Kim & Kim (2025) — Neural Processes Linking Interoception to Moral Preferences

Full title: Neural Processes Linking Interoception to Moral Preferences Aligned with Group Consensus
Authors: JuYoung Kim and Hackjin Kim
Affiliation: Laboratory of Social and Decision Neuroscience, Korea University, Seoul
Journal: Journal of Neuroscience (Behavioral/Cognitive section)
Year: 2025
Type: Behavioral + neuroimaging (resting-state fMRI); two independent experiments
Ingested: 2026-06-15

Summary

• Experiment 1: Self-report interoceptive awareness (K-MAIA, 32-item Korean Multidimensional Assessment of Interoceptive Awareness) + moral dilemmas task (self-paced vignettes on Cognition.run server)
• Experiment 2: Interoceptive accuracy via heartbeat counting task (Schandry 1981; 6 trials, 25–50 s each; IAcc = 1 − |actual − reported heartbeats| / actual heartbeats) + fMRI (Siemens 3T, 304 volumes at TR=1.35s) + same moral task
• Moral measure: Group consensus calculated as ratio of participants choosing utilitarian option per scenario; individual moral similarity = percentage of choices matching the majority-preferred option

Key Findings

1. Both self-report interoceptive awareness and objective interoceptive accuracy predict moral alignment with group consensus (significant positive correlation, Bonferroni-corrected p < 0.05 in both experiments)
2. Neural mediation: Fractional occupancies of resting-state brain states involving the ventromedial prefrontal cortex (vmPFC) and the precuneus mediated the interoception → moral alignment relationship
3. The relationship held across male and female participants; replicated across two independent experiments

Theoretical Framing

The paper draws on allostasis and predictive processing: conforming to majority-approved moral preferences reduces prediction error and the metabolic cost of inferring others' expectations. Interoception, as the system that tracks the body's allostatic state, provides the grounding for these "internalized models of what the majority expects" — high interoceptive awareness means more accurate sensing of the allostatic cost of social misalignment.

The moral intuition is literally felt in the body; interoceptive fidelity determines how strongly that felt signal influences choice.

Neural Pathway Clarification

The mediation runs: Interoceptive ability → resting-state vmPFC/precuneus activity → moral choice alignment. The insular cortex is the primary substrate for interoceptive processing (and thus the upstream source of the signal), but the study's explicit mediators are vmPFC and precuneus — not the insula directly. This distinction matters: the insula is where internal body signals are integrated; the vmPFC/precuneus are where those signals are translated into social/moral valuation.

Relation to Existing Wiki Pages

• entities/insular-cortex.md — The Somatic Marker section ("Moral Compass") was written citing this paper; the more precise claim is that interoceptive accuracy (not insular activation per se) predicts moral alignment, mediated by vmPFC/precuneus
• concepts/neuroception.md — Moral/social conformity as neuroceptive survival strategy (feeling "out of sync" as threat); this paper provides empirical neural evidence for that claim
• concepts/theory-of-constructed-emotion.md — Confirms that bodily state information shapes social/moral construction, not just emotional construction

Pages Updated During Ingest

• wiki/sources/kim-kim-2025-interoception-moral.md — Full page rewritten with actual methods/findings
• entities/insular-cortex.md — note to be added clarifying vmPFC/precuneus as the explicit mediators (insula is upstream interoceptive substrate)

Open Questions

• Does interoceptive training (body scan, HRV biofeedback) increase moral alignment with group consensus — or might it also increase independent moral judgment by reducing allostatic pressure to conform?
• How does orchid vs. dandelion sensitivity interact with this? High-IS orchids presumably have stronger interoceptive signals and thus stronger moral conformity pressure
• The vmPFC is also the hub of the Default Mode Network's self-referential processing — what is the functional overlap with the DMN-somatic decoupling found in meditation research?