Heartbeat-evoked potentials during interoceptive-exteroceptive integration are not consistent with precision-weighting

Abstract

AbstractInteroceptive-exteroceptive integration is fundamental for a unified interactive experience of the world with the body. Predictive coding accounts propose that these integrated signals operate predictively, with regulation by precision-weighting. Heartbeat-evoked potentials (HEPs) are one means to investigate integrated processing. In a previous study, consistent with predictive coding characterisations of precision-weighting, we observed modulation of HEPs by attention. However, we found no evidence of HEP modulation by participants’ interoceptive ability, despite the characterisation by predictive coding theories of trait abilities as a similar reflection of differential precision-weighting. In this study, we sought to more sensitively test the hypothesised trait-precision influences on HEPs by using an individually-adjusted measure of interoceptive performance. However, contrary to a precision-weighted predictive coding framework, we failed to find evidence in support of the HEP modulations by attentional-precision or trait-precision. Nonetheless, we observed robust HEP effects indicative of an expectation of a sound on the basis of a heartbeat -i.e. interoceptive-exteroceptive integration. It is possible that under our more individually-tailored task, participants relied less on attentional-precision to ‘boost’ predictions due to an enhanced perception of cardio-audio synchrony. Furthermore, assessing interoceptive ability is challenging, thus variations in performance may not accurately reflect trait-precision variations. Nevertheless, in sum, our findings are inconsistent with a precision-weighted prediction error view of the HEP, and highlight the need for clearer definitions of the manipulation and measurement of precision in predictive coding. Finally, our robust interoceptive-exteroceptive integration HEP effects may provide a valuable tool for investigating such integration in both clinical conditions and cognition.Impact statementWe investigate heart-evoked potentials during interoceptive-exteroceptive integration to determine whether cross-modal integrated processes operate under a precision-weighted predictive coding framework. Using a more sensitive individually-tailored task, we found no evidence of the modulation of cardio-audio expectation by attention or individual differences in interoceptive perception (i.e. by state or trait measures of precision). Nonetheless, we replicate evidence of cardiac-driven predictions of auditory stimuli, providing a potential tool for investigating their relationship with emotion and embodied selfhood.