Interoception is associated with the impact of eye contact on spontaneous facial mimicry

The study investigates how interoception relates to two foundational components of social cognition: the self-other boundary (the degree of overlap or distinction in processing self and others) and sensitivity to socially relevant cues. Prior work links interoception to emotional awareness and self-consciousness, and suggests associations with empathy and mentalizing, but mechanisms remain unclear. Social cognition depends not only on understanding others’ mental states but also on motor recognition, self-other boundary regulation, and detection of social cues. The present study examines whether individual differences in interoceptive accuracy (IAc), measured by a heartbeat counting task, are associated with spontaneous facial mimicry (SFM) as an implicit index of self-other overlap and with the enhancement of SFM by eye contact (direct vs. averted gaze) as an index of social-cue sensitivity. The authors hypothesize: (1) If IAc relates to self-other overlap (or distinction), it would be positively (or negatively) correlated with SFM; (2) If IAc relates to social-cue sensitivity, it would be positively correlated with the magnitude of the eye contact effect on SFM.

The paper reviews mixed evidence on how interoception relates to self-other boundary processes. Studies using imitation inhibition and automatic imitation tasks suggest higher interoceptive sensitivity is linked to greater susceptibility to others’ actions (greater overlap), while other work using imitation inhibition and body ownership illusions (rubber hand, enfacement) links higher interoception to greater self-other distinction. These inconsistencies may reflect task dependencies. Spontaneous mimicry offers a task-independent, implicit measure. Social-cue sensitivity is also reviewed with eye gaze as a canonical social signal: direct gaze captures attention, increases positive social judgments, and enhances mimicry. The authors draw on proposals that social cues may upweight interoceptive processing, potentially tuning socio-affective learning, but direct evidence is scarce. Thus, they propose using the eye contact effect on SFM as an index of social-cue sensitivity to test its association with interoception.

Participants: 80 Japanese adults (45 males; mean age = 24.50 years, SD = 9.53, range 15.30–57.70), with no reported psychiatric or neurological history. Ethics approval was obtained (The University of Tokyo, No. 18-50); written informed consent was provided. Procedure: Participants completed a heartbeat counting task, then an SFM task, followed by a self-report questionnaire (alexithymia; details in Appendix A). Heartbeat counting task (IAc): Participants sat upright and silently counted consciously perceived heartbeats during four pseudorandomized intervals (25, 35, 45, 100 s), guided by start/stop tones, without palpation. ECG was recorded via a wearable system (myBeat; Union Tool, Tokyo). Mean heart rate: 74.40 bpm (SD = 10.44, range 51.32–100.60). Two participants were excluded (experimenter error; pulse counting using hands), yielding N = 78 for IAc. IAc was computed using the standard heartbeat counting accuracy formula, scaled 0–100 (higher scores indicate greater accuracy). SFM task: Participants passively viewed short videos of adult Japanese models (male/female) whose facial expression changed from neutral to happy. Two gaze conditions: direct gaze (eye contact) vs. averted gaze (control). Stimuli were presented in 4 blocks (16 trials total; fixed within-block model order), with two blocks per gaze condition; block order counterbalanced. Each trial (2500 ms) comprised a neutral phase (1000 ms) followed by a happy expression (1500 ms); a fixation (500 ms with click) preceded each trial; intertrial gray background 5000 ms. Participants’ faces were video-recorded. Two trained, condition-blind coders applied FACS to identify lip-corner pull (AU12) as the SFM marker while the model displayed the happy expression; baseline was defined during the model’s neutral phase. Intercoder reliability (25% subset): Cohen’s κ = 0.925; primary coder’s data were used. SFM frequency per condition was the count of AU12 occurrences; the eye contact effect index was the difference (direct − averted). Note: 34% of participants (27/80) showed no SFM in either condition; analyses including/excluding non-responders yielded equivalent correlation results. Data analysis: Within-subject paired t-test compared SFM frequency between gaze conditions. Associations among IAc and SFM measures (direct, averted, difference) were tested using Spearman rank correlations. Twelve correlations were assessed; Bonferroni-corrected α = 0.0042. Robustness checks with Pearson correlations and alternative multiple comparison corrections (Holm, FDR) yielded the same conclusions. Gender effects: Welch’s t-test compared IAc by gender; a two-way repeated-measures ANOVA tested Condition (direct vs. averted; within) × Gender (male vs. female; between) on SFM frequency; correlations were also examined within each gender group.

- Heartbeat counting task: Mean IAc = 67.95 (SD = 19.43; range 27.41–98.25; N = 78). No significant correlation between IAc and heart rate (p = 0.106; 95% CI for correlation [−0.39, 0.041]). - SFM eye contact effect: SFM frequency was significantly higher in the direct-gaze than the averted-gaze condition; paired t-test t(79) = 4.501, p < 0.001, r = 0.45. - Correlations (Spearman): - Direct-gaze condition: IAc positively correlated with SFM frequency (r = 0.574, p < 0.001; 95% CI [0.40, 0.71]). - Averted-gaze condition: No significant correlation (r = 0.030, p = 0.794; 95% CI [−0.19, 0.25]). - Eye contact effect (direct − averted SFM): Positive correlation with IAc (r = 0.603, p < 0.001; 95% CI [0.43, 0.73]). - Gender analyses: - IAc did not differ by gender (males: M = 71.25, SD = 18.00; females: M = 63.44, SD = 20.65); Welch’s t(76) = 1.780, p = 0.079. - ANOVA on SFM: Main effect of Condition, F(1,78) = 18.758, p < 0.001, ηp² = 0.19; no Gender main effect, F(1,78) = 0.012, p = 0.912; no Condition × Gender interaction, F(1,78) = 1.337, p = 0.251. - Correlations within gender mirrored overall pattern: - Eye contact effect vs. IAc: males r = 0.461, p = 0.001 (95% CI [0.19, 0.66]); females r = 0.688, p < 0.001 (95% CI [0.45, 0.83]). - Direct-gaze SFM vs. IAc: males r = 0.430, p = 0.003 (95% CI [0.16, 0.64]); females r = 0.562, p = 0.001 (95% CI [0.27, 0.76]). - Averted-gaze SFM vs. IAc: males r = 0.021, p = 0.889 (95% CI [−0.27, 0.31]); females r = 0.160, p = 0.366 (95% CI [−0.19, 0.48]).

Findings show that interoceptive accuracy is associated with sensitivity to social cues: individuals with higher IAc exhibited a larger enhancement of spontaneous facial mimicry under direct gaze relative to averted gaze. The positive correlation between IAc and SFM occurred only under direct gaze, suggesting that any link between interoception and self-other overlap is context-dependent and may be driven by socially relevant signals. The authors discuss potential mechanisms: direct gaze may heighten arousal or selectively allocate attention, amplifying mimicry particularly among individuals with higher interoceptive sensitivity. Prior work showing that direct gaze can enhance IAc (without heart rate changes) supports a bidirectional relationship between gaze processing and interoception. Alternative interpretations are acknowledged: SFM might reflect general affective responses to smiling faces rather than self-other matching, and the eye contact effect might index the triggering of deeper embodied simulation rather than a pure social-cue sensitivity construct. Overall, the results address the hypotheses by demonstrating a robust association between IAc and the eye contact effect on SFM, while the association between IAc and self-other boundary per se appears to depend on social context.

The study advances understanding of how interoception relates to social cognition by showing that interoceptive accuracy predicts individual differences in the eye contact enhancement of spontaneous facial mimicry. This supports a role for interoception in sensitivity to social cues and suggests that any link to self-other overlap is modulated by context (e.g., gaze). Future research should: (1) disentangle task/context effects by probing multiple social cues (gaze, actions, voice, touch) and measures of social-cue sensitivity (attentional bias, detection thresholds, memory impacts); (2) examine contextual modulation of self-other boundary and the potential indirect contribution of interoception; (3) integrate physiological and neural measures (electrodermal activity, heartbeat-evoked responses) to parse arousal and attention contributions and to more directly index interoceptive processing.

- The design cannot determine whether the direct-gaze correlation reflects a default association or whether the effect is specifically driven by eye contact; causality cannot be inferred from correlational analyses. - SFM may index affective responses or embodied simulation rather than a pure measure of self-other boundary; functional interpretation of SFM and the eye contact effect is not unique. - A substantial subset of participants (34%) showed no SFM in either condition; although results were robust to inclusion/exclusion, non-responsiveness may limit sensitivity. - Potential arousal- and attention-related confounds were not directly measured; physiological validation is needed. - Generalization is limited to the tested stimuli and context (Japanese adult models; happy expressions; brief dynamic clips).