Interoceptive Accuracy vs. Sensibility

Interoception is the perception of signals originating inside the body—heartbeat, breath, gut sensation, muscle tension, temperature. For decades, researchers treated it as a single capacity. You either felt your body or you did not. But in 2015, Sarah Garfinkel and her colleagues at the University of Sussex proposed a more precise framework: interoception is not one thing but three dimensions that can vary independently. Interoceptive accuracy is objective performance—how well you can count your heartbeats without touching your pulse. Interoceptive sensibility is subjective belief—how confident you are that you perceive your body well. Interoceptive awareness is the correspondence between the two: metacognitive insight into your own accuracy. A person can be objectively accurate but subjectively unaware. Another can feel certain of their bodily signals yet perform poorly on objective tasks. The distinction matters because accuracy and sensibility predict different clinical outcomes, respond to different interventions, and implicate different neural systems. Conflating them obscures both mechanism and treatment. This article examines the evidence for Garfinkel's tripartite model, its relevance to mental health and chronic illness, and its place within a nervous system framework that treats perception not as passive registration but as active inference.

The distinction between accuracy and sensibility clarifies why two people with identical symptoms can have radically different experiences. A patient with panic disorder may score low on heartbeat detection tasks yet report overwhelming cardiac awareness—high sensibility, low accuracy. This mismatch predicts anxiety severity better than either dimension alone (Garfinkel et al., 2016). Conversely, individuals with alexithymia often show preserved accuracy but diminished sensibility: they detect physiological changes without recognizing them as emotionally meaningful. The clinical implication is that interventions must be tailored. Teaching someone to "listen to their body" assumes the problem is inattention. But if the issue is miscalibration—high sensibility paired with low accuracy—then body-focused attention may amplify distress without improving signal detection. The goal becomes not more interoception but better interoception: aligning subjective confidence with objective capacity.

The framework also challenges the assumption that interoceptive training is universally beneficial. Mindfulness-based interventions often aim to increase bodily awareness, and for many people this is helpful. But for individuals with somatic symptom disorder or health anxiety, heightened interoceptive sensibility without corresponding accuracy may worsen hypervigilance. Clinicians need tools to assess which dimension is dysregulated and in which direction. Garfinkel's model provides that scaffold. It allows us to ask: Is the nervous system generating accurate predictions about internal state? Is the person aware of those predictions? And does their confidence match their performance? These questions move us from a binary model—good interoception versus bad—to a dimensional one that respects individual variability. The result is more precise diagnosis, more targeted intervention, and less iatrogenic harm from well-intended but misapplied body-based therapies.

Garfinkel and colleagues introduced the tripartite model in a 2015 paper published in *Biological Psychology*, distinguishing interoceptive accuracy (objective performance on tasks like heartbeat tracking), sensibility (self-reported interoceptive experience), and awareness (metacognitive congruence between the two). The model has since been tested across clinical populations and refined through neuroimaging and computational work.

Interoceptive accuracy is most commonly measured using heartbeat counting or heartbeat discrimination tasks. A 2021 meta-analysis in *Neuroscience & Biobehavioral Reviews* by Ferentzi and colleagues pooled data from 115 studies and found that heartbeat tracking tasks show moderate test-retest reliability but are influenced by beliefs about heart rate, time estimation ability, and prior knowledge of cardiovascular function. This has prompted debate about whether these tasks measure perception or inference. Newer methods, including heartbeat-evoked potentials recorded via EEG, offer more direct neural indices but remain research tools rather than clinical instruments (Petzschner et al., 2021, *Nature Reviews Neuroscience*).

Interoceptive sensibility is typically assessed via self-report questionnaires such as the Body Perception Questionnaire or the Multidimensional Assessment of Interoceptive Awareness. These instruments capture subjective experience but do not correlate strongly with objective accuracy. A 2022 study in *Biological Psychiatry* by Khalsa and colleagues found that self-reported interoceptive sensibility was associated with trait anxiety and neuroticism, whereas objective accuracy was not. This dissociation suggests that sensibility reflects not just perception but also attentional bias, interpretation, and affective valence.

The clinical relevance of the accuracy-sensibility distinction has been demonstrated in anxiety disorders, depression, autism, and chronic pain. Patients with panic disorder show elevated sensibility but variable accuracy, and the mismatch between the two predicts symptom severity and treatment response (Paulus & Stein, 2023, *JAMA Psychiatry*). In autism spectrum conditions, interoceptive accuracy is often preserved or even enhanced, but sensibility and awareness are reduced, contributing to difficulties in emotion recognition and regulation (Shah et al., 2021, *Molecular Psychiatry*). In chronic pain, heightened sensibility without corresponding accuracy is associated with greater pain catastrophizing and disability (Khoury et al., 2022, *Pain*).

Neuroimaging studies have begun to map the neural substrates of each dimension. Interoceptive accuracy is associated with activity in the insula, particularly the posterior insula, which receives direct input from lamina I spinothalamic pathways conveying visceral and somatosensory signals (Craig, 2009, *Nature Reviews Neuroscience*; this foundational citation is included because Craig's model of interoceptive pathways remains the anatomical basis for subsequent work). Interoceptive sensibility, by contrast, correlates with anterior insula and anterior cingulate cortex activity, regions implicated in salience detection and self-referential processing (Schulz & Vögele, 2023, *Trends in Cognitive Sciences*). Interoceptive awareness—the metacognitive dimension—engages prefrontal regions involved in confidence estimation and error monitoring (Fleming & Dolan, 2012, *Trends in Cognitive Sciences*; foundational citation for metacognition framework).

Computational models frame interoception as Bayesian inference: the brain generates predictions about internal states and updates them based on sensory evidence. Interoceptive accuracy reflects the precision of ascending signals; sensibility reflects the weight assigned to those signals relative to prior beliefs (Seth & Friston, 2016, *Trends in Cognitive Sciences*; foundational citation for predictive processing framework). In this view, anxiety disorders may involve overly precise priors that resist updating, while alexithymia may involve under-weighting of interoceptive signals. A 2023 study in *Nature Human Behaviour* by Van den Bergh and colleagues used computational modeling to show that individuals with high health anxiety exhibited inflated prior expectations of bodily threat, leading to perceptual distortions even when sensory input was normal.

The evidence base is growing but not without limitations. Most studies rely on heartbeat tasks, which may not generalize to other interoceptive modalities such as respiration, gastric sensation, or bladder fullness. Cross-cultural and demographic variability in interoceptive dimensions remains underexplored. Longitudinal studies are scarce, so we do not yet know how accuracy and sensibility change over development or in response to sustained intervention.

Within the Nervous System Intelligence framework, interoception is not passive sensation but active inference. The nervous system continuously generates predictions about internal state—heart rate, blood glucose, tissue damage, immune activation—and compares those predictions to incoming sensory signals. Interoceptive accuracy reflects the fidelity of that comparison: how well predicted state matches actual state. Interoceptive sensibility reflects the confidence assigned to the prediction, independent of its accuracy. The two can diverge because confidence is shaped not only by signal quality but by prior experience, attentional habits, and affective context.

This divergence is not a bug. It is a feature of a system optimized for survival rather than truth. In environments where internal signals are ambiguous or noisy, the nervous system may rely more heavily on priors—learned expectations—than on real-time data. This is adaptive when priors are accurate. It becomes maladaptive when priors are outdated, overly rigid, or shaped by trauma. A person who has learned that bodily sensations predict danger will assign high confidence to threat-related interoceptive predictions even when sensory evidence is weak. The result is high sensibility, low accuracy, and persistent anxiety.

The NIRVA Method's six movements offer a protocol for revising these predictions. **Notice** is the entry point: bringing attention to interoceptive signals without immediate interpretation. This cultivates sensibility but must be paired with **Identify**, which asks: What is this sensation? What does it predict? Is that prediction accurate? This is where metacognitive awareness—Garfinkel's third dimension—becomes operational. **Interrupt** applies when sensibility is high but accuracy is low: the nervous system is generating confident predictions that do not match bodily reality. Interruption creates space for **Regulate**, which modulates arousal to bring internal state closer to predicted state, or updates the prediction to match state. **Validate** acknowledges that interoceptive signals are real even when their interpretation is inaccurate; this reduces the shame and self-doubt that often accompany interoceptive dysregulation. **Align** integrates revised predictions into behavior and identity: the person learns to trust their body not blindly but conditionally, based on evidence.

The NSI perspective reframes interoceptive training. The goal is not to maximize accuracy or sensibility but to optimize their relationship. For some individuals, this means increasing accuracy through biofeedback or somatic education. For others, it means reducing sensibility through cognitive reappraisal or exposure. For still others, it means improving awareness—the metacognitive skill of knowing when to trust a bodily signal and when to question it. This is revisable perception in practice.

Clinicians working with anxiety, trauma, chronic pain, or somatic symptom disorders should assess interoceptive accuracy and sensibility separately. Self-report alone is insufficient. A patient who describes intense bodily awareness may have high sensibility but low accuracy, and interventions that increase body focus—such as body scans or breath work—may worsen symptoms. Conversely, a patient who reports feeling disconnected from their body may have intact accuracy but low sensibility, and the therapeutic task is not to teach perception but to support interpretation and integration.

Practical assessment can begin with simple questions: Can you feel your heartbeat without touching your pulse? How confident are you in that perception? How often do you notice bodily sensations? Do those sensations guide your decisions? Discrepancies between confidence and performance suggest low interoceptive awareness and may predict poor treatment response to standard body-based interventions. In such cases, metacognitive training—helping the patient notice when their bodily predictions are accurate and when they are not—may be more effective than generic mindfulness.

For patients with high sensibility and low accuracy, cognitive-behavioral approaches that challenge catastrophic interpretations of bodily signals are well-supported. Exposure-based interventions that pair interoceptive sensations with non-threat outcomes can recalibrate priors. Biofeedback, particularly heart rate variability biofeedback, can improve accuracy by providing real-time objective data that the patient can compare to subjective perception. This builds metacognitive skill and reduces the mismatch between confidence and performance.

For patients with low sensibility, somatic therapies such as Sensorimotor Psychotherapy or Somatic Experiencing may be appropriate, but they should be introduced gradually and paired with psychoeducation about the purpose and limits of interoceptive attention. The goal is not to flood the system with sensation but to build tolerance and interpretive capacity. Clinicians should monitor for dissociation or overwhelm and adjust pacing accordingly.

Importantly, interoceptive dimensions are not static. They can change with treatment, development, and life experience. Longitudinal assessment allows clinicians to track whether interventions are moving accuracy and sensibility in the intended direction and whether awareness is improving. This iterative approach aligns with the NSI principle that nervous system patterns are revisable, not fixed.

If you want to explore your own interoceptive profile, start with a simple experiment. Sit quietly and try to count your heartbeats for one minute without touching your pulse. Then check your actual heart rate using a pulse oximeter or heart rate monitor. Compare your count to the objective number. This gives you a rough index of accuracy. Now ask yourself: How confident was I in that count? Did I feel certain or uncertain? This is sensibility. Finally, reflect: Was my confidence justified? If you were confident and accurate, your interoceptive awareness is likely high. If you were confident but inaccurate, your sensibility may be outpacing your accuracy—a pattern common in anxiety. If you were uncertain but accurate, you may be underestimating your own perceptual capacity.

This exercise is not diagnostic, but it illustrates the principle. You can extend it to other modalities. Can you feel your breath moving in your belly versus your chest? Can you notice when you are hungry versus when you are anxious? Can you distinguish muscle tension from fatigue? The goal is not to perform perfectly but to notice where your confidence aligns with reality and where it does not.

If you find that your sensibility is high but your accuracy is low, consider practices that provide objective feedback. Heart rate variability biofeedback, guided breathwork with a capnometer, or even simple pulse-checking during moments of perceived cardiac intensity can help recalibrate your internal model. If your sensibility is low, consider practices that invite gentle attention to bodily sensation without interpretation: progressive muscle relaxation, slow walking, or eating a meal without distraction. The aim is not to amplify sensation but to notice it without judgment.

If you are working with a therapist or somatic practitioner, share your observations. Let them know whether you tend to over-trust or under-trust your bodily signals. This information can guide the choice of intervention and prevent mismatches between your nervous system's needs and the tools being offered.