Not Whether the Body Is Heard, but How It Is Trusted: Interoception and Suicide Risk
- Scoping review and meta-analysis led from the Vanke School of Public Health, Tsinghua University (Beijing); 66 studies and 36,934 participants (63 observational, of which 59 cross-sectional and 4 longitudinal, plus 3 intervention studies), 41 contributing to the meta-analysis; searched to 1 April 2025.
- Interoception was split into four validated dimensions: accuracy, sensibility, awareness, and an "Other" (cognitive-emotional evaluation of bodily signals) category.
- Interoceptive accuracy, how well a person objectively detects bodily signals, showed weak and non-significant associations with self-harm and suicidal outcomes.
- Interoceptive sensibility told the opposite story: greater trust in bodily signals and greater self-regulation through the body were associated with lower suicidal ideation, fewer attempts and lower composite suicidal outcomes, with small-to-moderate effect sizes; within the "Other" dimension, interoceptive deficits were positively associated with non-suicidal self-injury and ideation, and neuroimaging implicated the insula and prefrontal cortex.
Suicidology has spent years searching for a risk marker with more traction than a checklist. This synthesis, the first quantitative one on interoception and suicide, offers a candidate domain and, more usefully, corrects a naive version of it. The question that matters is not whether a patient can accurately count their heartbeat. It is whether they trust their body and can use it to regulate distress.
What the data shows
Across 66 studies and nearly 37,000 people, the four faces of interoception behaved very differently. Accuracy, the lab-task ability to detect internal signals, was essentially flat against self-harm and suicide. That is a striking null, because accuracy is the dimension researchers most often reach for. The signal lived in sensibility, the subjective, self-reported relationship to bodily experience. Two of its facets stood out: trusting bodily signals, and the capacity to regulate distress by attending to the body. Higher scores on both tracked lower ideation, fewer attempts and lower overall suicidal outcomes, with small-to-moderate effects. In the "Other" dimension, capturing distorted cognitive-emotional appraisal of bodily signals, deficits ran the other way, positively associated with non-suicidal self-injury and ideation. Neuroimaging studies converged, if preliminarily, on the insula and prefrontal cortex, the machinery that maps and evaluates internal state.
The design ceiling is real: 59 of 63 observational studies were cross-sectional, so directionality is unsettled and the effects are associations, not predictions.
For your practice
This reorients where to look. A body-detachment story, "I don't feel anything, I don't trust what my body tells me," is more risk-relevant here than a patient's performance on any heartbeat task. Assess the relationship to the body, not the acuity of it: does the patient trust bodily signals, or dismiss and fear them? Can they down-regulate through attention to the body, or does turning inward only amplify threat? That maps directly onto interventions already in the room, interoceptive exposure, grounding, body-based affect regulation, and suggests their target is trust and usability of bodily signals rather than detection accuracy. Treat this as a domain to explore in formulation, not a screening instrument; the evidence base is still largely cross-sectional.
For suicide risk, the meaningful interoceptive variable is not how precisely the body is detected but whether it is trusted and can be used to steady distress.
Most included studies were cross-sectional (59 of 63 observational), so the associations cannot establish that interoceptive change precedes changes in risk; the neuroimaging evidence is described by the authors as preliminary.