Introduction

In recent years, autonomous sensory meridian response (ASMR) has emerged as a globally recognized digital phenomenon, captivating millions of individuals across diverse cultural and linguistic communities. ASMR is characterized by distinct tingling sensations that typically begin at the scalp and travel down the spine. These sensations are elicited through exposure to specific auditory and visual stimuli, often delivered via online platforms such as YouTube. ASMR content is often associated with familiar triggers such as gentle whispers, rhythmic tapping, and role-played scenarios simulating personal attention. However, its genres and formats are remarkably diverse, reflecting a wide range of cultural conventions, aesthetic styles, and individual preferences1,2. Interestingly, ASMR induces tactile-like sensations despite the absence of physical touch, making it comparable to an illusory sensory phenomenon3. Beyond its sensory appeal, the capacity of ASMR to simulate “digitally mediated intimacy” also highlights broader psychological and social implications4.

Despite its widespread popularity among the general population, many aspects of mechanisms underlying ASMR remain poorly understood5,6. Only a subset of individuals report experiencing distinct tingling sensations and the accompanying state of relaxation that ASMR aims to induce1,7. This selective responsiveness underscores notable individual differences in susceptibility to ASMR, raising an intriguing question: why do some individuals experience profound psychological and physiological reactions to ASMR stimuli, while others remain unaffected? Emerging research suggests that these differences may arise from a complex interplay of sensory and emotional experiences7,8,9, psychological traits10,11,12,13, and physiological factors14,15,16,17. However, precise mechanisms driving these individual differences remain largely unexplored.

The present study used a mixed approach to stimulus design to address concerns about the generalizability and cultural specificity of ASMR research18,19. We combined culturally familiar genres with perceptually defined categories of ASMR stimuli to balance ecological validity with experimental control. This approach was intended to minimize biases related to social or cultural expectations and to improve the reproducibility of ASMR responsiveness measures, thereby contributing to a deeper understanding of psychological and physiological factors underlying individual differences in ASMR experiences. Moreover, this design enabled us to investigate theoretical accounts that view ASMR as a surrogate for social or affective touch.

One such account is the social grooming hypothesis, which posits that ASMR mimics comforting effects of interpersonal touch within a simulated context14. Unlike the relaxation induced by exposure to natural environments, ASMR elicits feelings of intimacy and emotional connection by replicating sensory dynamics of close social interactions. A central mechanism in this process is affective touch—a gentle and emotionally soothing type of touch mediated by C-tactile (CT) afferents20. These unmyelinated fibers, primarily located in hairy skin such as the forearm, respond optimally to slow, light strokes within a velocity range (3–10 cm/s). Accumulating evidence shows that such stimulation is consistently associated with increased pleasantness and emotional comfort21,22,23,24. Based on these findings, we hypothesized that individuals with greater sensitivity to CT-mediated touch would show stronger ASMR.

This view complements a recent neurobiological model25, which suggests that ASMR arises from synesthetic cross-activation between the auditory cortex and the posterior insular cortex, particularly its representation of affective touch. This type of cross-modal integration has been hypothesized to promote parasympathetic activity and emotional calming, providing a theoretical framework for how ASMR stimuli may evoke both sensory and emotional responses.

Building on the role of tactile mechanisms, interoception—the perception of internal bodily signals—has been proposed as another contributor to ASMR experiences. It includes awareness of signals such as hunger, heart rate, respiration, and visceral sensations26, and plays a crucial role in emotional regulation and bodily homeostasis. Recent studies suggest a functional connection between interoceptive processing and ASMR. For instance, increased amplitude of heartbeat-evoked potentials has been observed during ASMR experiences, indicating heightened cortical processing of cardiac signals27. Interestingly, ASMR responders have been reported to show lower interoceptive accuracy—as measured by heartbeat counting—compared to non-responders28, though the implications of this finding remain open to interpretation. One possibility is that reduced accuracy reflects a greater reliance on exteroceptive or affective cues, rather than impaired bodily awareness per se. Moreover, heightened awareness of bodily signals has been associated with stronger emotional responses in other domains29,30. To further investigate whether interoceptive accuracy contributes to individual differences in ASMR responsiveness, the present study employed a heartbeat counting task and examined its association with tingling intensity during ASMR stimulation.

To examine how sensory and interoceptive traits contribute to ASMR responsiveness, we designed three tasks targeting tingling intensity, affective touch sensitivity, and interoceptive accuracy. ASMR responsiveness was assessed by having participants watch ASMR videos and continuously report the intensity of tingling sensations. Affective touch sensitivity was measured through pleasantness ratings of gentle strokes delivered to the forearm and palm at varying velocities. Interoceptive accuracy was evaluated using a heartbeat counting task, in which participants silently counted their heartbeats over specified time intervals and reported their counts. By analyzing relationships among these measures, this study aims to identify whether individual differences in ASMR responsiveness are associated with heightened sensitivity to affective touch and accurate perception of internal bodily signals. This approach provides a framework for understanding psychological and physiological mechanisms underlying ASMR and offers new directions for examining how social touch and interoceptive processing shape affective sensory experiences.

Methods

Ethics statement

Ethical approval for this study was granted by the Research Ethics Committee of Chukyo University (no. 2024–025). All procedures adhered to the Ethical Guidelines for Medical and Biological Research Involving Human Subjects. Written informed consent was obtained from all participants before the study began, and they received 1,000 JPY as compensation for their participation.

Participants

Forty-six participants (25 men and 21 women; mean ± SD age = 23.2 ± 4.0 years) were recruited for this study. Based on an a priori power analysis, the sample size (N = 46) was determined using G*Power (ver. 3.1.9.7). This analysis was designed to achieve a statistical power of 0.80 at an α-level of 0.05, targeting a medium effect size (r = 0.40) as suggested by previous studies31,32. Participants were free from any diagnosed neurological or cardiovascular disorders.

General procedures

All participants completed all three tasks regarding to ASMR, affective touch, and heartbeat counting. Since both the ASMR and affective touch tasks are linked to mood and relaxation, the heartbeat counting task was scheduled between the two to reduce potential mood carryover effects. Before starting each task, participants performed a practice trial to familiarize themselves with procedures. Stimulus presentation and data collection were controlled using a PC with PsychoPy (ver. 2023.2.2). The entire experiments took approximately 1 h.

ASMR task

The ASMR task, adapted from previous studies32,33, was designed to evaluate participant sensitivity to ASMR triggers through a series of 20 video clips (Fig. 1), each lasting 30 s. These clips were categorized into four categories: microphone interactions (Mic), peripersonal audiovisual stimuli (Proximal), eating scenes (Eating), and nature environments (Nature). The Mic, Proximal, and Eating categories featured an on-screen performer, while Nature clips excluded any human presence and served as a non-ASMR control. Although all participants reported prior experience with ASMR videos in general, none had previously encountered specific video clips used as stimuli in this experiment.

Fig. 1
figure 1

Representative video frames from four ASMR categories: Mic, Proximal, Eating, and Nature. The first three categories featured an on-screen performer engaging in ASMR trigger behaviors. Mic videos included scenes involving direct contact with microphones using fingers, ear picks, balloons, nails, or towels. Proximal videos depicted hand movements and object manipulations occurring in close proximity to the camera, evoking a sense of personal attention. Eating videos presented close-up sounds of consuming various foods. Nature videos served as a control condition and consisted of non-social audiovisual environments. Each video lasted 30 s, and five distinct videos were prepared for each category. Colored borders around video frames indicate individual video conditions and correspond to time-series lines in Fig. 2.

Full size image

Specifically, Mic clips depicted direct contact with binaural microphones using fingers, ear picks, balloons, nails, or towels. Proximal clips included audiovisual interactions occurring near microphones—such as scratching corkboards, rubbing gel-covered hands, plucking glass surfaces, tapping plastic containers, or flipping pages—without direct stimulation of recording devices. Eating clips portrayed scenes of consuming hard candy, honeycomb, aloe, pickles, or fried chicken. Nature clips displayed forest bathing, bubbling soda, rustling leaves, crackling fires, or flowing streams, reflecting calming audiovisual phenomena not explicitly designed to trigger ASMR.

These categories were selected to represent a blend of perceptually defined stimuli and culturally familiar ASMR genres. Mic and Proximal were constructed based on auditory spatial proximity and tactile simulation, while Eating reflected widespread genre conventions in ASMR media. This approach allowed us to balance ecological validity with experimental control, enabling us to examine a range of ASMR responses while minimizing social, emotional, and identity-related confounds. We deliberately excluded whispering and role-play scenarios, as these are heavily influenced by performer characteristics and cultural expectations.

ASMR videos were selected from a publicly available video-sharing platform (https://www.youtube.com/@HatomugiASMR) and evaluated by three independent raters (excluding the authors) with prior familiarity with ASMR media. The raters assessed candidate videos based on their perceived potential to elicit tingling sensations, and selected five videos for each category through consensus. The selected clips were then trimmed to 30-s segments to retain the most salient ASMR trigger sequences. The stimulus duration was chosen to balance effective ASMR induction, as reported in prior research17,32, with the need to minimize participant fatigue across repeated trials. Performer faces and voices were removed from all ASMR videos to minimize emotional and social confounds. Stimulus presentation was pseudo-randomized with the constraint that no two videos from the same category were shown consecutively, in order to minimize category-specific carryover effects.

Visual stimuli were presented on an LCD monitor with a temporal resolution of 60 Hz and a spatial resolution of 1280 × 720 pixels (visual angle of 16.0 × 9.0 degree) at a viewing distance of approximately 57 cm. Auditory stimuli were delivered dichotically through Sennheiser HD 599 headphones. A 3-s fixation cross was displayed between videos to ensure smooth transitions and signal the onset of the next video.

During each video, participants continuously reported the intensity of their tingling sensations on a 4-point Likert scale (0 = none, 1 = slight, 2 = moderate, 3 = intense) via designated keys on a keyboard. Participants were instructed to press and hold the key corresponding to the current intensity level, and to switch to another key whenever the perceived intensity changed. This procedure allowed for real-time tracking of moment-to-moment fluctuations in tingling sensations during each 30-s video presentation. The validity of this continuous rating method has been supported by previous research17, which reported high correlations between continuous and retrospective ASMR ratings. Although this approach required active engagement during stimulus presentation, we prioritized temporal resolution over passive immersion. Participants reported no notable discomfort or distraction. Data were recorded at a sampling rate of 1000 Hz. The task was conducted without rest periods between trials, maintaining a consistent flow and enabling standardized, continuous measurement across different video categories.

Affective touch task

This task design enabled a comprehensive assessment of participants’ affective touch sensitivity, considering variations across skin types and stroking velocities. To selectively stimulate CT afferents22, we compared two skin areas: the forearm (hairy skin, where CT afferents are densely distributed) and the palm (glabrous skin, where CT afferents are sparse). Stimulation was delivered using a cheek brush made of goat hair (J110, Hakuhodo, Hiroshima, Japan), for its consistent texture and soft contact properties. Gentle strokes were applied to the dorsal surface of the right forearm, spanning from the wrist to the midpoint of the forearm. Similarly, strokes were performed over the center of the right palm. Each stroke was standardized in length (approximately 9 cm) and maintained with uniform pressure across both areas.

Strokes were delivered at five predetermined velocities: 0.3 cm/s, 1 cm/s, 3 cm/s, 9 cm/s, and 27 cm/s. Based on previous findings20,21, velocities of 3 cm/s and 9 cm/s were hypothesized to be CT-optimal, typically perceived as highly pleasant. Each stroke lasted for 6 s, irrespective of the velocity. During the task, participants were seated comfortably in a chair in a relaxed posture. To prevent visual feedback, their right arm was obscured from view. The experimenter manually delivered strokes to specified skin areas. After each stroke, participants provided a verbal rating of the pleasantness of the touch on a 0–10 scale (0 = none, 10 = extreme). Each velocity was administered three times on each skin area in a randomized order, resulting in a total of 30 stroking trials.

To minimize variability in stroke application, all stimuli were administered by a single trained experimenter. The experimenter was instructed to maintain a consistent hand posture and stroke height across all trials. An elbow rest was used to stabilize the arm position and reduce variability in applied pressure. To ensure consistent stroking velocity, a visual pacing cue was presented on a monitor facing the experimenter: a red marker programmed in PsychoPy moved across the screen at the target speed. The experimenter practiced following this cue prior to the experiment and monitored it throughout each trial.

Heartbeat counting task

Interoceptive accuracy was assessed using a modified version of the heartbeat counting task34. Participants wore a Polar Verity Sense (Polar Electro, Kempele, Finland) wristband on their left arm. This device employed near-infrared technology to monitor heartbeats, transmitting data via Bluetooth to a PC application for real-time recording. The wristband was secured snugly to prevent participants from perceiving their pulse directly, adhering to guidelines from previous studies35,36. The task included three kinds of interval trials (25 s, 45 s, and 65 s), each repeated three times in a randomized order, for a total of nine trials. This design ensured a robust and standardized measurement of interoceptive accuracy.

Participants were instructed to sit with both hands on the table, refrain from crossing their legs, and focus on distinctly perceiving their heartbeats, including weaker sensations, without guessing36. At the start of the task, a 120-s baseline recording was conducted to determine resting heart rate levels. During each trial, participants silently counted their heartbeats without physically checking their pulse. At the end of each trial, they reported the number of heartbeats they counted. Simultaneously recorded actual heartbeat counts were used to calculate interoceptive accuracy by comparing reported counts with recorded counts. A 20-s rest period was provided between trials to allow participants to recover before the next trial.

Data analyses

For the ASMR task, each category consisted of five 30-s videos. Time-series data of participant ratings were downsampled to 25 Hz to improve efficiency in data processing and analysis. To calculate overall tingling intensity ratings for each category, a two-step averaging process was used. First, the initial 5 s of each video were excluded from the analysis to account for potential instability in participant responses immediately after stimulus onset. The remaining 25-s time-series data for each video were aggregated to compute mean ratings. Next, these video-level means were averaged across five videos within the same category. This procedure resulted in mean tingling intensity ratings for each category, enabling comparisons of tingling sensations across different ASMR categories.

Pleasantness ratings from the affective touch task were averaged separately for each velocity condition across skin areas. Velocities of 3 cm/s and 9 cm/s were considered CT-optimal speeds, whereas 0.3 cm/s and 1 cm/s served as baseline conditions. The affective touch sensitivity score was calculated as the ability to differentiate between highly pleasant and neutral tactile stimuli using Eq. (1):

$$\text{Affective Touch Sensitivity}={\text{Mean Pleasantness}}_{(\text{3,9 cm}/\text{s})}-{\text{Mean Pleasantness}}_{(0.3, 1\text{cm}/\text{s})}$$
(1)

This differential score quantified the participant sensitivity to variations in touch pleasantness across velocities and skin types. Data from the 27 cm/s condition were excluded from Eq. (1) because pleasantness ratings plateaued at the 9 cm/s condition.

Heartbeat data were converted to heart rates (beats per minute: bpm) and averaged for each interval condition before further analysis. In addition, interoceptive accuracy was calculated from heartbeat data using a calculation adapted from a previous study34. This evaluates how closely participants’ self-reported heartbeat counts match their actual heartbeat counts. For each of the trials, interoceptive accuracy score was determined using Eq. (2):

$$\text{Interoceptive Accuracy}=1- \frac{\text{Actual Heartbeat}-\text{Reported Heartbeat}}{\text{Actual Heartbeat}}$$
(2)

A score close to 1 indicates high interoceptive accuracy, reflecting minimal deviation between reported and actual heartbeats. Three participants reported that they could not perceive their heartbeats and were subsequently excluded from correlation analyses.

All statistical analyses were performed using JASP (version 0.19.2; https://jasp-stats.org/). Post-hoc comparisons following an analysis of variance (ANOVA) were conducted using Bonferroni correction. For sets of four comparisons, the adjusted significance threshold was set at α = 0.0125 to control for family-wise error rate. Relationships among three key variables—tingling intensity, affective touch sensitivity, and interoceptive accuracy—were first assessed for multivariate normality using the Shapiro–Wilk test. The results indicated that the overall data did not follow a normal distribution: W = 0.960, p = 0.013. Accordingly, Spearman’s rank correlations (rs) were used in main analyses, and 95% confidence intervals (CIs) were estimated using 10,000 bootstrap resamples. Notably, the pattern of results was consistent with that obtained from Pearson’s correlations.

To further examine predictive contributions of affective touch sensitivity and interoceptive accuracy to ASMR responsiveness, multiple linear regression analyses were conducted separately for each video category. In each model, forearm touch sensitivity and interoceptive accuracy served as predictors, with tingling intensity as the dependent variable. Palm touch sensitivity was excluded from regression models due to its limited theoretical relevance—given the low density of CT afferents in glabrous skin—and lack of statistical association with tingling intensity.

Results

ASMR task

Figure 2 shows time-series data of tingling intensity ratings across different categories. This analysis reveals dynamic fluctuations in tingling intensity, providing insights into category-specific trends. A one-way ANOVA revealed a significant main effect of category: F(3, 135) = 32.83, p < 0.001, \({\eta }^{2}\) = 0.42. Post-hoc multiple comparisons (mean ± SD) revealed the following pattern: Mic (1.26 ± 0.65) > Proximal (0.68 ± 0.46) and Eating (0.89 ± 0.81) > Nature (0.42 ± 0.42). The results indicate that participants experienced distinct tingling sensations elicited by ASMR videos, which were differentiated from general feelings of pleasantness. This finding supports the hierarchical structure of ASMR trigger potency, with nature stimuli being less effective in eliciting ASMR than stimuli designed to simulate close personal interactions. This is consistent with a previous study showing that tingling sensations are better predicted by acoustic features with shorter temporal windows compared to pleasant feelings32. These findings highlight unique psychological and emotional responses triggered by ASMR through specific temporal and spectral characteristics of the stimuli.

Fig. 2
figure 2

Time-series plots and corresponding boxplots of tingling intensity ratings for each category (N = 46). Line graphs show mean real-time ratings across 30-s clips, with each colored line representing a specific video from the corresponding category (as color-coded in Fig. 1). Gray-shaded regions indicate the initial 5 s of each video, which were excluded from statistical analysis. Boxplots summarize the distribution of average tingling intensities for five clips within each category across participants. Mic videos yielded the highest overall tingling ratings, followed by Eating, Proximal, and Nature.

Full size image

Affective touch task

Pleasantness ratings across skin areas and stroke velocities are shown in Fig. 3. A 2 (skin area) × 5 (velocity) repeated-measures ANOVA was conducted on pleasant ratings. The analysis showed no significant difference in overall pleasantness ratings between the forearm (mean ± SD: 3.57 ± 1.67) and the palm (3.87 ± 1.83): F(1, 45) = 2.30, p = 0.14, \({\eta }_{\text{p}}^{2}\) = 0.049. However, there was a significant main effect of stroking velocity, F(4, 180) = 45.28, p < 0.01, \({\eta }_{\text{p}}^{2}\) = 0.50, with the following pattern: 9 cm/s (4.93 ± 2.06) and 27 cm/s (4.64 ± 2.22) > 3 cm/s (4.07 ± 1.98) > 1 cm/s (2.88 ± 1.81) > 0.3 cm/s (2.10 ± 1.59). A significant interaction between skin area and stroking velocity was observed: F(4, 180) = 3.84, p = 0.005, \({\eta }_{\text{p}}^{2}\) = 0.079. Post-hoc comparisons revealed that on the forearm, pleasantness ratings for stroking velocities of 3 cm/s and 9 cm/s differed significantly, while this distinction was not observed for the palm.

Fig. 3
figure 3

Pleasantness ratings for different stroking velocities on the forearm and palm. Pleasantness ratings increased with an increase in stroking velocities, regardless of skin area. Error bars represent the standard deviation of mean pleasantness ratings. ***p < 0.001 (Bonferroni-corrected).

Full size image

Affective touch sensitivity was calculated based on pleasantness ratings. The results showed that sensitivity was higher for the forearm (2.26 ± 1.45) than for the palm (1.75 ± 1.65): t(45) = 2.42, p = 0.020, Cohen’s d = 0.36. This finding is consistent with anatomical evidence suggesting a higher density of CT afferents in hairy skin. The richer CT innervation likely contributes to the enhanced affective touch sensitivity observed in the forearm. A moderate positive correlation was found between affective touch sensitivity on the forearm and palm (r = 0.58, p < 0.001), suggesting that individuals vary consistently in their responsiveness to gentle tactile stimulation across skin areas. Taken together, these results support the view that individuals who report stronger tingling sensations may also exhibit heightened affective touch sensitivity. Although the present study did not classify participants as ASMR responders or non-responders, such associations may partly underlie individual differences in ASMR susceptibility.

Heartbeat counting task

Differences between actual and reported heart rates are shown in Fig. 4. A 2 (condition: actual vs. reported) × 3 (interval) ANOVA was conducted on heart rate. Reported heart rate (mean ± SD: 48.8 ± 23.0 bpm) was significantly lower than actual heart rate (73.4 ± 8.6 bpm): F(1, 42) = 47.55, p < 0.001, \({\eta }_{\text{p}}^{2}\) = 0.53. Participants underestimated their actual heart rate by more than 30% on average. The coefficient of variation (CV) differed markedly between conditions, with a CV of 0.12 for actual heart rate and 0.47 for reported heart rate. The main effect of interval was also significant, F(2, 84) = 7.15, p = 0.001, \({\eta }_{\text{p}}^{2}\) = 0.15, and a significant interaction between condition and interval was observed, F(2, 84) = 9.48, p < 0.001, \({\eta }_{\text{p}}^{2}\) = 0.18. Post-hoc comparisons revealed that reported heart rate was consistently lower than actual heart rate across all intervals. These results indicate a robust tendency for participants to underestimate their actual heart rate, consistent with previous findings37,38. This underestimation highlights potential differences in the processing of interoceptive signals and may reflect systematic biases in self-perceived physiological states.

Fig. 4
figure 4

Comparison of actual and reported heart rates. Reported heart rate was consistently lower than actual heart rate across 25-s, 45-s, and 65-s intervals. Error bars represent the standard deviation of mean heart rates. **p < 0.01, *p < 0.05 (Bonferroni-corrected).

Full size image

Tingling intensity as a function of sensory traits

To investigate whether affective touch sensitivity relates to ASMR tingling experiences, we conducted Spearman’s rank correlations between pleasantness ratings and tingling intensity for each video category (Fig. 5). Significant correlation was found for the Eating condition (rs = 0.37, p = 0.012, 95% CI [0.10, 0.57]), which remained significant after Bonferroni correction (adjusted α = 0.0125). The other conditions showed trends in the expected direction, but did not reach statistical significance: Mic, rs = 0.25, p = 0.091, 95% CI [− 0.09, 0.57]; Proximal, rs = 0.24, p = 0.109, 95% CI [− 0.05, 0.50]; Nature, rs = 0.25, p = 0.088, 95% CI [− 0.08, 0.51]. These results suggest that individuals with heightened affective touch sensitivity on the forearm tend to report more intense tingling sensations in response to ASMR stimuli. Full results for palm touch sensitivity, which did not show statistically meaningful associations, are presented in Supplementary Fig. 1.

Fig. 5
figure 5

Scatterplots showing associations between tingling intensity and sensory traits across four video categories. Points indicate individual participants. Spearman’s rank correlation coefficients (rs) are reported in each panel. Trend lines were estimated using LOWESS (locally weighted scatterplot smoothing) to illustrate monotonic but potentially non-linear associations. *p < 0.05 (Bonferroni-corrected).

Full size image

We also examined whether interoceptive accuracy was associated with tingling intensity across video categories. As shown in Fig. 5, all correlations were positive but did not reach statistical significance: rs < 0.26, p > 0.093. These findings suggest that while interoception may play a role in ASMR responsiveness, its effect appears weaker than that of affective touch sensitivity, particularly on the forearm.

To better understand the relative independence of these two sensory traits, we examined whether interoceptive accuracy and affective touch sensitivity were themselves correlated. No significant correlations were observed: r = 0.093, p = 0.55 for the forearm; r = 0.083, p = 0.60 for the palm. These null findings suggest that two processes may rely on distinct neural mechanisms. Interoceptive accuracy, often measured using heartbeat perception tasks, reflects the monitoring of internal bodily states (e.g., cardiac signals)39. In contrast, affective touch sensitivity is mediated by CT afferents and is grounded in the processing of external, socially relevant tactile cues, particularly in hairy skin areas20. These different mechanisms may account for their differential associations with ASMR experiences.

Modeling ASMR tingling from affective and interoceptive sensitivity

Although correlation analyses demonstrated that affective touch sensitivity is more strongly associated with tingling sensations than interoceptive accuracy, it remained unclear how these two sensory traits jointly contribute to ASMR responsiveness. To address this question, we conducted separate multiple linear regression analyses for each stimulus category (Mic, Proximal, Eating, and Nature), using forearm touch sensitivity and interoceptive accuracy as predictors of tingling intensity. This approach allowed us to assess the independent contribution of each trait variable while accounting for shared variance among predictors. Palm touch sensitivity was excluded from regression models, as it showed no meaningful association with tingling ratings in correlation analyses and is less relevant from the perspective of CT afferent distribution. Regression models including palm touch are reported in Supplementary Table 1 for completeness.

The regression results are summarized in Table 1. For the Eating condition, the model was significant, F(2, 40) = 5.02, p = 0.011, R2 = 0.200, with forearm touch sensitivity emerging as a significant predictor (β = 0.380, p = 0.011). In contrast, regression models did not reach statistical significance for the other three conditions: Mic, F(2, 40) = 1.07, p = 0.354, R2 = 0.051; Proximal, F(2, 40) = 2.91, p = 0.066, R2 = 0.127; Nature, F(2, 40) = 1.74, p = 0.188, R2 = 0.080. These results suggest that, among four stimulus categories, tingling sensations evoked by eating-related ASMR content are most strongly associated with individual differences in affective tactile sensitivity.

Table 1 Multiple regression models predicting ASMR tingling intensity across stimulus categories (N = 43).
Full size table

Discussion

This study examined relationships among tingling intensity, affective touch sensitivity, and interoceptive accuracy to investigate somatosensory and interoceptive mechanisms underlying individual differences in ASMR experiences. Among four video categories, a significant positive correlation was observed between tingling intensity and affective touch sensitivity in the Eating condition, even after correction for multiple comparisons. Similar positive trends were observed for the Mic and Nature conditions, although they did not reach statistical significance under the corrected threshold. These findings suggest that affective tactile processing may contribute to ASMR responsiveness, particularly in response to eating-related stimuli. In contrast, no significant correlations were found between tingling intensity and interoceptive accuracy, indicating that affective touch sensitivity may play a more prominent role than interoceptive processes in shaping ASMR experiences.

The positive correlation between tingling intensity and affective touch sensitivity suggests an overlap in sensory mechanisms underlying these phenomena. Individuals with heightened sensitivity to affective touch may be predisposed to stronger ASMR sensations when exposed to typical triggers. This association could be explained by the role of the CT afferent system, which mediates emotional comfort and relaxation during gentle tactile stimulation20. These findings are consistent with previous studies indicating that ASMR responders often show heightened sensitivity to positive social touch and mirror-touch synesthesia7,8. Individuals who are more attuned to interpersonal touch experiences, such as gentle caresses, also experience heightened emotional responses to ASMR triggers12. Although ASMR and affective touch involve distinct sensory modalities, their shared ability to evoke emotional regulation and relaxation through sensory processing suggests a common underlying mechanism40, which may account for observed correlations.

Although ASMR is typically elicited by auditory or visual stimuli33, accumulating evidence suggests that it recruits the same somatosensory and emotional systems activated during affective touch. This notion points to a shared neurobiological basis3,41. Such a connection is consistent with the social grooming hypothesis, which posits that affective touch evolved as mechanisms to strengthen affiliative bonds in primates14. ASMR may simulate this affiliative function, evoking feelings of social connection and emotional comfort through neural pathways originally adapted for tactile interaction, even in the absence of direct physical contact. In this view, ASMR could serve as a modern analog to grooming behaviors, using auditory and visual triggers to replicate comforting and bonding effects of touch.

This functional resemblance is further supported by a neurobiological model25, which suggests that ASMR arises from synesthetic cross-activation between the primary auditory cortex and the posterior insular cortex—particularly its map for affective touch. According to the model, auditory stimuli can activate interoceptive representations of gentle, affective touch. These signals are then integrated within the anterior insula to generate a “global emotional moment,” characterized by increased parasympathetic tone and enhanced subjective wellbeing. This model offers a plausible framework for understanding how ASMR evokes emotionally rich bodily sensations in response to ordinary auditory cues. It also complements our empirical findings, which highlight the role of affective touch sensitivity in modulating tingling sensations.

Tactile stimuli that simulate touch are among the most potent ASMR triggers, often evoking intense tingling sensations and underscoring the critical role of affective touch in ASMR experiences18. This tactile route to relaxation may be biologically grounded: affective touch has been associated with the release of oxytocin, a neuropeptide linked to relaxation and emotional comfort, both of which are frequently reported by ASMR experiencers42. Furthermore, affective touch has been shown to reduce pain perception, a finding that is consistent with reports of ASMR providing relief for chronic pain1. These parallels suggest that ASMR and affective touch may engage overlapping physiological mechanisms that support calming and analgesic effects—possibly through oxytocin-mediated pathways. Taken collectively, these findings emphasize that individual differences in tactile processing significantly influence ASMR intensity. Moreover, this influence is not limited to the tactile domain; it extends to auditory and visual modalities, highlighting the inherently multisensory nature of ASMR.

Interoceptive accuracy did not show a significant correlation with tingling intensity, contrasting with the observed relationship involving affective touch sensitivity. This finding suggests that ASMR may depend more on exteroceptive sensory processes, rather than on interoceptive processes, particularly interoceptive accuracy. Recent studies indicate that interoceptive contributions to ASMR may be more nuanced than initially assumed. For example, although no significant differences in interoceptive awareness were found between ASMR responders and non-responders, larger heartbeat-evoked potential amplitudes were observed in ASMR responders, suggesting a potential role for unconscious interoceptive processes27. Moreover, some researchers identified a negative relationship between ASMR and interoceptive awareness, proposing that ASMR responders may compensate for reduced interoceptive precision by relying more heavily on exteroceptive cues18. It is important to note, however, that interoceptive accuracy and interoceptive awareness are distinct constructs43. As such, the influence of interoception on ASMR responsiveness warrants careful examination to avoid conflating these measures and their respective contributions.

Another notable finding was the lack of significant differences in pleasantness ratings between the forearm and the palm. Previous studies have emphasized the role of CT afferents in hairy skin, such as the forearm. However, tactile pleasantness in glabrous skin, such as the palm, has been increasingly attributed to alternative pathways, such as Aβ afferents44,45. Moreover, affective touch pleasantness is known to vary across individuals and contexts, influenced by factors such as attention, emotional associations, and multisensory integration. High pleasantness ratings observed for the palm in this study likely reflect the involvement of both Aβ-mediated mechanisms and additional factors, including familiarity with and emotional engagement in the stimuli46,47. These findings suggest that tactile pleasantness is shaped by a complex interplay of neural and contextual factors, extending beyond the anatomical distribution of CT afferents.

Interestingly, among four stimulus categories, only the Eating condition yielded a significant regression model, with affective touch sensitivity emerging as a significant predictor of tingling intensity. This finding may reflect the multisensory richness and social familiarity associated with eating sounds, which are often accompanied by implicit tactile expectations (e.g., chewing, licking, or the sensation of food texture). In contrast, although the Mic and Proximal categories were designed to simulate tactile proximity, they may have elicited more varied perceptual interpretations or depended more on auditory spatial cues, thereby reducing the influence of touch-related traits on ASMR responsiveness. The Nature category, lacking direct human interaction cues, likely evoked weaker tactile-emotional associations overall. These results suggest that affective touch sensitivity plays a particularly salient role in shaping ASMR responsiveness when the stimuli are strongly linked to embodied, socially grounded experiences.

This study has some limitations that suggest valuable directions for future research. The ASMR task employed commonly recognized triggers; however, individual differences in trigger preferences indicate that less conventional or niche stimuli may have been overlooked. This omission could potentially limit the generalizability of our findings to the broader spectrum of ASMR experiences. Beyond the digital and laboratory contexts, cross-cultural studies of children’s folklore have documented traditional touch-based practices—known as folk illusions—performed on the forearms and palms, which closely resemble ASMR-inducing tactile experiences48. These examples suggest that ASMR-like sensations are not limited to digital media but may reflect enduring, culturally embedded forms of social touch. Future research may benefit from integrating these embodied practices to deepen our understanding of ASMR across developmental, cultural, and sensory domains.

In the affective touch task, manual application of tactile stimuli introduced variability in stroke force and velocity, which may have influenced pleasantness ratings. Automating the delivery of tactile stimuli with devices that precisely control speed and force could mitigate this issue. Additionally, a small subset of participants reported sensations of pruritus during the task, probably affecting their ability to provide accurate assessments of pleasantness. Nevertheless, the overall pattern of pleasantness ratings remained robust across participants, suggesting that these minor sources of variability did not substantially undermine the reliability of the results.

The heartbeat counting task also presented methodological challenges. Participants often reported fatigue or disengagement during extended intervals, which might have impaired their ability to accurately perceive their heartbeats. Furthermore, there is the possibility that variables such as body mass index affect interoceptive accuracy49. Although such tasks can help identify alternative strategies (e.g., reliance on time perception) that may influence heartbeat counting accuracy, we chose to omit this component in order to reduce participant burden and maintain procedural simplicity. Future research may benefit from including a time estimation task to further dissociate interoceptive accuracy from cognitive estimation strategies.

Previous research has rarely examined ASMR through the lens of contrasts between exteroceptive and interoceptive mechanisms. The present study contributes to this gap by elucidating how individual differences in affective touch sensitivity and interoceptive accuracy relate to ASMR responsiveness. Our findings emphasize the central role of sensory mechanisms—particularly those related to affective touch—in shaping tingling sensations during ASMR experiences. Although interoceptive accuracy was not significantly associated with ASMR intensity, this does not preclude a role for other interoceptive dimensions, such as implicit bodily awareness or affective integration. Moreover, our results highlight the importance of contextual factors and individual variability in shaping ASMR, reinforcing the need for more personalized approaches to leveraging this phenomenon. Collectively, these findings refine our understanding of sensory mechanisms of ASMR and support its potential utility in promoting wellbeing, reducing stress, and informing therapeutic applications.