Being asked to notice your heartbeat sounds straightforward. Turning that experience into a dependable research measurement is harder. A person might feel individual beats, estimate a familiar resting rate, or confidently report a number without distinguishing the two. Research on interoception—the sensing of signals from inside the body—has to make those possibilities visible.
The instructions can change the result
In a 2018 experiment involving 123 healthy participants, researchers compared the usual heartbeat-counting task with instructions emphasizing that people should report only beats they actually felt. Accuracy scores fell by approximately half under the more restrictive instructions.
Participants described several approaches to the original task, including feeling beats in a particular location, sensing them more diffusely and estimating heart rate. That means a good score could arise through different routes. The experiment did not establish that body sensations are untrustworthy. It showed that a test can blend perception with estimation while giving both answers the same label.
Estimating a rate is a separate ability
A 2026 study of the Heart Rate Matching task makes that distinction explicit. Participants adjusted an animated heart until its pace matched their estimated heart rate. The investigators compared the task with other measurements across three experiments.
In the first laboratory experiment, 41 participants completed the relevant session; the tendency to overestimate or underestimate heart rate did not significantly correlate with two tasks intended to measure perception of individual heartbeats. In a larger online experiment, some aspects of matching responses related to original counting scores, although the size of a person’s estimation error did not. The authors interpret their new measure principally as an assessment of beliefs about heart rate. They also acknowledge that limited statistical power could help explain some differences between experiments.
A repeatable answer is not automatically proof that a test measures the intended sensation. This study received European Research Council funding; its authors declared no conflicts.
A newer task asks about timing
The May 2026 PAT 2.0 methods report describes another approach. People move a virtual dial until tones seem aligned with their heartbeat. Consistent choices across trials provide evidence about perception of timing; knowing an average heart rate is insufficient to solve that timing problem.
The revised task uses detected beats, improves instructions and checks whether someone can perform a comparable task using external signals. That check matters: misunderstanding the task should not be mistaken for difficulty sensing a heartbeat. This is a measurement-development report, including reanalysis of existing data, rather than a trial showing that training improves health.
Support included the Medical Research Council, Oxford and the NIHR. Jennifer Murphy and Geoffrey Bird disclosed paid interoception consultancy for Healios.
Read the task before the headline
When a study reports better body awareness, ask what participants actually did. Counting, estimating, judging timing and rating confidence should not be treated as interchangeable. Then ask whether the claim stays within that task and body signal.
These examples support a modest reading habit: replace a broad label with the concrete measurement underneath it. A cardiac laboratory score can illuminate a specific ability without becoming a verdict on how well someone understands their whole body.