier 1, Ina Luttenberger 1, Aljoscha Dreisoerner 1,2, Bence Szaszkó 3,✉
PMCID: PMC12865878  PMID: 41631519

ABSTRACT

Chronic stress has well‐documented adverse effects on physical and psychological health. Beyond contributing to the development of fatigue, its impact is intensified by social stressors such as loneliness, making the development of effective interventions crucial. Our randomised controlled trial therefore investigated whether a 14‐day self‐soothing touch (SST) intervention reduces stress, fatigue, and loneliness compared to a minimal‐instruction meditation control in 78 chronically stressed individuals (M Age  = 22.2 years; 81% female). We assessed acute (change pre‐to‐post session) and cumulative effects (across days) using Ecological Momentary Assessment (EMA), while also collecting retrospective self‐reports at baseline, post‐intervention, and 4‐week follow‐up. For EMA outcomes, we additionally tested moderation by attachment anxiety and avoidance. Using linear mixed‐effects models, both SST and meditation significantly reduced momentary stress (SST: b = −0.41, SE = 0.08, t = −4.79, p < 0.001; Control: b = −0.56, SE = 0.09, t = −6.43, p < 0.001), as well as fatigue (p SST < 0.001, p Control < 0.001) and loneliness (p SST ≤ 0.011, p control = 0.004) from pre‐to‐post session, with no significant group differences (all ps ≥ 0.212). SST but not meditation yielded a decrease in pre‐session fatigue across the intervention period (b = −0.06, SE = 0.02, p < 0.001), with stronger reductions among individuals higher in attachment avoidance. In contrast, neither intervention had effects on retrospective measures (all ps ≥ 0.117). Overall, SST emerged as a feasible and accessible approach, comparable to brief meditation in reducing stress, fatigue, and loneliness, with additional benefits particularly for those high in attachment avoidance.

Keywords: ecological momentary assessment, fatigue, loneliness, meditation, self‐soothing touch, stress

1. Introduction

Chronic stress is a prevalent issue with substantial impacts on mental and physical health, as well as daily functioning (Terehov and Yakovlev 2023). It stems from a persistent inability to adapt to stressors due to an imbalance between stressors and coping factors, altering regulatory systems and contributing to physiological and psychopathological consequences (Dantzer et al. 2008; Dunlavey 2018; Miller et al. 2007). Sustained glucocorticoid exposure through the hypothalamic‐pituitary‐adrenocortical (HPA) axis is neurotoxic, affecting glutamate regulation and the autonomic nervous system (ANS; Arnsten 2009; Thayer and Lane 2000) and thereby promoting chronic low‐level inflammation (deRijk et al. 1994). Of note, elevated cortisol (the end product of the HPA axis) is linked, among others, to depressive symptoms (Iob et al. 2020), generalised anxiety disorder (Juruena et al. 2020), and neurodegenerative diseases (Esch et al. 2002).

Chronic stress is also a central pathophysiological factor in fatigue (Nater et al. 2011), with dysregulated stress response systems and inflammatory biomarkers repeatedly found in patients suffering from fatigue (Papadopoulos and Cleare 2011; Steptoe et al. 2007). It is characterised by severe tiredness, lack of energy, and impaired concentration (Landmark‐Høyvik et al. 2010). In contrast to short‐term tiredness, fatigue is persistent and cannot be alleviated by rest or sleep with its cognitive, emotional, and physical consequences affecting everyday life. Fatigue may reflect an adaptive response to sickness that helps the body to prioritise recovery (Dantzer et al. 2008) but remains an unpleasant, avoidance‐motivating state that significantly compromises quality of life (Boksem and Tops 2008).

Stress response systems are also activated by loneliness. Loneliness is known to act as a potential social‐evaluative stressor engaging the response systems dysregulated by chronic stress: Lonely individuals show alterations of the HPA axis (Adam et al. 2006), diminished autonomic flexibility in response to stress (Song et al. 2025), and a pro‐inflammatory immune profile (Jaremka et al. 2013); longitudinal work further links loneliness to the symptom cluster of pain, depression, and fatigue, potentially contributing to sustained high levels of fatigue (Jaremka et al. 2014).

Against this backdrop, simple, accessible, and easily implementable strategies mimicking the stress‐buffering features of supportive contact and mitigating the numerous negative impacts of chronic stress are particularly attractive. We propose self‐soothing touch (SST) (Dreisoerner et al. 2021) as a potential intervention that does not require extensive training, equipment, or another person. SST is a conscious physical gesture, associated with increased self‐compassion, defined as treating oneself with kindness during times of suffering (Neff 2003). As a physical expression of self‐compassion, SST is intended to generate warm, affiliative feelings similar to those elicited by comforting interpersonal contact (Dreisoerner et al. 2021; Morrison 2016). Self‐compassion is a crucial resource for coping with distress (Neff 2023) and has been linked to protection from psychopathology, including depression and anxiety (Han and Kim 2023; MacBeth and Gumley 2012; Neff 2023). Furthermore, higher self‐compassion has been associated with lower fatigue (Babenko et al. 2019), and training programs have decreased symptoms of burnout, one of which is fatigue (Delaney 2018; Eriksson et al. 2018). Self‐compassion may also strengthen coping with loneliness (Andel et al. 2021), while self‐compassion focused programs have demonstrated reductions in loneliness in older adults (Patapoff et al. 2024). Likewise, affective touch showed beneficial effects on these dimensions: Touch‐based interventions have been associated with stress‐reducing and stress‐buffering effects (Dreisoerner et al. 2021; Morrison 2016; Packheiser et al. 2024); affective touch has also been shown to reduce feelings of social exclusion in experimental paradigms (von Mohr et al. 2017).

Attachment patterns may also shape responses to SST. As SST mimics affective touch, individual differences in how social touch is appraised may also moderate the benefits of SST. Attachment styles influence both, the pleasantness rating of touch (Spitoni et al. 2020), and the frequency of affectionate touch (Debrot et al. 2021), with greater touch frequency linked to higher well‐being. Moreover, individuals high in attachment anxiety show reduced pleasantness discrimination between affective and non‐affective touch (Krahé et al. 2018), whereas high attachment avoidance predicts lower desire for touch (Jakubiak et al. 2021).

Based on these findings, we evaluated effects of SST on stress, fatigue, and loneliness. To do so, we generated longitudinal data using psychological self‐report scales and Ecological Momentary Assessment (EMA) that captured both pre‐to‐post session changes and cumulative changes leading to reduced recall bias and increased ecological validity (Shiffman et al. 2008; Trull and Ebner‐Priemer 2013). Our study aimed to test the feasibility of the intervention and research design as a pilot for follow‐up studies that will incorporate biological stress and inflammation markers with the ultimate aim to provide evidence of an intervention that mitigates negative effects of chronic stress.

We hypothesised that our two‐week SST intervention would significantly reduce retrospective self‐reported stress (H1a) and momentary stress (with all momentary measures assessed by EMA; H1b), retrospective self‐reported fatigue (H2a) and momentary fatigue (H2b), as well as retrospective self‐reported loneliness (H3a) and momentary loneliness (H3b). For EMA outcomes, effects were defined as within‐session pre‐post change (Group × Measurement) and a potential cumulative day‐to‐day change (Group × Measurement × Day). We also hypothesised that changes in momentary stress and fatigue would be moderated by attachment anxiety (H4a/H4b) and attachment avoidance (H5a/H5b). We used a time‐ and expectation‐matched meditation as an active control group to control for non‐specific effects (e.g., relaxation, expectancy, routine), allowing us to test whether SST provides incremental benefits on the variables in question.

2. Method

2.1. Participants

We conducted simulation‐based power analyses in RStudio (Version 2025.05.0 + 496, Posit Team 2025) with R (Version 4.5.0, R Core Team 2025) to determine the sample size required for detecting a three‐way interaction (Group × Measurement × Moderator), assuming a medium effect size of d = 0.5, an alpha of 0.05%, and 80% power, resulting in 75 participants.

Only chronically stressed individuals with a PSS‐10 score of > 12.79 (weighted median value of the population between 20 and 59 years of age; Klein et al. 2016) were included. We excluded all individuals fulfiling at least one of the following criteria, based on our prescreening questionnaire: younger than 18 years of age; incapable of giving consent; hearing‐impaired; present alcohol abuse (> 15 drinks per week on average for men and > 8 for women); use of cannabis or psychotropic medication within the past 2 weeks; use of illicit drugs within the last year; substance‐induced disorder (active or in remission < 2 years); eating disorder (active or in remission < 5 years); current or past record of bipolar disorder, schizophrenia, or borderline personality disorder. From April to June 2025, Participants were recruited via the University's recruiting system in exchange for partial course credit. Participation was voluntary and written consent was obtained in person.

2.2. Procedure

Our study followed a randomised controlled design with two groups and evaluated a 14‐day daily 5‐min SST intervention compared to a meditation practice of the same length in time. First, participants completed an online prescreening questionnaire to assess eligibility and were then randomly assigned to either self‐soothing touch (intervention group) or meditation (control group) using computerised block randomisation (block size 4) with a randomisation ratio of 1:1 using online randomisation. After randomisation, participants attended their first on‐site session at the University, where they received in‐person instructions for performing their assigned activity and were encouraged to complete all daily assessments that followed during the intervention period. During the following 14‐day intervention period, participants engaged daily in their instructed activity remotely. One day after finishing the intervention, participants were asked to complete the postintervention questionnaire; 4 weeks later, they were asked to fill out the follow‐up questionnaire. The study procedure together with all constructs measured is depicted in Figure 1.

FIGURE 1.

FIGURE 1

Open in a new tab

Study Design. Participants completed retrospective questionnaires at baseline (day 1), postintervention (day 16), and follow‐up (day 44). During the intervention period (days 2–15), participants practiced their assigned activity (SST or minimal instruction meditation) and completed pre‐ and post‐session Ecological Momentary Assessment (EMA) using a visual analogue scale (VAS; 0–100). Icons made by Kiranshastry and Freepik from www.flaticon.com.

3. Intervention

Comparability of the control and intervention condition was ensured by both procedures being performed for 5 min after waking up, following instruction audios guided by the same voice. During on‐site briefing participants were informed that the exercise was a self‐touch intervention to orient their attention toward the tactile component in the subsequent practice. In the intervention group, participants placed their right hand on the heart and their left hand on the abdomen (Dreisoerner et al. 2021) and were asked to close their eyes, take deep breaths and direct their focus to the sensation of touch ‐ to feel the pressure, weight, and warmth of their hands. Every 15–20 s, refocus cues were provided, for example: ‘come back to the feeling of your hands on your body’. The intention was to foster a comforting, warm awareness of self‐touch, with the participants' attention directed toward the sensory experience.

Participants in the control group performed an exercise that was originally conceptualised as sham mindfulness meditation (adapted from Zeidan et al. 2010). This procedure was intended to make participants believe they were engaging in mindfulness meditation to control for expectation effects and to isolate the effects of SST. This specific protocol was chosen because Zeidan et al. (2010) demonstrated that participants believed they were truly meditating, indicating high credibility. Additionally, this protocol allowed us to match the duration, sensory reduction, and ritualised daily practice of the intervention. To ensure comparable expectation, participants were additionally told during on‐site briefing, that they will be engaging in a mindfulness practice. Unlike in traditional mindfulness meditation or in the present SST intervention, participants were not explicitly instructed on where to direct their focus. Instead, they were just instructed to close their eyes and told to ‘take deep breaths as you meditate’, repeated every minute. However, analyses of the present study indicated that this exercise may have produced effects similar to actual meditation. Together with recent findings, showing that the same sham protocol increased state mindfulness (Davies et al. 2025), we assume that the task likely contains active components. We therefore refer to it as meditation in the following to avoid misrepresenting it as purely inactive sham.

3.1. Measures

3.1.1. EMA Measurements

To capture potential immediate effects of each session as well as day‐to‐day fluctuations in stress, fatigue, and loneliness, we used EMA. EMA enables participants to rate their current experience in real‐time (Shiffman et al. 2008) rather than relying on retrospective reporting. Participants were asked: ‘How stressed/fatigued/lonely are you feeling right now?’ immediately before and after the intervention. Answers were rated on a visual analogue scale (VAS) from 0 (not at all) to 100 (very much), depicting the constructs as a continuum without relying on arbitrary categorisation. This allowed us to detect even small amounts of variation over time, and daily fluctuations (Paul‐Dauphin et al. 1999). The EMA survey was embedded with the audio‐guided intervention into a single platform that guided participants through the entire procedure from start to finish. To increase adherence, participants received a daily email containing a link to the survey and intervention.

3.1.2. Retrospective Self‐Report Measurements

Complementing the EMA data, participants also completed standardised self‐report measures to observe intermediate effects. Retrospective self‐reported stress was assessed by the German 10‐item version of the Perceived Stress Scale (PSS‐10; Cohen et al. 1983; Schneider et al. 2020). Participants were asked about the prevalence of certain feelings and thoughts over the past month on a 5‐point Likert scale, with responses ranging from 1 (never) to 5 (very often). One example item is: ‘In the last month, how often have you felt difficulties were piling up so high that you could not overcome them?’. The PSS‐10 showed high internal consistency (Cronbach's α = 0.84) and good validity.

To assess retrospective self‐reported fatigue, we used the Multidimensional Fatigue Inventory (MFI‐20; Westenberger et al. 2022). It consists of 20 items (e.g., ‘I feel fit’) grouped into 5 subscales: general fatigue, physical fatigue, reduced activity, reduced motivation, and mental fatigue. Responses are given on a 5‐point Likert scale ranging from 1 (Yes, i.e. true) to 5 (No, i.e. not true). Cronbach's α values for the subscales ranged from 0.79 to 0.90, indicating high reliability. Satisfactory convergent validity of the MFI‐20 has been previously demonstrated (Westenberger et al. 2022).

Retrospective self‐reported loneliness was measured using the 20‐item German version of the UCLA Loneliness Scale (Döring and Bortz 1993). Respondents rate items—for example, ‘I feel isolated from others’—on a four‐point Likert scale ranging from 1 (never) to 4 (often). Internal consistency was considered high, with Cronbach's α values reported between 0.89 and 0.94 (Russell 1996). Convergent validity has been demonstrated through associations with measures of social support, depression, and self‐esteem, indicating good construct validity.

Attachment avoidance and attachment anxiety were measured using the German version of the Experiences in Close Relationships Scale (ECR‐G‐10; Neumann et al. 2023). It contains 10 items and is based on two dimensions: Attachment avoidance and attachment anxiety. It uses a 7‐point Likert scale from 1 (strongly disagree) to 7 (strongly agree), and both scales showed acceptable‐to‐high internal consistency (Attachment avoidance: Cronbach's α = 0.75–0.81, Attachment anxiety: Cronbach's α = 0.72–0.77) and are independent of each other. Correlational evidence supports its construct validity (Neumann et al. 2024).

Covariates assessed at baseline included age in years, gender (male/female/diverse/other), substance use (daily average consumption of alcohol, caffeine, nicotine in standardised units), the presence of a major stressor in the last 4 weeks (yes/no), and levels of self‐compassion (mean score on the Self‐Compassion Scale [SCS‐D, Cronbach's α = 0.91]; Hupfeld and Ruffieux 2011). Covariates assessed at posttest included the liking of the intervention that was assessed using a 6‐point Likert scale 1 (do not agree at all) to 6 (fully agree) for seven distinct adjectives (e.g., calming, unsettling; Dreisoerner et al. 2021), as well as intervention fidelity, assessed with a single item (‘On how many days did you follow the intervention as instructed?’).

3.2. Data Analysis

For data cleaning and analysis, RStudio (Version 2025.05.0 + 496, Posit Team 2025) with R (Version 4.5.0, R Core Team 2025) was used. All analyses followed intention‐to‐treat principles. We fitted linear mixed‐effects models for both questionnaire and EMA outcomes. Data missing due to incomplete assessments were retained through maximum‐likelihood estimation. Satterthwaite approximations were used for estimating degrees of freedom. All models included Group and Measurement as fixed effects, together with grand‐mean centred covariates listed above. Importantly, the factor Measurement denoted different aspects depending on the data source: For questionnaires with retrospective self‐report measures, it referred to study phase (baseline, postintervention, follow‐up), whereas for EMA it indicated daily ratings immediately before and after each practice (pre‐ and post‐session). For EMA outcomes, Day (1–14) and its interactions were additionally considered to capture cumulative trends. Random intercepts for participants were included in all models.

For hypothesis testing, interaction models were compared to main‐effects models via likelihood ratio tests. To test moderation, Attachment Anxiety and Avoidance were entered individually as interaction terms in two separate models to avoid multicollinearity; each model was compared against the respective model without the interaction involving the moderator. Final models were refitted with restricted maximum likelihood (REML) and estimated marginal means (EMMs) with planned contrasts, simple slopes, and linear trends were calculated. When analyses were not preregistered as planned contrasts, we applied false discovery rate (FDR) correction to adjust for multiple testing (Benjamini and Hochberg 1995) and reported Cohen's d as a measure of effect size.

4. Results

The CONSORT chart in Figure 2 provides an overview of participant flow throughout the study, with a total of 78 individuals recruited and randomised. Participant characteristics of the intention‐to‐treat sample are provided in Table 1. Participants did not differ on most baseline variables, except for attachment avoidance. In the following, we examined how the intervention affected levels of stress, fatigue, and loneliness, and whether these effects were moderated by distinct dimensions of attachment.

FIGURE 2.

FIGURE 2

Open in a new tab

CONSORT Flow Diagram. Adapted from Hopewell et al. (2025), licenced under CC BY 4.0.

TABLE 1.

Baseline characteristics of the intention‐to‐treat sample.

Variable Total N = 78 Intervention group (n = 39) Control group (n = 39) p
M (SE)/n (%) M (SE)/n (%)
Gender > 0.99
Female 63 (80.8%) 31 (79.5%) 32 (82.1%)
Male 14 (17.9%) 7 (17.9%) 7 (17.9%)
Diverse 1 (1.3%) 1 (2.6%) 0 (0%)
Age (in years) 22.2 (0.3) 22.1 (0.5) 22.2 (0.5) 0.82
Major stressor 20 (25.6%) 10 (25.6%) 10 (25.6%) > 0.99
Alcohol 0.42 (0.06) 0.43 (0.09) 0.40 (0.08) 0.83
Caffeine 1.03 (0.10) 1.00 (0.14) 1.07 (0.15) 0.72
Nicotine 0.53 (0.27) 0.85 (0.52) 0.22 (0.13) 0.24
Stress (PSS‐10) 28.01 (0.69) 27.85 (0.84) 28.18 (1.10) 0.81
Fatigue (MFI‐20) 2.61 (0.08) 2.57 (0.11) 2.66 (0.11) 0.56
Loneliness (UCLA) 3.15 (0.01) 3.15 (0.02) 3.15 (0.02) 0.85
Attachment anxiety 3.81 (0.15) 3.88 (0.15) 3.75 (0.25) 0.66
Attachment avoidance 2.32 (0.14) 2.04 (0.18) 2.59 (0.20) 0.038*
Self‐compassion 3.11 (0.06) 3.18 (0.08) 3.03 (0.09) 0.20
Fidelity 13.00 (0.23) 13.15 (0.26) 12.85 (0.39) 0.51
Open in a new tab

Note: Values are presented as means (M) with standard errors (SE) for continuous variables and as frequencies with percentages (n, %) for categorical variables. Group comparisons for continuous variables were conducted using independent‐samples t‐tests. For categorical variables, chi‐squared tests were used; for gender, analyses were restricted to male and female participants due to the small number of participants identifying as diverse.

4.1. Stress

Looking at PSS‐10 scores indicating retrospective self‐reported stress (H1a), the interaction between Group and Measurement, tested alongside covariates, did not significantly improve model fit and, thus, support a role of the intervention, compared to the main‐effects model, X2 (1) = 3.03, p = 0.082. The interaction model explained 52.2% of the marginal variance (ICC = 0.290), 0.6% more than the model without the interaction (ICC = 0.266), providing no evidence for an intervention effect. Figure 3 (Panel A) shows levels of retrospective self‐reported stress (PSS‐10) per group and measurement. Planned comparisons revealed that follow‐up scores of the control group relative to baseline did not significantly change, b = −0.69, SE = 0.79, t(77.7) = −0.87, p = 0.386, d = −0.20. The same was true for the intervention group, b = 1.28, SE = 0.81, t(79.7) = 1.59, p = 0.117, d = 0.37.

FIGURE 3.

FIGURE 3

Open in a new tab

Group Differences in Retrospective Self‐Reported Stress, Loneliness, and Fatigue. Estimated marginal means (± standard errors) for retrospective self‐reported perceived stress (A), fatigue (B), and loneliness (C) across measurement points in the control (red) and intervention (blue) groups. Error bars represent 95% confidence intervals.

Fitting models on the EMA data showed a similar picture, with the Group × Measurement model (H1b) not significantly improving model fit compared to the main‐effects model, X2 (1) = 1.54, p = 0.214, explaining 16.6% compared to 16.7% of marginal variance (ICCGxM = 0.396, ICCmain = 0.396); the same was true for the model involving the threefold interaction between Group, Measurement, and Day ‐ that additionally takes into account changes over the timecourse of the interventions ‐ compared to the Group × Measurement model, X2 (3) = 4.67, p = 0.197, R2 marg_GxMxD = 16.7%, ICCGxMxD = 0.397. Figure 4 (Panel A) displays momentary stress levels by group across the intervention days, separately for measurements taken immediately before and after each daily practice.

FIGURE 4.

FIGURE 4

Open in a new tab

EMA Measures Over the Course of the Intervention Period. Estimated marginal means for momentary stress, fatigue, and loneliness across the 14‐day intervention period in the intervention and control groups, measured with visual analogue scales (0–100). Blue solid lines represent daily pre‐session, orange dashed lines daily post‐session estimates; shaded ribbons indicate 95% confidence intervals.

Planned contrasts showed a significant decrease in momentary stress levels in the intervention group, b = −0.41, SE = 0.08, t(1950) = −4.79, p < 0.001, d = −0.30, and the control group, b = −0.56, SE = 0.09, t(1950) = −6.43, p < 0.001, d = −0.41, indicating that SST and meditation both decreased momentary stress levels pre‐to‐post session, with no significant differences between groups, b = −0.15, SE = 0.12, t(1950) = −1.25, p = 0.212, d = −0.12. This effect decreased in magnitude over the course of the intervention, as visible in Figure 4 (Panel A). Looking at estimated trends over time, there was a significant decrease in pre‐session, but not post‐session stress levels over the course of 2 weeks in both control and intervention groups; however, after correcting for multiple comparisons, the pre‐session effect of the intervention group just failed to reach significance (p = 0.052).

Notably, self‐compassion emerged as a significant covariate for momentary stress, predicting lower PSS‐10 scores, b = −5.39, SE = 0.88, t(75.34) = −6.10, p < 0.001, and lower momentary stress, b = −0.68, SE = 0.30, t(77.02) = −2.24, p = 0.028.

Taken together, these findings provide no evidence that the intervention reduced retrospective self‐reported stress compared to meditation, thus not supporting H1a. Our EMA results show that both SST and meditation were effective in acutely reducing stress levels, but there was no evidence that SST outperformed meditation, providing no support for H1b.

4.2. Fatigue

Our models did not indicate an influence of our intervention on retrospective self‐reported fatigue scores operationalised by MFI‐20 sum score (H2a), as the interaction between Group and Measurement did not significantly improve model fit compared to the main‐effects model, X2 (2) = 0.46, p = 0.794, with the interaction model explaining 22.6% marginal variance (ICC = 0.634), only 0.1% more than that of the main‐effects model (ICC = 0.633). Figure 3 (Panel B) shows MFI‐20 scores at baseline, postintervention, and at follow‐up. Differences between groups relative to baseline were neither significant at postintervention, b = 0.07, SE = 0.12, t(154.26) = 0.60, p = 0.552, nor at follow‐up, b = 0.07, SE = 0.12, t(154.66) = 0.58, p = 0.563. Planned comparisons revealed no differences within each group from baseline to postintervention or baseline to follow‐up, all ps > 0.441.

Our EMA data also revealed no effects of SST compared to meditation on momentary fatigue (H2b), as indicated by the nonsignificant model comparison between the interaction and the main‐effects model, X2 (3) = 0.23, p = 0.625 (R2 marg_GxM = 10.8% vs. R2 marg_main = 10.9%, ICCGxM = 0.472 vs. ICCMain = 0.472), as well as between the threefold interaction model (R2 marg_GxMxD = 10.9%, ICCGxMxD = 0.473) and the Group × Measurement model, X2 (3) = 5.78, p = 0.123. Planned contrasts indicated a significant decrease in momentary fatigue pre‐to‐post session in both intervention, b = −0.37, SE = 0.09, t(1956) = −4.15, p < 0.001, d = −0.27, and control group, b = −0.33, SE = 0.09, t(1956) = −3.65, p < 0.001, d = −0.24, with differences between groups not being significant, b = 0.04, SE = 0.13, t(1956) = 0.31, p = 0.761, d = 0.04. Post‐hoc comparisons indicated a significant decrease in pre‐session fatigue for the intervention group, while this was not the case for the control group (see Table 2). Nevertheless, group differences failed to reach significance, b = −0.03, SE = 0.22, t(1950) = −1.32, p = 0.187.

TABLE 2.

Changes in momentary stress and fatigue within groups over time.

Contrast Momentary stress Momentary fatigue
b SE t(1950) p b SE t(1956) p
Intervention: Pre‐session −0.03 0.15 −2.23 0.052 −0.06 0.02 −3.95 < 0.001*
Intervention: Post‐session −0.21 0.15 −1.39 0.220 −0.03 0.02 −1.97 0.066
Control: Pre‐session −0.05 0.15 −3.63 0.001* −0.03 0.02 −2.01 0.066
Control: Post‐session −0.01 0.15 −0.74 0.457 −0.01 0.02 −0.51 0.609
Open in a new tab

Note: Values represent estimated marginal trends showing mean change per day (1–14) in momentary stress and fatigue separately for each group pre‐ and post‐session. Reported estimates (b) are unstandardised coefficients from linear mixed‐effects models; p‐values are FDR‐corrected.

Abbreviation: SE, standard error.

Taken together, the intervention did not reduce retrospective self‐reported fatigue (MFI‐20) compared to the control condition, and H2a was not supported. For momentary levels of fatigue (EMA), both groups showed significant pre‐to‐post session reductions, but these effects did not differ between intervention and control group, providing no support for H2b. However, unlike the meditation intervention, the SST intervention significantly reduced pre‐session levels of momentary fatigue over the course of the intervention.

4.3. Loneliness

Linear mixed‐effects models of UCLA Loneliness Scale mean scores revealed no intervention effect on retrospective self‐reported loneliness (H3a), as the Group × Measurement model did not fit significantly better than the main‐effects model, X2 (2) = 0.60, p = 0.741, explaining 7.0% of the marginal variance (ICC = 0.419), virtually identical to the main‐effects model (R2 marg = 6.8%, ICC = 0.418). Differences between groups relative to baseline were not significant at postintervention, b = −0.00, SE = 0.13, t(153.76) = −0.03, p = 0.998, and at follow‐up, 0.02, SE = 0.03, t(154.46) = 0.67, p = 0.502. Planned within‐group comparisons from baseline to postintervention and to follow‐up yielded one significant effect: a decrease in loneliness from baseline to follow‐up in the control group, b = −0.05, SE = 0.02, t(158) = −2.17, p = 0.031, d = −0.50, while all other effects were nonsignificant (all ps > 0.231). Figure 3 (Panel C) shows loneliness scores across measurement points.

Compared to stress and fatigue, a slightly different picture emerged when looking at EMA results concerning momentary loneliness (H3b; see Figure 4, Panel C): the model including the Group × Measurement × Day interaction appeared as the best‐fitting model (R2 marg = 11.0%, ICC = 0.438), predicting data significantly better than the Group × Measurement model (R2 marg = 10.5%, ICC = 0.436), X2 (3) = 38.92, p < 0.001, and the main‐effects model (R2 marg = 10.5%, ICC = 0.436), X2 (4) = 19.00, p < 0.001. However, this was mainly driven by the significant interaction of Group × Day, b = 0.43, SE = 0.17, t(1945) = 2.57, p = 0.010, while the threefold interaction between Group × Measurement × Day was not significant, b = 0.15, SE = 0.24, t(1944) = 0.63, p = 0.530.

Momentary loneliness decreased independent of Measurement (pre‐vs. post‐session) in the control group across the 14 days, b = −0.49, SE = 0.08, t(1951) = −5.76, p < 0.001, while it did not significantly change in the intervention group, b = 0.02, SE = 0.08, t(1951) = 0.22, p = 0.828, showing that only the meditation, contrary to SST, decreased cumulative levels of loneliness. Planned comparisons indicated a significant pre‐to‐post session decrease in momentary loneliness in both the intervention, b = −1.70, SE = 0.67, t(1950) = −2.54, p = 0.011, d = −0.16, and the control group, b = −1.98, SE = 0.69, t(1950) = −2.89, p = 0.004, d = −0.18, with no significant group differences, b = 0.28, SE = 0.96, t(1950) = 0.29, p = 0.772, d = 0.03. Higher levels of self‐compassion again led to a decrease in momentary loneliness levels, b = −7.48, SE = 2.59, t(77.23) = −2.89, p = 0.005, but not of retrospective self‐reported loneliness, b = −0.01, SE = 0.03, t(77.43) = −0.20, p = 0.846.

Overall, there was no evidence that the intervention reduced retrospective self‐reported loneliness compared to the control condition; thus, H3a was not supported. Our EMA results showed a small but significant pre‐to‐post session reduction in momentary loneliness in both groups, with effects of SST comparable to meditation and no group differences, providing no support for H3b. Meditation was superior to our SST intervention in decreasing momentary loneliness levels over the course of the 14‐day‐intervention.

4.4. Moderation by Attachment Anxiety and Attachment Avoidance

When examining whether Attachment Anxiety or Avoidance moderated the effect of the intervention on momentary stress, neither moderator reached significance (all ps > 0.328). We also tested whether Attachment Anxiety and Avoidance moderated the intervention's effects on momentary fatigue: While our data did not indicate a moderating effect of Attachment Anxiety, X2 (7) = 3.08, p = 0.878, there was a moderation effect of Attachment Avoidance: The model including the interaction Group × Measurement × Day × Attachment Avoidance revealed the best fit, explaining significantly more variance than the model with the threefold interaction and Attachment Avoidance as a fixed effect, X2 (7) = 18.34, p = 0.011 (R2 Anxiety_mod = 11.6, ICCAnxiety_mod = 0.473; R2 Anxiety_main = 11.0, ICCAnxiety_main = 0.472). For the moderating effect of Attachment Avoidance, planned contrasts indicated that the change in momentary fatigue across the 14 days varied significantly with avoidance level in the intervention group, F(2, 1956.5) = 16.49, p < 0.001, indicating that reductions in fatigue became stronger with higher avoidance. No such moderation was observed in the control group, F(2, 1956.5) = 2.20, p = 0.111. The corresponding day trends and their 95% confidence intervals are displayed in Figure 5. Beyond the preregistered three‐way interaction, we included Day as an exploratory factor in the model, as preliminary analyses indicated systematic changes across days. As the study was formally powered for the three‐way interaction, the moderation analyses involving Day should be interpreted as exploratory.

FIGURE 5.

FIGURE 5

Open in a new tab

Moderating Effects of Attachment Avoidance on the Intervention's Effect on Fatigue. Estimated marginal means by group, intervention day, and baseline attachment avoidance. Attachment avoidance is displayed at the sample mean (blue solid lines), one standard deviation below the sample mean (–1 SD; red dotted lines), and one standard deviation above the mean (+1 SD, green dashed lines). Shaded areas indicate ± 95% CI.

When comparing for between‐group differences, we observed higher decreases in fatigue in the intervention group than in the control group at mean levels of Attachment Avoidance, b = −0.04, SE = 0.02, t(1957) = −2.23, p = 0.026, d = −0.33, and +1 SD, b = −0.10, SE = 0.02, t(1957) = −4.09, p < 0.001, d = −0.89, whereas no differences were observed at −1 SD, b = −0.02, SE = 0.02, t(1957) = 1.12, p = 0.265, d = 0.23. In the control group, a significant change was found only at −1 SD, with no effects at mean or +1 SD levels.

In summary, Attachment Anxiety did not moderate momentary stress (H4a) or fatigue (H4b) and Attachment Avoidance did also not moderate momentary stress (H5a), providing no support for these hypotheses. By contrast, Attachment Avoidance significantly moderated momentary fatigue (H5b): participants higher in Avoidance exhibited stronger reductions in fatigue over time in the intervention group, whereas no moderation was observed in the control group. Thus, H5b was supported.

5. Discussion

The present study investigated acute and cumulative effects of a two‐week self‐soothing touch intervention compared to a meditation control on self‐reported measures of stress, fatigue, and loneliness. Both groups showed immediate reductions in pre‐to‐post session momentary stress, fatigue, and loneliness measured by EMA, with SST being comparable to meditation but differences between groups not being significant. SST participants also showed a highly significant cumulative reduction in pre‐session momentary fatigue over the intervention period. This was moderated by attachment avoidance, with higher avoidance predicting stronger reductions across the 14 days only in the SST group. Contrary to EMA‐based results, neither SST nor meditation had a significant effect on any retrospective self‐reported measure (perceived stress, fatigue, and loneliness). It should be noted that the control was initially designed as a sham mindfulness meditation to control for expectancy effects. However, consistent with our findings and recent studies using the same protocol (Davies et al. 2025), it appeared to elicit effects comparable to brief meditation.

5.1. Acute and Cumulative Beneficial Effects of SST and Meditation on Stress

The intervention did not decrease retrospective stress compared to meditation (H1a). This result aligns with Dreisoerner et al. (2021), who reported that a single SST session did not decrease self‐reported stress, albeit decreasing levels of the stress biomarker cortisol. There was a (nonsignificant) trend increase in stress in the SST group compared to the control group, most probably present due to elevated levels of academic stress in our student sample, as follow‐up questionnaires were assessed during examination period. The absence of this rise in the control group suggests meditation may have exerted a small stress‐buffering effect, as also reported in previous research (Manigault et al. 2021; Morton et al. 2020). Larger samples are needed to determine whether this effect would reach statistical significance.

Contrary to retrospective measures, EMA results on momentary stress indicated a significant pre‐to‐post session reduction of both SST and meditation, in line with prior research suggesting that repeated and prolonged practice amplifies positive effects of touch‐based interventions (Packheiser et al. 2024) and that even brief contemplative elements can influence levels of arousal (Tang et al. 2015). The lack of between‐group differences expected (H1b) are therefore not surprising against an active control incorporating elements of mindfulness (e.g., breath awareness, sitting still) that can lower sympathetic arousal and perceived strain. In addition, simply spending 5 minutes a day in quiet rest without stimulation ‐ as in both interventions ‐ may itself have contributed, as even brief periods of unguided rest are psychologically potent and can alter subjective experience (Wilson et al. 2014). This is also a potential reason why our meditation protocol, originally designed as sham intervention (Zeidan et al. 2010) ‐ a neutral control accounting for nonspecific factors such as time, attention, expectancy, instructor effects, and the ritual of daily practice ‐ seemed to contribute to outcomes in ways that went beyond controlling for nonspecific influences.

Exploratively looking at cumulative effects on momentary stress, we saw a highly significant decrease in pre‐session levels in the control group and a trend decrease in the SST group after correcting for multiple comparisons (p = 0.052). Of note, between‐group differences were not significant for momentary stress, suggesting that the cumulative improvements were largely driven by nonspecific components shared by both practices (e.g., paced breathing, expectancy) rather than SST‐specific mechanisms.

5.2. Attachment Avoidance Moderates Cumulative Fatigue Reductions in Response to SST

While there were no effects of the intervention on retrospective measures of fatigue (H2a), momentary levels of fatigue (H2b) decreased similarly to stress in both groups. Fatigue is tightly coupled to stress physiology and motivational regulation, as high levels of glucocorticoids affect alertness and vigilance (Joëls and Baram 2009) while also potentially contributing to increased ANS responses to stress (Koob 1999) that can increase inflammation and fatigue (deRijk et al. 1994). Research on SST as potential intervention for fatigue reduction is lacking, although previous studies have shown beneficial effects of mindfulness trainings (Park et al. 2024) and touch‐based interventions on fatigue (Karagozoglu and Kahve 2013; Packheiser et al. 2024). It is therefore consistent with our expectations that decreases in momentary stress parallelled decreases in momentary fatigue, suggesting that both interventions engaged overlapping mechanisms of short‐term stress regulation. Affiliative self‐touch may contribute to downregulation of arousal via mechanoreceptors stimulation and parasympathetic engagement buffering stress (Löken et al. 2009; Morrison 2016) and thereby dampening HPA‐axis responses (Dreisoerner et al. 2021) linked to lower perceived fatigue. Crucially, our results have shown that individuals high in attachment avoidance—that are known to report lower desire for interpersonal touch (Jakubiak et al. 2021) while also appraising it differently (Krahé et al. 2018; Spitoni et al. 2020)—particularly profit from SST, as higher levels of attachment avoidance predicted larger decreases in momentary fatigue in the present study. For these individuals, self‐delivered touch may be more acceptable than receiving touch from others, making it possible to engage in a (self‐)compassion without interpersonal discomfort (Debrot et al. 2021).

Cumulative effects showed a similar pattern for momentary fatigue as for stress but reversed across groups: SST produced a highly significant pre‐session decrease in fatigue across the 14 days, while the effects in the meditation group failed to reach significance (albeit between‐group differences were nonsignificant). Thus, SST can be seen as a promising and easy‐to‐implement alternative to existing and well‐established mindfulness practices.

5.3. Loneliness: Comparable Acute Effects; Cumulative Effect Restricted to Meditation

Both SST and meditation failed to reduce retrospectively assessed loneliness (H3a) but have shown highly significant acute decreases in momentary loneliness (H3b). As for cumulative effects, we found a comparatively strong decrease in both pre‐ and post‐session in the meditation, but not the SST condition, showing that even short daily sessions of meditation can induce important changes on aspects of psychological wellbeing. While SST elicited short‐term reductions in loneliness, consistent with findings on social touch (Heatley Tejada et al. 2020; Noone and McKenna‐Plumley 2022), it showed no lasting cumulative effects, likely due to the absence of interpersonal interaction and social function of external touch, emphasised as central to the effectiveness in those studies. In contrast, we observed both acute and cumulative effects of meditation, aligning with prior studies showing that mindfulness interventions can reduce loneliness (Saini et al. 2021) and improve well‐being (Lahtinen and Salmivalli 2020).

The discrepancy between retrospective measures and EMA that was observed in all outcomes of this study likely reflects differences in measurement sensitivity, influenced by timeframe: The UCLA Loneliness Scale taps into more global, trait‐like perceptions that typically shift slowly and are vulnerable to recall/averaging biases (Shiffman et al. 2008; Trull and Ebner‐Priemer 2013), whereas VAS ratings in EMA offer fine‐grained, state‐level resolution and are more sensitive to small changes (Paul‐Dauphin et al. 1999). Similar mechanisms could be at play for the other retrospective measures assessed within this study, albeit to a different degree: While the MFI‐20 is rather intended to capture aggregated (tonic) fatigue over a longer time period (Westenberger et al. 2022), the PSS‐10 explicitly references the past four weeks (Cohen et al. 1983); nevertheless, it is still prone to biases inherent to retrospective measures and provides considerably less data than EMA. Accordingly, all retrospective questionnaires reflect global cognitive appraisal over several weeks and typically require sustained behavioural or appraisal change (Cohen et al. 1983; Shiffman et al. 2008). This is particularly relevant in our study because follow‐up questionnaires were completed during examination period, when overall stress levels in students are typically elevated.

Short‐term effects may also reflect rapid, physiologically mediated mechanisms characteristic of affective touch (Morrison 2016) and self‐touch (Dreisoerner et al. 2021), which reduce momentary stress but do not necessarily produce sustained changes in cognitive appraisal. Brief meditation exercises—such as the minimal‐instruction meditation and SST intervention used here—have similarly been shown to produce primarily short‐term, state‐level changes (Davies et al. 2025). In contrast, longer‐term improvements in stress and well‐being have been demonstrated in mindfulness‐ and compassion‐based programs lasting 4–8 weeks (Neff and Germer 2013). Therefore, the 14‐day duration of the present study was likely sufficient to elicit acute improvements but not to yield changes detectable in retrospective measures.

5.4. The Role of Self‐Compassion in Decreasing Feelings of Stress, Fatigue, and Loneliness

In our data, higher self‐compassion predicted lower stress outcomes ‐ both retrospectively and momentarily ‐ in line with previous findings (Han and Kim 2023; Neff 2003, 2023). This also applied to retrospectively assessed fatigue and momentary loneliness, consistent with studies reporting negative associations between self‐compassion and fatigue symptoms (Behilak et al. 2024; Babenko et al. 2019; Eriksson et al. 2018) and loneliness (Akin 2010; Ghezelseflo and Mirza 2020). Self‐compassion may have been a key mechanism why our intervention had comparable effects to meditation in decreasing stress and fatigue, as previous research suggests that a compassionate mindset during practice may be necessary for soothing touch to be effective (Neff and Germer 2013). However, the absence of cumulative effects of SST on loneliness—despite strong associations between self‐compassion and momentary loneliness—suggests that self‐compassion alone may not be sufficient to impact this domain long‐term, or that additional factors (e.g., social–evaluative focus of meditation practices) contributed more directly to loneliness reduction in the control condition.

5.5. Strengths, Limitations, and Future Directions

Albeit the present study has several strengths ‐ including (a) combining EMA with standardised questionnaires to capture both acute within‐session changes and cumulative trends, (b) an expectation‐ and time‐matched active control, and (c) a simple five‐minute at‐home protocol supporting adherence—it is not without limitations. Foremost, a central limitation concerns our control condition: Although conceived as sham intervention modelled after Zeidan et al. (2010), it appeared to exert own active effects rather than functioning as an inert comparator. In other words, what was intended as a non‐meditative control seemingly incorporated meditative elements, limiting its suitability as a sham. To provide the correct conceptual framework for interpreting the results, we refer to it as meditation rather than a sham throughout the manuscript. We included an active rather than an inactive control group because active comparators help balance expectancy and credibility. Future research may incorporate both inactive and active controls, while recognising that inactive controls can artificially inflate observed effects.

Crucially, as our study was designed as a pilot exploring feasibility of the research design for future larger‐scale investigations, stress outcomes were solely assessed through self‐report measures susceptible to biases. Expectancy effects may also have contributed, as participants were primed to believe they were engaging in a potentially beneficial practice. Future trials should therefore incorporate objective stress markers (e.g., cortisol, alpha‐amylase, and/or HRV) and larger samples to strengthen validity and reliability.

Since our secondary hypotheses concerning moderation involved analyses with four‐way interactions, a larger sample size would have been needed to achieve adequate statistical power (even though the moderation effect of attachment avoidance could still be demonstrated). The study was powered to detect the pre‐registered three‐way interaction (Group × Measurement × Attachment Avoidance), for which the sample size was appropriate, but not to detect the exploratory four‐way interaction additionally involving Day. Thus, the primary confirmatory analyses were adequately covered, whereas the moderated day trends should be interpreted as exploratory.

As we observed baseline differences in attachment avoidance, stratified randomisation on this variable is warranted in a successive confirmatory trial. In addition, although we included feasibility checks, these self‐reports cannot ensure procedural fidelity given that both SST and meditation could be performed anywhere. Finally, our sample predominantly consisted of young female university students, limiting its generalisability.

6. Conclusion

Our 14‐day brief self‐soothing touch intervention produced acute pre‐to‐post session reductions in momentary stress, fatigue, and loneliness; however, compared to a brief meditation intervention, we found no evidence of between‐group differences for these effects. SST also showed a cumulative decrease in pre‐session momentary fatigue across the intervention period, with this trajectory moderated by attachment avoidance: Participants higher in avoidance showed stronger fatigue reductions under SST, an effect not observed in the meditation group. None of the retrospective scales on perceived stress, fatigue, and loneliness (UCLA Loneliness Scale) showed group differences, indicating that the intervention did not produce longer‐term effects across any measured retrospective outcomes. SST thus appeared highly comparable to brief meditation in reducing momentary stress, fatigue, and loneliness. This indicates that SST serves as a feasible, low‐barrier intervention for short‐term, momentary stress regulation in daily life.

Funding

The study was preregistered on the Open Science Framework (https://osf.io/degtq) on April 1, 2025, approved by the ethics board of the University of Vienna (reference number 01299), and was conducted in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting Information S1

SMI-42-e70145-s001.docx (32.5KB, docx)

Acknowledgements

ChatGPT 5 (OpenAI 2025) was used to revise the language of the manuscript. Open Access funding provided by Medizinische Universitat Wien/KEMÖ.

Maier, Franziska , Luttenberger Ina, Dreisoerner Aljoscha, and Szaszkó Bence. 2026. “Self‐Soothing Touch Reduces Momentary Stress, Fatigue, and Loneliness Comparable to Brief Meditation: A Randomised Controlled Trial,” Stress and Health: e70145. 10.1002/smi.70145.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

References

  1. Adam, E. K. , Hawkley L. C., Kudielka B. M., and Cacioppo J. T.. 2006. “Day‐To‐Day Dynamics of Experience‐‐Cortisol Associations in a Population‐Based Sample of Older Adults.” Proceedings of the National Academy of Sciences of the United States of America 103, no. 45: 17058–17063. 10.1073/pnas.0605053103. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Akin, A. 2010. “Self‐Compassion and Loneliness.” International Online Journal of Educational Sciences 2, no. 3: 702–718. [Google Scholar]
  3. Andel, S. A. , Shen W., and Arvan M. L.. 2021. “Depending on Your Own Kindness: The Moderating Role of self‐Compassion on the Within‐Person Consequences of Work Loneliness During the COVID‐19 Pandemic.” Journal of Occupational Health Psychology 26, no. 4: 276–290. 10.1037/ocp0000271. [DOI] [PubMed] [Google Scholar]
  4. Arnsten, A. F. 2009. “Stress Signalling Pathways That Impair Prefrontal Cortex Structure and Function.” Nature Reviews Neuroscience 10, no. 6: 410–422. 10.1038/nrn2648. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Babenko, O. , Mosewich A. D., Lee A., and Koppula S.. 2019. “Association of Physicians’ self‐Compassion With Work Engagement, Exhaustion, and Professional Life Satisfaction.” Medical Science 7, no. 2: 29. 10.3390/medsci7020029. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Behilak, S. , Abdullah S., Ahmed G. K., and Saraya O. A. A. E. F. A.. 2024. “Influence of self‐Compassion on Fatigue and Psychological Wellbeing Among Psychiatric Nurses.” Egyptian Journal of Neurology, Psychiatry and Neurosurgery 60, no. 1: 113. 10.1186/s41983-024-00891-z. [DOI] [Google Scholar]
  7. Benjamini, Y. , and Hochberg Y.. 1995. “Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing.” Journal of the Royal Statistical Society: Series B 57, no. 1: 289–300. 10.1111/j.2517-6161.1995.tb02031.x. [DOI] [Google Scholar]
  8. Boksem, M. A. , and Tops M.. 2008. “Mental Fatigue: Costs and Benefits.” Brain Research Reviews 59, no. 1: 125–139. 10.1016/j.brainresrev.2008.07.001. [DOI] [PubMed] [Google Scholar]
  9. Cohen, S. , Kamarck T., and Mermelstein R.. 1983. “A Global Measure of Perceived Stress.” Journal of Health and Social Behavior 24, no. 4: 385–396. 10.2307/2136404. [DOI] [PubMed] [Google Scholar]
  10. Dantzer, R. , O'connor J. C., Freund G. G., Johnson R. W., and Kelley K. W.. 2008. “From Inflammation to Sickness and Depression: When the Immune System Subjugates the Brain.” Nature Reviews Neuroscience 9, no. 1: 46–56. 10.1038/nrn2297. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Davies, B. , Juden R., Greville J., and Hooper N.. 2025. “The Impact of Understanding and Openness on State Mindfulness Following a Brief Meditation: A Sham‐Controlled Study.” Mindfulness 16, no. 8: 2282–2291. 10.1007/s12671-025-02624-6. [DOI] [Google Scholar]
  12. Debrot, A. , Stellar J. E., MacDonald G., Keltner D., and Impett E. A.. 2021. “Is Touch in Romantic Relationships Universally Beneficial for Psychological Well‐Being? The Role of Attachment Avoidance.” Personality and Social Psychology Bulletin 47, no. 10: 1495–1509. 10.1177/0146167220977709. [DOI] [PubMed] [Google Scholar]
  13. Delaney, M. C. 2018. “Caring for the Caregivers: Evaluation of the Effect of an Eight‐Week Pilot Mindful Self‐Compassion (MSC) Training Program on Nurses’ Compassion Fatigue and Resilience.” PLoS One 13, no. 11: e0207261. 10.1371/journal.pone.0207261. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. deRijk, R. H. , Boelen A., Tilders F. J., and Berkenbosch F.. 1994. “Induction of Plasma interleukin‐6 by Circulating Adrenaline in the Rat.” Psychoneuroendocrinology 19, no. 2: 155–163. 10.1016/0306-4530(94)90005-1. [DOI] [PubMed] [Google Scholar]
  15. Döring, N. , and Bortz J.. 1993. “Psychometrische Einsamkeitsforschung: Deutsche Neukonstruktion Der UCLA Loneliness Scale [Psychometric Loneliness Research: German Reconstruction of the UCLA Loneliness Scale].” Diagnostica 39, no. 3: 224–239. 10.1037/t08689-000. [DOI] [Google Scholar]
  16. Dreisoerner, A. , Junker N. M., Schlotz W., et al. 2021. “Self‐Soothing Touch and Being Hugged Reduce Cortisol Responses to Stress: A Randomized Controlled Trial on Stress, Physical Touch, and Social Identity.” Comprehensive Psychoneuroendocrinology 8: 100091. 10.1016/j.cpnec.2021.100091. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Dunlavey, C. J. 2018. “Introduction to the Hypothalamic–Pituitary–Adrenal Axis: Healthy and Dysregulated Stress Responses, Developmental Stress and Neurodegeneration.” Journal of Undergraduate Neuroscience Education 16, no. 2: R59–R60. [PMC free article] [PubMed] [Google Scholar]
  18. Eriksson, T. , Germundsjö L., Åström E., and Rönnlund M.. 2018. “Mindful Self‐Compassion Training Reduces Stress and Burnout Symptoms Among Practicing Psychologists: A Randomized Controlled Trial of a Brief Web‐Based Intervention.” Frontiers in Psychology 9: 2340. 10.3389/fpsyg.2018.02340. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Esch, T. , Stefano G. B., Fricchione G. L., and Benson H.. 2002. “The Role of Stress in Neurodegenerative Diseases and Mental Disorders.” Neuroendocrinology Letters 23, no. 3: 199–208. [PubMed] [Google Scholar]
  20. Ghezelseflo, M. , and Mirza M.. 2020. “The Role of Self‐compassion in Predicting Loneliness and Self‐Efficacy in the Elderly.” Iranian Journal of Ageing 15, no. 2: 212–223. 10.32598/sija.13.10.630. [DOI] [Google Scholar]
  21. Han, A. , and Kim T. H.. 2023. “Effects of self‐Compassion Interventions on Reducing Depressive Symptoms, Anxiety, and Stress: A Meta‐Analysis.” Mindfulness 14, no. 7: 1553–1581. 10.1007/s12671-023-02148-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Heatley Tejada, A. , Dunbar R. I. M., and Montero M.. 2020. “Physical Contact and Loneliness: Being Touched Reduces Perceptions of Loneliness.” Adaptive Human Behavior and Physiology 6, no. 3: 292–306. 10.1007/s40750-020-00138-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Hopewell, S. , Chan A. W., Collins G. S., et al. 2025. “CONSORT 2025 Statement: Updated Guideline for Reporting Randomized Trials.” Nature Medicine 31, no. 6: 1776–1783. 10.1038/s41591-025-03635-5. [DOI] [PubMed] [Google Scholar]
  24. Hupfeld, J. , and Ruffieux N.. 2011. “Validierung Einer Deutschen Version der Self‐Compassion Scale (SCS‐D).” Zeitschrift für Klinische Psychologie und Psychotherapie 40, no. 2: 115–123. 10.1026/1616-3443/a000088. [DOI] [Google Scholar]
  25. Iob, E. , Kirschbaum C., and Steptoe A.. 2020. “Persistent Depressive Symptoms, HPA‐Axis Hyperactivity, and Inflammation: The Role of Cognitive‐Affective and Somatic Symptoms.” Molecular Psychiatry 25, no. 5: 1130–1140. 10.1038/s41380-019-0501-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Jakubiak, B. K. , Fuentes J. D., and Feeney B. C.. 2021. “Individual and Relational Differences in Desire for Touch in Romantic Relationships.” Journal of Social and Personal Relationships 38, no. 7: 2029–2052. 10.1177/02654075211003331. [DOI] [Google Scholar]
  27. Jaremka, L. M. , Andridge R. R., Fagundes C. P., et al. 2014. “Pain, Depression, and Fatigue: Loneliness as a Longitudinal Risk Factor.” Health Psychology 33, no. 9: 948–957. 10.1037/a0034012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Jaremka, L. M. , Fagundes C. P., Peng J., et al. 2013. “Loneliness Promotes Inflammation During Acute Stress.” Psychological Science 24, no. 7: 1089–1097. 10.1177/0956797612464059. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Joëls, M. , and Baram T. Z.. 2009. “The Neuro‐Symphony of Stress.” Nature Reviews Neuroscience 10, no. 6: 459–466. 10.1038/nrn2632. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Juruena, M. F. , Eror F., Cleare A. J., and Young A. H.. 2020. “The Role of Early Life Stress in HPA Axis and Anxiety.” Advances in Experimental Medicine and Biology 1191: 141–153. 10.1007/978-981-32-9705-0_9. [DOI] [PubMed] [Google Scholar]
  31. Karagozoglu, S. , and Kahve E.. 2013. “Effects of Back Massage on Chemotherapy‐Related Fatigue and Anxiety: Supportive Care and Therapeutic Touch in Cancer Nursing.” Applied Nursing Research 26, no. 4: 210–217. 10.1016/j.apnr.2013.07.002. [DOI] [PubMed] [Google Scholar]
  32. Klein, E. M. , Brähler E., Dreier M., et al. 2016. “The German Version of the Perceived Stress Scale—Psychometric Characteristics in a Representative German Community Sample.” BMC Psychiatry 16, no. 1: 159. 10.1186/s12888-016-0875-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Koob, G. F. 1999. “Corticotropin‐Releasing Factor, Norepinephrine, and Stress.” Biological Psychiatry 46, no. 9: 1167–1180. 10.1016/s0006-3223(99)00164-x. [DOI] [PubMed] [Google Scholar]
  34. Krahé, C. , von Mohr M., Gentsch A., et al. 2018. “Sensitivity to CT‐optimal, Affective Touch Depends on Adult Attachment Style.” Scientific Reports 8: 14544. 10.1038/s41598-018-32865-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Lahtinen, O. , and Salmivalli C.. 2020. “The Relationship Between Mindfulness Meditation and Well‐Being During 8 Weeks of Ecological Momentary Assessment.” Mindfulness 11, no. 1: 255–263. 10.1007/s12671-019-01248-x. [DOI] [Google Scholar]
  36. Landmark‐Høyvik, H. , Reinertsen K. V., Loge J. H., et al. 2010. “The Genetics and Epigenetics of Fatigue. PM&R.” Journal of Injury, Function, and Rehabilitation 2, no. 5: 456–465. 10.1016/j.pmrj.2010.04.003. [DOI] [PubMed] [Google Scholar]
  37. Löken, L. S. , Wessberg J., Morrison I., McGlone F., and Olausson H.. 2009. “Coding of Pleasant Touch by Unmyelinated Afferents in Humans.” Nature Neuroscience 12, no. 5: 547–548. 10.1038/nn.2312. [DOI] [PubMed] [Google Scholar]
  38. MacBeth, A. , and Gumley A.. 2012. “Exploring Compassion: A Meta‐Analysis of the Association Between Self‐Compassion and Psychopathology.” Clinical Psychology Review 32, no. 6: 545–552. 10.1016/j.cpr.2012.06.003. [DOI] [PubMed] [Google Scholar]
  39. Manigault, A. W. , Slutsky J., Raye J., and Creswell J. D.. 2021. “Examining Practice Effects in a Randomized Controlled Trial: Daily Life Mindfulness Practice Predicts Stress Buffering Effects of Mindfulness Meditation Training.” Mindfulness 12, no. 10: 2487–2497. 10.1007/s12671-021-01718-1. [DOI] [Google Scholar]
  40. Miller, G. E. , Chen E., and Zhou E. S.. 2007. “If it Goes Up, Must it Come Down? Chronic Stress and the HPA Axis in Humans.” Psychological Bulletin 133, no. 1: 25–45. 10.1037/0033-2909.133.1.25. [DOI] [PubMed] [Google Scholar]
  41. Morrison, I. 2016. “Keep Calm and Cuddle on: Social Touch as a Stress Buffer.” Adaptive Human Behavior and Physiology 2, no. 4: 344–362. 10.1007/s40750-016-0052-x. [DOI] [Google Scholar]
  42. Morton, M. L. , Helminen E. C., and Felver J. C.. 2020. “A Systematic Review of Mindfulness Interventions on Psychophysiological Responses to Acute Stress.” Mindfulness 11, no. 9: 2039–2054. 10.1007/s12671-020-01386-7. [DOI] [Google Scholar]
  43. Nater, U. M. , Maloney E., Heim C., and Reeves W. C.. 2011. “Cumulative Life Stress in Chronic Fatigue Syndrome.” Psychiatry Research 189, no. 2: 318–320. 10.1016/j.psychres.2011.07.015. [DOI] [PubMed] [Google Scholar]
  44. Neff, K. D. 2003. “Self‐Compassion: An Alternative Conceptualization of a Healthy Attitude Toward Oneself.” Self and Identity 2, no. 2: 85–101. 10.1080/15298860309032. [DOI] [Google Scholar]
  45. Neff, K. D. 2023. “Self‐Compassion: Theory, Method, Research, and Intervention.” Annual Review of Psychology 74, no. 1: 193–218. 10.1146/annurev-psych-032420-031047. [DOI] [PubMed] [Google Scholar]
  46. Neff, K. D. , and Germer C. K.. 2013. “A Pilot Study and Randomized Controlled Trial of the Mindful self‐compassion Program.” Journal of Clinical Psychology 69, no. 1: 28–44. 10.1002/jclp.21923. [DOI] [PubMed] [Google Scholar]
  47. Neumann, E. , Rohmann E., and Sattel H.. 2023. “The 10‐Item Short Form of the German Experiences in Close Relationships Scale (ECR‐G‐10): Model Fit, Reliability, and Validity.” Behavioral Sciences 13, no. 11: 935. 10.3390/bs13110935. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Neumann, E. , Rohmann E., and Sattel H.. 2024. “ECR‐G‐10. Experiences in Close Relationships Scale ‐ German 10‐Item Short Form [Verfahrensdokumentation, Fragebogen].” In Leibniz‐Institut Für Psychologie (ZPID) (Hrsg.), Open Test Archive . ZPID. 10.23668/psycharchives.15202. [DOI] [Google Scholar]
  49. Noone, C. , and McKenna‐Plumley P. E.. 2022. “Lonely for Touch? A Narrative Review on the Role of Touch in Loneliness.” Behaviour Change 39, no. 3: 157–167. 10.1017/bec.2022.12. [DOI] [Google Scholar]
  50. OpenAI . 2025. “ChatGPT (GPT‐5) [Large language model].” https://chat.openai.com/.
  51. Packheiser, J. , Hartmann H., Fredriksen K., Gazzola V., Keysers C., and Michon F.. 2024. “A Systematic Review and Multivariate Meta‐Analysis of the Physical and Mental Health Benefits of Touch Interventions.” Nature Human Behaviour 8, no. 6: 1088–1107. 10.1038/s41562-024-01841-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Papadopoulos, A. S. , and Cleare A. J.. 2011. “Hypothalamic–Pituitary–Adrenal Axis Dysfunction in Chronic Fatigue Syndrome.” Nature Reviews Endocrinology 8, no. 1: 22–32. 10.1038/nrendo.2011.153. [DOI] [PubMed] [Google Scholar]
  53. Park, B. J. , Choi Y., Lee J. S., Ahn Y. C., Lee E. J., and Son C. G.. 2024. “Effectiveness of Meditation for Fatigue Management: Insight From a Comprehensive Systematic Review and Meta‐Analysis.” General Hospital Psychiatry 91: 33–42. 10.1016/j.genhosppsych.2024.08.001. [DOI] [PubMed] [Google Scholar]
  54. Patapoff, M. A. , Jester D. J., Daly R. E., Mausbach B. T., Depp C. A., and Glorioso D. K.. 2024. “Remotely‐Administered Resilience and Self‐Compassion Intervention Targeting Loneliness and Stress in Older Adults: A Single‐Case Experimental Design.” Aging & Mental Health 28, no. 2: 369–376. 10.1080/13607863.2023.2262411. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Paul‐Dauphin, A. , Guillemin F., Virion J.‐M., and Briançon S.. 1999. “Bias and Precision in Visual Analogue Scales: A Randomized Controlled Trial.” American Journal of Epidemiology 150, no. 10: 1117–1127. 10.1093/oxfordjournals.aje.a009937. [DOI] [PubMed] [Google Scholar]
  56. Posit Team . 2025. RStudio: Integrated Development Environment for R: Computer software. https://posit.co/.
  57. R Core Team . 2025. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing: Computer software. https://www.R‐project.org/. [Google Scholar]
  58. Russell, D. W. 1996. “UCLA Loneliness Scale (Version 3): Reliability, Validity, and Factor Structure.” Journal of Personality Assessment 66, no. 1: 20–40. 10.1207/s15327752jpa6601_2. [DOI] [PubMed] [Google Scholar]
  59. Saini, G. K. , Haseeb S. B., Taghi‐Zada Z., and Ng J. Y.. 2021. “The Effects of Meditation on Individuals Facing Loneliness: A Scoping Review.” BMC Psychology 9, no. 1: 88. 10.1186/s40359-021-00585-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Schneider, E. E. , Schönfelder S., Domke‐Wolf M., and Wessa M.. 2020. “Measuring Stress in Clinical and Nonclinical Subjects Using a German Adaptation of the Perceived Stress Scale.” International Journal of Clinical and Health Psychology 20, no. 2: 173–181. 10.1016/j.ijchp.2020.03.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Shiffman, S. , Stone A. A., and Hufford M. R.. 2008. “Ecological Momentary Assessment.” Annual Review of Clinical Psychology 4: 1–32. 10.1146/annurev.clinpsy.3.022806.091415. [DOI] [PubMed] [Google Scholar]
  62. Song, Y. , Sun Z., Luo F., and Yu B.. 2025. “Loneliness Is Associated With Diminished Heart Rate Variability Reactivity to Acute Social Stress in Younger Adults.” Biological Psychology 194: 108963. 10.1016/j.biopsycho.2024.108963. [DOI] [PubMed] [Google Scholar]
  63. Spitoni, G. F. , Zingaretti P., Giovanardi G., et al. 2020. “Disorganized Attachment Pattern Affects the Perception of Affective Touch.” Scientific Reports 10, no. 1: 9658. 10.1038/s41598-020-66606-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  64. Steptoe, A. , Hamer M., and Chida Y.. 2007. “The Effects of Acute Psychological Stress on Circulating Inflammatory Factors in Humans: A Review and Meta‐Analysis.” Brain, Behavior, and Immunity 21, no. 7: 901–912. 10.1016/j.bbi.2007.03.011. [DOI] [PubMed] [Google Scholar]
  65. Tang, Y.‐Y. , Hölzel B. K., and Posner M. I.. 2015. “The Neuroscience of Mindfulness Meditation.” Nature Reviews Neuroscience 16, no. 4: 213–225. 10.1038/nrn3916. [DOI] [PubMed] [Google Scholar]
  66. Terehov, A. , and Yakovlev M.. 2023. “Mathematical Modeling of the Risks of Stress‐Related Diseases: A Review.” Bulletin of Rehabilitation Medicine 22, no. 4: 159–166. 10.38025/2078-1962-2023-22-4-159-166. [DOI] [Google Scholar]
  67. Thayer, J. F. , and Lane R. D.. 2000. “A Model of Neurovisceral Integration in Emotion Regulation and Dysregulation.” Journal of Affective Disorders 61, no. 3: 201–216. 10.1016/s0165-0327(00)00338-4. [DOI] [PubMed] [Google Scholar]
  68. Trull, T. J. , and Ebner‐Priemer U.. 2013. “Ambulatory Assessment.” Annual Review of Clinical Psychology 9, no. 1: 151–176. 10.1146/annurev-clinpsy-050212-185510. [DOI] [PMC free article] [PubMed] [Google Scholar]
  69. von Mohr, M. , Kirsch L. P., and Fotopoulou A.. 2017. “The Soothing Function of Touch: Affective Touch Reduces Feelings of Social Exclusion.” Scientific Reports 7, no. 1: 13516. 10.1038/s41598-017-13355-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  70. Westenberger, A. , Nöhre M., Brähler E., Morfeld M., and de Zwaan M.. 2022. “Psychometric Properties, Factor Structure, and German Population Norms of the Multidimensional Fatigue Inventory (MFI‐20).” Frontiers in Psychiatry 13: 1062426. 10.3389/fpsyt.2022.1062426. [DOI] [PMC free article] [PubMed] [Google Scholar]
  71. Wilson, T. D. , Reinhard D. A., Westgate E. C., et al. 2014. “Just Think: The Challenges of the Disengaged Mind.” Science 345, no. 6192: 75–77. 10.1126/science.1250830. [DOI] [PMC free article] [PubMed] [Google Scholar]
  72. Zeidan, F. , Johnson S. K., Gordon N. S., and Goolkasian P.. 2010. “Effects of Brief and Sham Mindfulness Meditation on Mood and Cardiovascular Variables.” Journal of Alternative & Complementary Medicine 16, no. 8: 867–873. 10.1089/acm.2009.0321. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supporting Information S1

SMI-42-e70145-s001.docx (32.5KB, docx)

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Articles from Stress and Health are provided here courtesy of Wiley

RESOURCES