In a breathing experiment, a participant notices that one inhalation feels harder than the next. The researcher knows that a small resistance was added on some breaths. A correct answer could reflect sensitivity to pressure, a change in breathing effort, a sound from the apparatus or an expectation about the trial sequence. A confident answer could be accurate or confidently mistaken. Measuring respiratory interoception starts by deciding which of these signals and judgments the study intends to capture.
Respiratory interoception concerns perception and interpretation of signals associated with breathing. It may involve airway resistance, chest movement, effort, air hunger or the timing of a breath. It is not one directly observable faculty. A laboratory threshold task, a report of breath awareness and a judgment of confidence each answer different questions. This Methods & Measures guide explains how to align tasks with claims, distinguish sensitivity from bias and metacognition, and avoid treating a score as a diagnosis or as proof of general bodily awareness.
Specify the respiratory signal
“Feeling the breath” is too broad for a measurement protocol. The researcher should specify whether the target is detecting an inspiratory resistance, judging breath phase, tracking spontaneous respiration, noticing effort, or describing a subjective sensation. These experiences recruit different cues. A participant can notice the chest moving without accurately detecting a small imposed load. Another can detect a load while rarely attending to ordinary breathing in daily life.
The signal is also produced by an interaction between body and apparatus. A mouthpiece, tubing or filter can add baseline resistance and change the experience of breathing. A mask may leak or make sounds. The participant’s own breathing depth and speed change pressure and effort. A load that is nominally identical across trials may therefore produce different sensory evidence if the breath changes. Recording respiratory flow, pressure and timing helps interpret what the participant could actually detect.
External cues matter. A researcher changing a filter may make an audible click, alter airflow temperature or move equipment within view. If the participant can use those cues, the task no longer isolates the intended bodily signal. Rieger and colleagues developed remote, automated and MRI-compatible inspiratory resistive loading to improve control over repeated perturbations. Blinding the participant to trial type, hiding apparatus changes and testing for audible or visual differences are practical validity checks. They do not make the signal purely internal, but they narrow the plausible response routes.
Respiratory sensation is often called interoceptive because it informs a person about the body’s state. Yet some cues, such as airflow at the nose or movement of clothing, may also be cutaneous or external. Nikolova and colleagues acknowledge this mixed sensory character while developing a controlled resistance task. The useful claim is not that every correct response arose from one internal receptor; it is that performance under specified respiratory conditions indexes sensitivity to the task’s breathing-related change.
Threshold, decision bias and confidence
A single correct-versus-incorrect score combines several processes. If a person says “load present” on every trial, they may detect all true loads but also produce many false alarms. If they use a cautious criterion, they may miss weak loads while avoiding false alarms. A design with loaded and unloaded trials can estimate sensitivity separately from response bias. Varying resistance across levels allows a psychometric function that relates stimulus strength to detection probability.
The Filter Detection Task developed by Harrison and colleagues was designed to measure breathing-related perceptual sensitivity, decision bias and metacognitive performance while managing a common problem: if trials are far too easy or too hard, confidence comparisons become difficult. Adaptive procedures can keep performance in a range that is informative for estimation. The details of the algorithm, stimulus levels and response options should be reported, because different settings can produce different estimates.
Nikolova and colleagues’ Respiratory Resistance Sensitivity Task uses controlled resistance changes and psychophysical modeling to estimate sensitivity, bias and metacognitive quantities. Its value is methodological separation. Two participants with the same overall accuracy can have different thresholds or decision criteria. Two participants with similar thresholds can differ in how well their confidence tracks their correct and incorrect responses. Those differences should not be collapsed into a generic “good interoception” label.
Metacognitive performance asks whether confidence discriminates one’s own correct from incorrect trials, not merely whether the participant tends to sound confident. A person may be well calibrated on easy loads but uncertain near threshold. Confidence scales also require comprehension and may be influenced by social expectations. A report should state how confidence was elicited, whether it followed each trial, and how task difficulty was controlled before interpreting confidence-accuracy correspondence.
Garfinkel and colleagues distinguished objective task accuracy, self-evaluated interoceptive sensibility and metacognitive awareness in heartbeat work. Their evidence is from a different organ system, so it cannot validate a respiratory load task by itself. It does provide a warning against treating a questionnaire about bodily attention as a substitute for a detection threshold. The respiratory studies then supply the domain-specific methods needed to test that distinction for breathing.
Breath attention is a different task
A resistance-detection task asks whether a controlled change occurred. A breath-following task asks whether someone can track an ongoing pattern of respiration. Daubenmier and colleagues developed a task intended to assess sustained attention to respiratory sensations and compared experienced meditators with nonmeditators. The design is informative because attending to naturally occurring breaths differs from noticing a discrete external manipulation. Results from one should not be renamed as results from the other.
Following the breath without changing it is especially difficult to test. Paying attention can alter respiratory rate, depth or pauses. An instruction to “breathe normally” may itself increase self-monitoring. If a participant actively controls inhalation to match a task, a good score could reflect motor control rather than passive sensory tracking. A measurement design should record respiratory behavior and identify whether changes during the task are part of the construct or a confound.
Training history matters. A person who has practiced meditation, breathwork, music or endurance sport may know more language for describing breathing and may use strategies in a task. That does not guarantee greater sensory accuracy. Conversely, a person unfamiliar with introspective language may detect loads accurately but describe the sensation vaguely. Researchers should separate task performance from familiarity with descriptions and avoid assuming that a practice improves every facet of respiratory interoception.
General body-awareness questionnaires can characterize how people report noticing, regulating and trusting bodily sensations. The Multidimensional Assessment of Interoceptive Awareness, developed by Mehling and colleagues, is a self-report instrument covering several such dimensions. It was not built as a direct respiratory threshold test. A high score can coexist with mediocre load detection; a low score need not imply poor sensitivity to a specific resistance. Both measures may be valuable when their different targets are named.
Respiratory physiology belongs in the analysis
Breathing is not a passive signal source. Participants can change rate, depth and route between nose and mouth, and those changes can affect task difficulty. A load presented during a shallow inhalation may feel different from one during a deep inhalation. The phase of a breath and the interval since the previous load may also affect expectation and sensation. A protocol should standardize what can be standardized and record what cannot.
Baseline airway resistance varies among individuals and across sessions. A participant with congestion or an altered breathing pattern may face a different effective stimulus even when the laboratory load is identical. The equipment itself has resistance. Thus an absolute load value is not automatically a comparable sensory challenge across people. Individualized threshold estimation or calibrated relative changes can help, but the exact procedure must be reported for reproducibility.
Accuracy estimates depend on the reference standard. In a physical detection task, the investigator knows whether the resistance device was engaged, but must still verify that it functioned as intended on each trial. In a subjective sensation task, there may be no single externally correct label for “air hunger” or “ease.” The measurement problem changes. One can estimate agreement with a standardized stimulus or consistency of a person’s report, but those are not identical forms of validity.
Symptoms and distress add interpretive complexity. A person may be highly attentive to breathing because it is uncomfortable or frightening. More attention does not necessarily mean more accurate discrimination of subtle changes, and lower accuracy does not mean the sensation is unreal. Studies involving respiratory illness or anxiety should use appropriate clinical oversight and avoid turning an experimental perceptual score into a diagnostic verdict. The present page is about research measurement, not self-assessment or treatment.
Design a test of the intended construct
First state the claim at the level of observation. “Participants detected added inspiratory resistance at lower levels” is more precise than “participants became more interoceptive.” If confidence is the focus, state whether the outcome is average confidence, calibration or metacognitive discrimination. If breath attention is the focus, define what counts as successful tracking and how deliberate breath control is identified.
Second use appropriate comparison trials. Include unloaded breaths, vary stimulus level and consider sham changes in the apparatus that preserve possible sounds and movements without changing resistance. Randomize order so participants cannot infer the answer from a simple pattern. Check whether the participant noticed unintended cues. Report exclusions, missed responses and whether practice trials altered the strategy used during the main task.
Third inspect reliability. A threshold estimated from too few informative trials can be noisy. Repeated sessions can test whether a person-level rank persists, while remembering that respiratory state may genuinely vary. A measure may be sensitive to an experimental manipulation but unreliable for ranking individuals. Confidence and threshold estimates have separate uncertainty. Presenting both without intervals can make small differences look more meaningful than they are.
Fourth check convergent and discriminant evidence. A load threshold might relate to a second resistance task, but should not be assumed to correlate strongly with a broad questionnaire. Breath-following performance may align with sustained attention more than with load sensitivity. Measuring attention, respiratory physiology and task comprehension alongside the target outcome helps explain convergence or divergence. A weak correlation among tasks may indicate different respiratory capacities rather than a defective instrument.
Fifth specify the population. A protocol studied in healthy adults cannot be assumed to have the same difficulty or interpretation in children, people with lung disease, or participants using different breathing aids. Apparatus comfort, safety and accessible instructions become more important in such settings. Translation of confidence scales and sensation words also requires checking meaning, not just replacing vocabulary.
Finally, predefine how performance will be summarized. Report threshold and slope of the psychometric function if estimated, response bias, accuracy at specific levels, and confidence measures separately. State whether the analysis is within-person change or between-person comparison. Do not select whichever index gives the most favorable result after the data are seen. A transparent analysis lets later researchers know which aspect of breathing perception changed.
What the evidence does not show
The cited studies do not provide one universal meter for “breath awareness.” Resistance sensitivity, breath following, self-reported attention and confidence are related but distinct observations. A person who detects small loads is not thereby shown to interpret every respiratory signal accurately in daily life. A person who reports frequent breath awareness is not thereby shown to have a low detection threshold.
Nor do laboratory scores establish psychological virtue, clinical resilience or the value of a contemplative practice. The Daubenmier study tested specific tasks and groups; it does not show that every experienced practitioner has superior respiratory perception. The Garfinkel study’s distinctions arose from heartbeat measures and should be transferred to respiratory research as questions to test, not as automatic equivalences.
No trial can fully eliminate changes in attention, strategy and breathing behavior caused by being measured. Psychophysical methods make some uncertainties explicit and reduce some confounds, but they do not reveal a private sensation without interpretation. A careful conclusion names the stimulus, equipment, response, sample and uncertainty, then stops where the evidence stops.
Read the result at the right level
Return to the participant who noticed a harder inhalation. The strongest conclusion may be that they discriminated a calibrated resistance from sham breaths under controlled conditions, with a measured threshold and a particular confidence pattern. That is a useful result. It is also narrower than a claim about their general relationship to the body. The related Learning Hub entries preserve that wider context while this guide keeps the methods distinct.
Good measurement makes the research question smaller and the answer stronger. It asks which respiratory change was sensed, by which route, with what decision rule and how reliably. It lets a null result remain a null result rather than declaring a person disconnected from their breath. That care matters because breathing is simultaneously a physiological process, a sensation, a possible object of attention and an action people can change while being studied.
Research references
- Harrison, O. K., et al. (2021). The Filter Detection Task for measurement of breathing-related interoception and metacognition. Biological Psychology, 165, 108185. View source
- Nikolova, N., et al. (2022). The respiratory resistance sensitivity task: An automated method for quantifying respiratory interoception and metacognition. Biological Psychology, 170, 108325. View source
- Daubenmier, J., Sze, J., Kerr, C. E., Kemeny, M. E., & Mehling, W. (2013). Follow your breath: Respiratory interoceptive accuracy in experienced meditators. Psychophysiology, 50, 777–789. View source
- Garfinkel, S. N., Seth, A. K., Barrett, A. B., Suzuki, K., & Critchley, H. D. (2015). Knowing your own heart: Distinguishing interoceptive accuracy from interoceptive awareness. Biological Psychology. View source
- Mehling, W. E., et al. (2012). The Multidimensional Assessment of Interoceptive Awareness (MAIA). PLOS ONE, 7(11), e48230. View source
- Rieger, S. W., Stephan, K. E., & Harrison, O. K. (2020). Remote, automated, and MRI-compatible administration of interoceptive inspiratory resistive loading. Frontiers in Human Neuroscience, 14, 161. View source