Associations between self-reported interoception and resting-state EEG markers in panic disorder: heartbeat-evoked...

Panic disorder (PD) is associated with altered interoceptive experience, but it remains unclear how self-reported interoception relates to neurophysiological indices of bodily-signal processing. This study examined whether associations between self-reported interoception and resting-state electroencephalography (EEG) markers differ between participants with PD and healthy controls (HC).

This analysis included 19 PD participants and 21 HCs from a previously published resting-state EEG dataset. Self-reported interoception was assessed using the Body Perception Questionnaire-Very Short Form (BPQ-VSF) and the Multidimensional Assessment of Interoceptive Awareness (MAIA). Resting-state EEG signals were recorded during eyes-closed and eyes-open conditions. Heartbeat-evoked potentials (HEP) and relative band power (RBP) were extracted to index cardiac-related cortical processing and frequency-specific resting-state activity. Associations between self-report measures and EEG indices were examined using Pearson correlations with 95% confidence intervals, Bayes factors, and false-discovery-rate (FDR) correction across scalp regions within predefined analysis families. Age-adjusted sensitivity analyses and group × self-report interaction models were used to assess robustness and formally test group differences in association patterns.

Compared with HCs, PD participants reported higher body-focused attention and lower scores on several adaptive interoceptive dimensions, including Not-Distracting, Not-Worrying, and Trusting. In the 235-301 ms HEP window, age-adjusted interaction models showed that BPQ-VSF HEP coupling differed between groups, especially under eyes-open conditions. Within-group analyses showed an exploratory opposite-direction BPQ-VSF HEP pattern, with positive associations in HC and negative associations in PD. The clearest FDR-corrected HEP association within panic disorder involved Not-Worrying under eyes closed. RBP analyses showed PD-concentrated θ (theta)- and β (beta)-band association profiles. Trusting was positively associated with θ power in PD, and BPQ-VSF showed positive β-band associations under eyes closed that remained robust after age adjustment. Not-Distracting showed negative θ-band associations in the primary correlation analyses, but these were attenuated after controlling for age.

These findings provide preliminary FDR-controlled and age-informed evidence that PD involves altered, condition- and frequency-dependent mappings between subjective bodily experience and resting-state EEG markers. The clearest formal group-difference evidence was observed for BPQ-VSF HEP coupling, whereas PD-specific associations were most robust for Not-Worrying-HEP, Trusting-θ power, and BPQ-VSF β power. Given the modest sample size and exploratory correlational design, replication in larger samples is needed.