Method at a glance
- Modality: respiratory / attentional (behavioural, often with concurrent fMRI and breath-belt physiology)
- Measures: Sustained, self-directed attention to the breath — operationalized through counting accuracy. Participants silently count breaths in repeating cycles and press keys to mark counts, cycle-ends, and self-caught lapses; miscounts and resets index moments of inattention (mind-wandering), giving an objective, meditation-like measure of on-task vs off-task attention without external stimuli.
- Procedure: Participant counts each breath, pressing one button per breath (or per cycle), a second button to mark the end of a fixed cycle (commonly nine breaths), and a third ‘reset’ button when they notice they have lost count. Correct cycles, miscounts, and resets are recorded. When paired with a breath belt, actual breath peaks are logged so that on-task periods (a cycle with the expected number of presses and breaths) can be separated from off-task periods (miscounts, resets) for within-task analysis.
- Also called: BCT, breath counting, breath-counting paradigm
Strengths
- Objective, behaviourally scored proxy for focused-attention meditation that needs no self-report probe to detect a lapse — a miscount or reset is a self-caught mind-wandering event.
- Uses only the endogenous breath as its object, so it isolates internally-directed attention without confounding external stimuli; compatible with fMRI and physiological recording.
- Yields several dissociable outcomes (correct cycles, miscounts, resets) that relate differently to sustained attention and mind-wandering measures (Wong et al. 2018).
Limitations
- Attention is only sampled at cycle terminations, so a lapse anywhere earlier in the cycle is invisible until the count is checked — the temporal resolution of ‘inattention’ is coarse.
- Counting itself, and the button presses marking it, add a working-memory and motor load absent from silent breath meditation; breathing changes during the task can confound BOLD signals unless physiologically controlled (RETROICOR).
- Performance metrics correlate only moderately with trait mindfulness and self-reported attention, and — per Treves et al. (2025) — the neural states accompanying the task need not track within-task attention or individual differences.
Connections
A behavioural paradigm (Levinson et al. 2014) that turns focused-attention meditation into something scoreable. Instead of asking a meditator whether their mind wandered, the BCT has them count breaths in fixed cycles and press keys to mark each count, the end of each cycle, and any moment they notice they have lost count. A miscount or a reset is a mind-wandering event, caught by the participant and time-stamped — so the task supplies an objective index of on-task versus off-task attention using nothing but the breath.
Why the wiki has it
The BCT is the operational bridge between the wiki’s contemplative material (breath-focused body-focused contemplative training) and its attention/mind-wandering material. It is the paradigm behind Treves et al. (2025), the wiki’s first first-hand fMRI study of breath attention, and it recurs across the breathing-attention literature the wiki holds by citation (Harrison et al. 2021; Farb et al. 2023; Wong et al. 2018).
What the outcomes mean
Three behavioural outcomes fall out of the task and they are not redundant: correct cycles, miscounts (the count drifted without being noticed until cycle-end), and resets (the participant noticed the loss of count and acknowledged it). Wong et al. (2018) showed these relate differently to sustained attention and mind-wandering, so a BCT is not a single accuracy score but a small profile of how attention fails.
The physiological add-on
Pairing the task with a breath belt (as in Treves et al.) does two things: it lets RETROICOR regress out breathing-driven BOLD noise, and it lets the analyst define on-task periods (a cycle whose button presses and recorded breath peaks agree) against off-task periods (miscounts, resets) for within-task neural analysis. This is what makes the BCT usable for dynamic connectivity work that asks not just “task vs rest” but “attending vs not, moment to moment.”
The caution the wiki attaches
Treves et al. (2025) is the standing caveat: the BCT robustly separated focused attention from rest at the level of dynamic brain states, but the states that marked the task were not the states that marked within-task attention, and none predicted individual differences in attentiveness. The task is a good behavioural instrument; whether any single neural signature of “attending to the breath” underlies it across people is, on current evidence, unsettled. See Treves et al. (2025).