Anxiolytic Effects of Social Touch: Moderating Influences of Trauma and Social Interaction Experiences

Social touch is essential for both physical and mental health. It plays a vital role in human communication and bonding, contributing to an overall sense of connection to others. This study is particularly concerned with the anxiolytic effects of social touch. Previous research has provided evidence that social touch can positively influence oxytocin levels (Li et al., 2019), systolic blood pressure (Holt-Lunstad et al., 2008), heart rate variability (Triscoli et al., 2017), and general mental health (Packheiser et al., 2023). Under conditions of acute stress, social touch has been shown to have the potential to decrease self-reported distress (Von Mohr et al., 2017), heart rate (Ditzen et al., 2007) and cortisol levels (Berretz et al., 2022; Ditzen et al., 2007, 2009; Dreisoerner et al., 2021). Trauma-exposed individuals often display a pattern of touch aversion and avoidance (Stevens et al., 2024). This effect is especially pronounced for fast touch compared to slow touch (Maier et al., 2020). Surprisingly, however, it remains unclear whether trauma experiences are associated with altered anxiolytic effects of social touch. The present study therefore aimed to examine the relationship between trauma experiences and the anxiolytic effects of social touch. Furthermore, the perceived pleasantness of touch, as well as its anxiolytic effects, are known to be influenced by the quality of the relationship between the touch provider and the receiver (Coan et al., 2006; Grewen et al., 2005; Kraus et al., 2019). Accordingly, we incorporated a Positive Interaction Paradigm (PIP) to test whether prior interaction experiences may moderate the association between trauma experiences and the anxiolytic effects of social touch. Neural mechanisms underlying the stress-reducing effects of social touch likely involve prefrontal regulation of limbic and autonomic responses. The prefrontal cortex (PFC), particularly the dorsolateral and ventromedial subregions, plays a key role in stress processing and emotion regulation by exerting top-down control over the amygdala and other stress-related systems (Buhle et al., 2014; Etkin et al., 2011). Tentative evidence from functional neuroimaging studies suggests that affiliative touch and physical contact can modulate prefrontal activity during stress (Coan et al., 2006), indicating that the PFC may contribute to the anxiolytic effects of social touch. In trauma-exposed individuals, emotion regulation networks involving the PFC are often disrupted, leading to either hyperreactivity to social cues or emotional numbing associated with increased recruitment of dorsolateral prefrontal areas (Akiki et al., 2017; Yehuda et al., 2015). Hence, prefrontal neuroimaging measures and self-report assessments of emotion regulation will be included to capture both neural and subjective aspects of stress modulation through social touch. Participants completed two in-person laboratory sessions at Ruhr University Bochum and three online questionnaire assessments. The first (in person) screening session included eligibility assessment via the Mini-International Neuropsychiatric Interview (M.I.N.I.; Sheehan et al., 1998; German version: Ackenheil et al., 1999), and completion of baseline questionnaires (listed under 13. Study Material). Participants also completed a “Touch Map” task, indicating on a topographical body map the regions where they would feel comfortable being touched (Nummenmaa et al., 2014; Suvilehto et al., 2015). The main experimental session (~180 minutes) included a series of physiological, behavioural, and neuroimaging tasks: the PIP, a social distance estimation task, a Touch Value Task and a Threat of Shock Task. Participants were randomized to a PIP or control condition. Physiological monitoring - Electrodermal Activity (EDA), Electrocardiogram (ECG) and respiration - and three saliva samples were collected before (t1), during (t2), and after (t3) the experimental tasks. Participants completed social distance estimations, in which they were asked to indicate comfortable and uncomfortable speaking distances (Scheele et al., 2012) and rated their mood and affect using the Profile of Mood States (POMS; McNair et al., 1971; German: Albani et al., 2005) and Positive and Negative Affect Schedule (PANAS; Watson et al., 1988; German: Breyer et al., 2016) both before and after a manipulation phase (PIP/control). The PIP comprised a semi-structured 10-minute conversation between the participant and an unfamiliar experimenter (Lieberz et al., 2021) The participants were asked to talk about pleasurable topics such as a fictive lottery win and their hobbies and interests to establish a positive relationship with the experimenter. In addition, they took part in a movement imitation task with the same experimenter designed to induce the release of the neuropeptide oxytocin (Spengler et al., 2017). Half of the participants were assigned to the control group, where they completed a written task about the same pleasurable topics instead of discussing them with an experimenter. Subsequently, they took part in a non-social computer-based movement imitation control task. After the PIP/control, participants completed a computer-based task designed to assess the perceived pleasantness of interpersonal touch by assigning a subjective monetary value to different touch stimuli. The experimenter (either the experimenter they had previously interacted with during the PIP or – if they were part of the control group - an unfamiliar experimenter they had not met before), seated behind a curtain, administered light touch stimuli to a 10 cm area on the participant’s left forearm. The experiment consisted of two parts. In the first part, participants were introduced to two types of touch that differed in stroking velocity: slow (affective, CT-optimal at 3cm/s) and fast (neutral, non CT-optimal at 10cm/s). Each touch type was paired with a distinct image displayed on the screen for 5 seconds while the corresponding touch was administered. Each image–touch pairing was presented once to allow participants to learn the associations. In the second part, participants completed 20 randomized trials, with each touch type presented ten times. Each trial began with a fixation cross displayed for 4–6 seconds, followed by the presentation of a cue image indicating the upcoming touch. In half of the trials, participants rated how much they would be willing to pay to receive the touch (“wanting” rating) using a visual analogue scale (0–100). In the other half of the trials, they rated how much the touch was worth to them (“liking” rating) on the same scale after the touch was delivered. Participants were then tested with a Threat of Shock Task. The intensity of the electric shock was individually calibrated and adjusted to a subjective 6 (on a scale from 1 “not unpleasant at all” to 10 “as unpleasant as imaginable”). The task consisted of three blocks (with 12, 24, and 12 trials) and touch was administered in the second block. There was a 5-min break between blocks during which participants completed additional emotion regulation questionnaires adapted from the Difficulties in Emotion Regulation Scale (Gratz & Roemer, 2004; German: Gutzweiler & In-Albon, 2018), and Emotion Regulation Questionnaire (Gross & John, 2003; German: Abler et al., 2009), asking them to assess their use of emotion regulation strategies specifically during the block they just completed, hence at three points (after block 1, after block 2 and after block 3). During the Threat of Shock Task participants were presented with cues indicating either a potential shock (Shock) or no shock (No Shock). Each cue was followed by a 4-second anticipatory period before the corresponding outcome was delivered. Shocks were administered in only 33.3% of the Shock trials. There was a 2-3 second inter stimulus interval preceding each trial during which a fixation cross was displayed. Participants received CT-optimal touch (3 cm/s) from an experimenter (not visible to the participant) on their left arm from cue onset until stimulus delivery during the second block, while there was no touch in the first and third block. Trials were followed by ratings asking participants to estimate tension during countdown (on a scale from 1, not tense at all, to 10, very tense) as well as unpleasantness of shock (on a scale from 1, not unpleasant at all, to 10, very unpleasant) and pleasantness of touch (on a scale from 1, not pleasant, to 10, very pleasant) in the corresponding trials. Participants were equipped with two Ag/AgCl electrodes filled with isotonic electrolyte medium on the centre of the right forearm to enable shock transmission using a constant voltage stimulator (STM200, BIOPAC Systems, Inc. Goleta, CA, USA). Another set of electrodes was placed on the hypothenar surface of the left hand to measure EDA. Functional near-infrared spectroscopy (fNIRS) data were recorded using the NIRSport2 system (NIRx Medical Technologies, LLC, Berlin, Germany). The NIRSport2 is a continuous-wave fNIRS device employing dual-wavelength light sources (760 nm and 850 nm) and silicon photodiode detectors to measure changes in oxy- and deoxyhemoglobin concentrations in cortical tissue. For this study, a prefrontal 8x8 optode configuration was used. This configuration provided bilateral coverage of the dorsolateral and medial prefrontal cortices. The optodes were mounted in a flexible NIRx cap (NIRScap) using spring-loaded holders to ensure stable and comfortable contact with the scalp. Optode positions were aligned according to the international 10–20 system, with the array centered approximately around Fpz and extending laterally toward F3/F4 and superiorly toward Fz. Each channel consisted of a source–detector pair separated by approximately 3 cm (long-separation channels). The exact source–detector geometry followed the predefined NIRx template for the prefrontal 8x8 setup. Before data acquisition, signal quality was optimized using the automatic gain adjustment routine provided by the Aurora software. All recordings were monitored in real time to ensure consistent optical coupling and acceptable signal-to-noise ratios across channels. Event markers were synchronized with the fNIRS data stream via the Lab Streaming Layer (LSL) interface supported by the NIRSport2 system. Raw data was saved in .snirf format, including metadata describing source and detector positions, wavelengths, and timing information. Participants also completed additional tasks that are not relevant to this specific study but will be preregistered and published separately.