Olivier Desmedt
A Belgian/Swiss psychologist (UCLouvain → University of Lausanne) who has become the field’s most systematic measurement critic of interoception. Where most of the wiki’s researchers propose what interoception is or does, Desmedt’s programme is adversarial and psychometric: it asks whether the words “interoceptive accuracy,” “sensibility” and “awareness” name anything a measure actually captures, and concludes largely that they do not.
The position
Two claims, both developed at length in Desmedt et al. (2023) and both now load-bearing on this wiki:
-
The definition is self-contradictory. Researchers define interoception both as the processing of internal signals (phenomenon-based) and as the activation of homeostatic pathways (physiological), in the same breath. Since non-homeostatic Aβ somatosensory pathways also carry internal-state information — demonstrably, in the cardiac, respiratory and gastric systems — no physiological pathway is necessary and sufficient, so physiology cannot ground the definition. See interoception-exteroception-boundary.
-
The dimensions are detached from the measures. The consensus taxonomies are too broad; the measures assigned to a single dimension converge weakly (the HCT and HDT share 4.4% of variance; sensibility questionnaires split into five factors). Because low convergence makes cross-measure inconsistency statistically expected (Carlson & Herdman 2012), giving different measures the same label manufactures a spurious replication crisis. See interoceptive-taxonomy, is-interoception-domain-general.
The constructive half is a phenomenon-based definition (“processes … from within itself and below the skin, across conscious and non-conscious levels”) and a hierarchical framework organised by construct specificity rather than processing depth.
The empirical spine
Desmedt’s critique is not armchair. His earlier work (with Olivier Corneille) is the source of the wiki’s cleanest evidence that the heartbeat counting task is contaminated by prior beliefs about heart rate rather than perception — the guessing-strategy mechanism at the heart of is-the-heartbeat-counting-task-valid. Alongside it sits Desmedt et al. (2022), the criterion-validity half of the case: a pre-registered meta-analysis of 133 studies / 11,524 adults in which HCT performance predicts none of the psychological outcomes (trait anxiety, depression, alexithymia) the “accuracy” label was built to explain, while tracking only cardiac-signal/physiological variables (heart rate, BMI, sex) — the guessing/cardiodynamic pattern at scale. He is thus one of the rare figures who both makes the conceptual argument and supplies the data that forces it.
Relation to the figures he critiques
Notably collegial for an adversary: sahib-khalsa is thanked for early feedback on the 2023 review that dismantles his roadmap, and Desmedt treats sarah-garfinkel & hugo-critchley’s dimensional model as the best available target rather than a straw man — his hierarchical framework is offered as its successor, not its refutation. His nearest ally in spirit is gary-berntson & Khalsa’s own admission that interoceptive “sensitivity is not a generalizable feature.”