The Autonomous Sensory Meridian Response Activates the Parasympathetic Nervous System


 The autonomous sensory meridian response (ASMR) produces a tingling sensation with positive feelings when binaural sounds are presented through headphones. This crossmodal response is believed to be related to sensory integration, but its psychological and physiological mechanisms are largely unknown. We conducted two experiments (on thirty and forty-two participants, respectively) to measure the subjective intensity of ASMR, finger photoplethysmography (PPG), and shift in mental states during their ASMR experience. We found that the intensity of ASMR was greater when auditory and visual stimuli were presented simultaneously than when only auditory stimuli were presented. This suggests the involvement of sensory-based, bottom-up processing. Furthermore, their pulse rates decreased and PPG amplitudes increased during the ASMR experience. Self-reported measures showed that the ASMR experience shifted participant’s negative moods to neutral mental states. These results suggest that the parasympathetic nervous system is dominant during the ASMR experience.


Introduction
Auditory events occurring in close proximity to the head can evoke skin-related sensations. We may experience a tingling sensation across the scalp and back of the neck, termed the autonomous sensory meridian response (ASMR), while listening to binaural sounds with headphones 1 . Such cross-modal bodily sensations provide clues to investigate underlying interaction between the auditory and somatosensory systems. However, it is unclear how auditory inputs shape body schema.
ASMR seems to differ from music-induced chills, often characterized as a subjective response accompanied by goose bumps and strong emotions. From the psychological perspective, the ASMR can be induced by various types of stimuli and affected by sound localization 1,2 , whereas musical chills strongly depend on sound properties 3-5 . ASMR stimuli frequently induce a feeling of relaxation and aid sleep 1,6 , whereas musical chills are associated with a profound emotional experience and shivers 7,8 . From the physiological perspective, experience of musical chills is often accompanied with goose bumps on the arms 9 , which suggests the involvement of the sympathetic nervous system 10 . However, it is unresolved whether the ASMR induces substantial changes in physiological states. We show that the ASMR activates the parasympathetic nervous system. We conducted two experiments to clarify the psychological and physiological characteristics of the ASMR experience. In Experiment 1, we investigated whether sound-induced ASMR is enhanced or reduced by simultaneous presentation of visual stimuli ( Figure 1A). Bodily sensations can be modi ed by the integration of multisensory information 11 . In the rubber-hand illusion, observers watch a rubber hand being stroked synchronously with their own hand and fail to localize the perceived position of their hand 12 . The spatial resolution of passive touch is better when the arm is visible than when it is not 13 . These ndings indicate a close relationship between visual and tactile modalities. Given that the ASMR is derived from bottom-up processing, bodily sensations could be increased with the addition of visual information. It is also known that skin-related modalities are in uenced by one's attention and expectation 14,15 . Thus, it is possible that participant curiosity about invisible sound sources shapes the ASMR. In such a case, the subjective intensity of the ASMR could be reduced by the prior knowledge of sound sources (i.e., visual information). We tested these hypotheses in Experiment 1.
We used nger photoplethysmography (PPG) to examine physiological responses during the ASMR in Experiment 2 ( Figure 1B). Several studies have demonstrated that musical chills are associated with increases in electrodermal activity, pupil diameter, or heart rate 8,16 . These ndings indicate that the sympathetic nervous system is dominant during musical chills. A recent study using ASMR stimuli showed that the skin conductance level increases and heart rate decreases while participants watch ASMR videos 17 . This means that the ASMR is associated with activation and deactivation of the autonomic nervous system. As described above, ASMR users frequently want to experience a feeling of relaxation, rather than excitement. Given that the ASMR activates the parasympathetic nervous system, it is possible that, due to blood vessel dilation, the pulse wave amplitude increases; due to slowing of the heart, pulse rate decreases. Using the Pro le of Mood States 2nd Edition (POMS 2), we checked participant mood states before and after they experienced ASMR stimuli. We examined transformation processes from sensory inputs to bodily sensations by means of changes in ASMR estimates, physiological responses, and mood states.

Participants
Thirty-two college students were recruited for Experiment 1. Two of the participants were excluded from subsequent analyses because they did not perceive tingling nor tickling sensations, leaving data from 30 participants (12 males and 18 females; mean ± SD age = 20.5 ± 0.13 years, range [19][20][21][22][23][24][25][26]. Participants were randomly assigned to either the audiovisual or audio-only condition. Numbers of males and females were counterbalanced between the two conditions to control for gender effects. Another group of 44 college students participated in Experiment 2. Data from two participants were lost because of technical problems, leaving data from 42 participants (14 males and 28 females; mean ± SD age = 21.9 ± 2.3 years, range 20-33). For each experiment, 12 and 30 participants reported that they regularly use ASMR media, but they had never watched the ASMR stimuli used in this study. According to a priori power analyses with a power of 0.8 (α-level = 0.05), we required 32 participants for Experiment 1 to detect a main effect (effect size = 0.4; mixed-design analysis of variance: ANOVA) and 46 participants for Experiment 2 to detect signi cant correlations (r = 0.4; bivariate normal model). This study was approved by the Research Ethics Committee of Chukyo University (approval no. RS19-015) and carried out in accordance with the Ethical Guidelines for Medical and Health Research Involving Human Subjects. All participants were right-handed with normal hearing and with normal or corrected-to-normal vision. They gave written informed consent after the procedures had been fully explained to them.
Stimuli and procedures Experiment 1 consisted of audiovisual and audio-only conditions, including ve 2-min trials for each ( Figure 1A). ASMR videos with Creative Commons licenses were taken from YouTube (https://www.youtube.com/watch?v=7MZtaAgqoTY and https://www.youtube.com/watch? v=asGLp12NSIE&t=2031s). Audiovisual stimuli were edited from segments of the videos that gave tingling sensations to all of the authors. There was no facial information in the stimuli to exclude confounding effects of facial expressions on participant responses. Participants could explicitly observe sound sources of the stimuli under the audiovisual condition. For the audio-only condition, visual information was removed from the stimuli. Participants were instructed to continuously indicate their tingling sensation using a 4-point Likert scale: "no ASMR", "weak ASMR", "moderate ASMR", and "strong ASMR." Stimulus presentation and data collection were controlled using a PC with Presentation software (Neurobehavioral Systems, Berkeley, CA, USA). Visual stimuli were presented on an LCD monitor with a temporal resolution of 60 Hz. Participants viewed the stimuli with a spatial resolution of 1,280 × 720 pixels (visual angle, 16.0 × 9.0 degree) at a distance of approximately 57 cm. They dichotically listened to the stimuli through Sennheiser HD 599 headphones. Responses of a participant's right hand were collected via four keys on a computer keyboard with a sampling rate of 1,000 Hz. A key press indicating a response was held until a subsequent key press. In interviews conducted after Experiment 1, participants reported the degree of tingling and tickling sensations throughout the experiment using a visual analogue scale (range 0-100).
In Experiment 2, the test period consisting of three 2-min trials was placed between 2-min rest trials ( Figure 1B). There was no gap between the trials. Participants experienced audiovisual stimuli while PPG signals were recorded using MP36 data acquisition units (Biopac Systems, Goleta, CA, USA) with the SS4LA PPG transducer. The stimuli used in trials 2, 3, and 4 in Experiment 1 were presented during the test period. The task procedure was the same as that under the audiovisual condition of Experiment 1. A PPG sensor was attached to a participant's left middle nger. PPG data were recorded at a sampling rate of 2,000 Hz using the software Biopac Student Lab Pro (version 4.1). Participants were instructed to minimize their body movements throughout the 10-min data-recording period.
Before and after Experiment 2, we obtained self-reported measures of participant mood states using the POMS 2 38

Data analyses
For behavioral data, participant responses were converted to an ASMR estimate (0, 1, 2, or 3). The intensity of the ASMR for each participant was computed by averaging the ASMR estimate over the test period (1-ms bins). The intensity of tingling and tickling sensations throughout the experiment was derived from the data of a visual analogue scale. The Total Mood Disturbance score for the POMS 2 was calculated by subtracting the VA subscore from the AH, CB, DD, FI, and TA subscores. We used the Bonferroni correction for post-hoc comparisons (α-level = 0.05) when conducting ANOVA. Statistical analyses were carried out with IBM SPSS Statistics (ver. 25).
PPG data were preprocessed using high-and low-pass ltering at 0.5-and 35-Hz cutoff frequencies.
Using the preprocessed data, pulse rate was evaluated for each participant. We extracted time points of peaks for all spikes and computed peak-to-peak intervals (PPIs). PPI values that deviated more than 3 SDs from the mean were removed as outliers in subsequent analyses. The pulse rate was calculated as the inverse of the PPI. The SD of the pulse rate was used as a measure of pulse rate variability. We also computed the pulse wave amplitude as the difference between the maximum and minimum values of the PPG signal in each spike interval. Figure 2A shows time-series data of the ASMR estimate. We conducted a mixed-design 2 (condition types) × 5 (trial types) ANOVA on the ASMR estimate ( Figure 2B). The ASMR estimate (mean ± standard error) was greater for audiovisual conditions (1. We evaluated tingling and tickling sensations throughout the experiment ( Figure 2C). The intensity of tingling and tickling sensations was greater under audiovisual conditions than under audio-only conditions: 69.0 ± 2.2 and 49.9 ± 6.6 for tingling, t = 2.76, p = 0.010, Cohen's d = 1.01; 61.7 ± 5.7 and 42.1 ± 7.6 for tickling, t = 2.05, p = 0.050, Cohen's d = 0.75. We examined the degree to which these measures were associated with the averaged ASMR estimate for each participant ( Figure 2D). The ASMR estimate was positively correlated with the intensity of tingling and tickling sensations: r = 0.57, p = 0.001; r = 0.58, p = 0.001. Consistent with previous ndings, the pattern of the correlations indicates that the ASMR experience is accompanied with tingling and tickling sensations.

Experiment 2: ASMR and PPG
Thus far, we have demonstrated additive effects of multisensory inputs on ASMR estimates. We chose the audiovisual stimuli from the ASMR videos used in Experiment 1. PPG and ASMR results are shown in Figure 3A. Pulse wave amplitude increased during test periods, compared with rest periods: F(4, 164) = 6.89, p < 0.001, = 0.14. In contrast, the pulse rate decreased during test periods, compared with rest periods: F(4, 164) = 14.20, p < 0.001, = 0.26. The pattern of the results indicates that the ASMR experience is associated with dominance of the parasympathetic nervous system.
We focused on intra-individual variability of the measures ( Figure 3B). Pulse wave amplitude was larger during the test period (0.554 ± 0.043) than during the rest period (0.495 ± 0. We investigated whether these measures during the test period were associated with the ASMR estimate ( Figure 3C). The ASMR estimate showed a signi cant correlation with the pulse rate (r = -0.49, p = 0.001), but not with the pulse wave amplitude (r = 0.16, p = 0.30). The correlations between the ASMR estimate and variability measures did not reach statistical signi cance: |r| < 0.06. Thus, pulse rate is more reliable for predicting subjective intensity of the ASMR experience than pulse wave amplitude.

Experiment 2: ASMR and mood states
We investigated effects of the ASMR experience on changes in participant mood states. The Total Mood Disturbance score decreased after the experiment (25.1 ± 4.1) compared to before the experiment (38.7 ± 5.1): t(41) = 5.08, p < 0.001, Cohen's d = 0.46. All subscores after the experiment were lower than those before the experiment (Figure 4). Speci cally, the differences in the CB, DD, and TA subscores reached statistical signi cance: t(41) > 2.79, p < 0.008, Cohen's d > 0.27. The reduction of negative moods re ects a bias that the ASMR experience moves participant moods toward neutral mental states.

Discussion
Behavioral results demonstrated that the ASMR estimate was greater under audiovisual conditions than under audio-only conditions. This additive effect of visual information suggests that the ASMR is largely affected by sensory-based, bottom-up processing. The ASMR estimate was associated with enhancement in pulse wave amplitude and with reduction of the pulse rate. The physiological results indicate a parasympathetic shift. Self-reported measures showed that the ASMR experience moved participants from negative moods to neutral mental states. Therefore, we conclude that the ASMR activates the parasympathetic nervous system. We revealed that multisensory (i.e., visual and auditory) inputs strengthened the ASMR, accompanied with tingling and tickling sensations. The ASMR is similar to the rubber-hand and body-swap illusions because sensory inputs in uence bodily awareness and sensations. In these illusions, the ownership of an arti cial hand and body can be induced by tactile stimulation on the hand or body and its visual information 12,18 . Several researchers have pointed out that the rubber-hand illusion is based on Bayesian perceptual learning 19,20 . The sensitivity to ASMR stimuli, as well as to the rubber-hand illusion, varies widely among individuals, but is preserved within individuals 2 . Thus, an individual's preference for ASMR videos probably represents a convergence of accumulating experience of various videos and sharpening the positive emotional aspects of ASMR videos.
Multisensory interactions frequently occur in the space surrounding one's body 21,22 because such a mechanism is critical to grasp the peripersonal space as a point of contact with the environment 23 . A tingling sensation was enhanced by sounds moving around the head 24 . Also, a strong tickling sensation occurred when the sound of an ear being stroked was presented with headphones 25 . Intriguingly, there is a case in which external stimulation is not a prerequisite for tingling sensation 15 . Focusing one's attention on a body part can give rise to various spontaneous sensations, including tingling 26,27 . A spontaneous tingling sensation more frequently occurred on ngers and toes than on the trunk 28 . The most intense sensations were reported at the tips of the ngers 26 . These results are probably related to the gradient of peripheral receptor density. Several researchers have argued that left-hand dominance of the spontaneous tingling sensation indicates right-hemisphere central processes 26,27 . In contrast, the ASMR mainly originates on the scalp and back of the neck, and laterality of body parts does not occur 1,2 . Thus, the ASMR differs from a spontaneous sensation triggered by endogenous attention.
The pulse wave amplitude increased and pulse rate decreased during the ASMR experience, indicating the dominance of the parasympathetic nervous system. This is consistent with previous ndings showing that the ASMR experience leads to a reduction of participant negative moods, including depression and anxiety 1 . A previous study demonstrated that the heart rate decreased for ASMR responders, but did not for non-responders, when participants watched ASMR videos 17 . Physiological responses observed in this study were not derived from the effect of anticipation in each individual because none had watched our audiovisual stimuli previously. In addition, we demonstrated that the pulse rate was a critical index to predict the subjective intensity of the ASMR. Therefore, our results provide general principles of physiological responses during the ASMR experience.
We presented empirical evidence supporting the report that Internet users frequently want ASMR videos to relax and calm down. In previous studies, 82% of respondents agreed that they used ASMR videos to aid sleep 1 and 63% of respondents answered that they watched such videos before sleeping 29 . Anecdotal reports about introducing sleep are consistent with our physiological ndings that the ASMR calms the body. Speci cally, we showed that an individual's ASMR intensity was closely linked to his/her pulse rate. Social isolation under the COVID-19 pandemic has led to an increase of mental illnesses, such as stress, anxiety, and depression 30 . Thus, the pulse rate can be a useful biomarker for assessment of psychological health, because PPG sensors are commonly implemented in wearable watch devices.
Several studies of emotion have used music to clarify neural correlates of chills, arguing that musical chills are closely linked to activation of paralimbic areas for emotion and reward processing, rather than those for auditory processing itself 31,32 . Heart rate increases during musical chills 8,31,33 . Chills are known as a phenomenon in which auditory inputs frequently cause goose bumps on the arms 7,9 , whereas only 15% of ASMR responders reported tingling sensation on their arms 2 . Although the ASMR and musical chills accompany pleasant emotions, the balance of autonomic functions could account for the different physiological responses of these phenomena. Even though misophonia induces negative emotional reactions to particular sounds (i.e., the opposite of the ASMR), several researchers have noted the similar sensitivity between the ASMR and misophonia 1,29 . However, misophonic triggers seem to increase the heart rate 34 . Hence, given that we specify acoustic features affecting the ASMR, chills, and misophonia, it should be possible to design comfortable sound landscapes to improve the quality of life 35 .
We showed that the ASMR experience relies on bottom-up processing, induced parasympathetic shift, and reduced negative moods. Investigating the ASMR can initiate not only new multisensory research, but also clinical investigation into treatments for depressive moods and mental stress. It has been argued that chills re ect threatening situations 36