Abstract
Interoceptive accuracy is frequently assessed using the Heartbeat Counting Task (HCT), requiring participants to count the number of times their heart beats. The HCT validity has been questioned, as participants may perform the task by estimating, rather than counting, their felt heartbeats. Participants could estimate the time or use their knowledge of their heart rate. Some research ruled out the contribution of time estimation in HCT performance. However, we believe these studies relied on a problematic analytic rationale. We revisited this question by relying on new analytic strategies, and by examining the role of estimation in HCT performance, while varying task instructions. The findings support the role of time and knowledge-based estimations under original instructions. They also highlight the critical impact of instructions on HCT validity. Given the many limitations of the HCT, we urge researchers to test the robustness of published effects and to reconsider the interpretation of replicable results.