Increasing familiarity with the heartbeat counting task does not affect performance

Background: Interoception is typically defined as the processing and perception of internal signals. A common evaluation of interoceptive abilities is the heartbeat counting task, during which participants count their heartbeats over a period of time. It has been argued recently, however, that performance in this task may reflect processes other than cardiac sensation. Objective: This study aimed to: 1) observe heartbeat counting task performance changes across multiple repetitions of the task; and 2) compare performance in the heartbeat counting task with that for a visual counting task to investigate generalised propensities for how uncertain stimuli are reported. Methods: Seventy-nine healthy participants performed seven blocks each of the heartbeat and visual counting tasks. Scores for each of the tasks were compared across blocks to identify any change in performance. Performance in the two tasks was then correlated. The study was preregistered at https://doi.org/10.17605/OSF.IO/GWAB7. Results: Heartbeat counting task performance did not change over blocks. In contrast, scores for the visual counting task reduced over time. A positive correlation was seen between performance in the two tasks (ρ(79) = .27). Conclusion: Heartbeat counting task performance is stable across repetitions of the task, not changing with increasing familiarity. This suggests that non-interoceptive factors, such as beliefs, may be involved. The correlation between the heartbeat and visual counting tasks may point to a general propensity in counting uncertain stimuli across sensory domains. Together, these results raise questions about the interpretation of the heartbeat counting task as a measure of interoception.