Research gap at a glance
- Question: When objective interoceptive accuracy is measured repeatedly within the same people over days, what share of its variance lies within rather than between persons, and does that share differ from believed accuracy and from bodily attention?
- Why it matters: Nearly every correlation in the wiki treats a single-session accuracy score as a trait. If accuracy behaves like the self-report measure that has been partitioned, single-session estimates carry an error as large as the signal and small correlations in disputes such as Pollatos versus Dunn are attenuated by an unknown amount.
- Theme: Measurement
- Status: open
What the wiki already knows
- Self-reported interoception sampled ten times a day for a week in 70 adults had an intra-class correlation of 0.51, so about half its variance was within people (Wallman-Jones et al. (2023), State vs trait interoception).
- Of 17 panic patients accurate at baseline, only 8 were accurate after treatment (Does et al. (2000)); within persons, counting error fell as state anxiety rose (Zoellner & Craske (1999)).
- One acute cold pressor raised counting accuracy and lowered discrimination accuracy in the same session, so the accuracy score is demonstrably movable within a session (Schulz et al. (2013)).
- Believed accuracy may be more dispositional: the Interoceptive Accuracy Scale (IAS) had a 30-day test-retest of r = .754, and trait and state versions agreed in one study but not another (Murphy et al. (2020)).
- No study has run an objective heartbeat or Filter Detection Task (FDT) repeatedly within days and partitioned the variance (State vs trait interoception).
What would answer it
- Administer short heartbeat-discrimination and respiratory-threshold blocks several times a day for a week (or across many lab sessions) to a few hundred adults, alongside momentary ratings of believed accuracy and bodily attention, using Ecological Momentary Assessment (EMA) and multilevel models to estimate the intra-class correlation for each construct and to test whether within-person fluctuation in accuracy tracks heart rate, physical activity or state anxiety.
Connections
- Raised on: State vs trait interoception, Schulz et al. (2013), Wallman-Jones et al. (2023), Emotional granularity
- Related concepts: State vs trait interoception, Interoceptive sensitivity, Interoceptive propensity, Physical activity, sedentary behaviour and interoception
- Related debates: Is the heartbeat-counting task a valid measure of cardiac perception?, Is more interoceptive awareness better?
State vs trait interoception makes the case that the field writes about interoception as a property of persons while its one variance partition says half of it belongs to moments. That partition was of a self-report attention measure, borrowed from a mindfulness scale, in young healthy adults. The page’s own verdict on what is missing for accuracy is blunt: essentially everything.
The scattered evidence on accuracy points the same way without settling it. Panic patients drift in and out of the accurate category across treatment, counting error moves with momentary anxiety, and a single stressor moves counting and discrimination scores in opposite directions. Each of these is open to the cardiodynamic reading, where the heart rather than the perceiver changes. For a state account that does not matter, because the quantity really moves; for an accuracy interpretation it is decisive, which is why heart rate belongs in the design.
There is also a construct-specific hypothesis worth testing. Murphy et al. (2020) suggest how accurately people believe they read their body may behave like a disposition while how much attention is on the body right now behaves like a state. The same open question is recorded for Emotional granularity and for Interoceptive propensity, so one repeated-measures design could partition several constructs at once.