Technical comment on Gunderson, ten Brinke, and Sokol-Hessner (2023). When the body knows: Interoceptive accuracy enhances physiological but not explicit differentiation between liars and truth-tellers. Personality and Individual Differences, 204, 112039

Gunderson, ten Brinke, and Sokol-Hessner (Personality and Individual Differences, 204, 2023; henceforth: GBH) reported that persons who had perfect interoception of their bodily reactions could detect liars with an accuracy of up to 68 % - far exceeding the 50 % predicted by chance and well above what participants achieved in explicit, conscious judgments. GBH conclude that the body knows whether somebody is lying, although humans fail to leverage this knowledge in their explicit judgments. Because GBH used videos of true murderers lying about their deeds, the study has high practical relevance. GBH base their arguments on an upper bound intro-duced in this literature by Franz & von Luxburg (Psychological Science, 26, 2015). However, they misinterpret the meaning of this upper bound and therefore come to overly-optimistic estimates of lie-detection accuracy. Reanalysis of their data with appropriate methods shows that the bodily reactions only allowed close-to-chance lie-detection accuracy, similar to participants' performance in explicit judgments. We conclude that GBH do not provide evidence for appreciable lie-detection capability in bodily reactions. Quite to the contrary, the study demonstrates how difficult it is to detect a liar - even if a real-world, high stakes scenario is used.