A person learns to notice when a worry interrupts reading. After several weeks of mindfulness practice, they finish more pages. Did practice enlarge their attentional capacity, or did it reduce the interference that had consumed the capacity they already had? Both stories could fit a better reading score. Luis Cásedas and colleagues’ 2024 paper asks researchers to distinguish them rather than treating any improvement as proof that meditation builds a larger mental engine.
An integrative framework for the mechanisms underlying mindfulness-induced cognitive change is a Perspective in Nature Reviews Psychology. Its capacity-efficiency mindfulness framework, or CEM, proposes that mindfulness training may improve performance by reducing cognitive-affective interference from mind wandering and negative affect. It integrates prior research and makes predictions for further tests. This guide examines that proposal as a research framework, not as an established explanation for every reported benefit of meditation.
The problem the authors try to solve
Many studies ask whether mindfulness programs improve cognition. Fewer can say why an improvement occurred. A vigilance score may rise after a course, but the course involves instruction, attention to experience, expectations, practice time and sometimes group contact. A person might become less distracted, less worried by errors, more familiar with the task or better rested. If a theory calls all of these “attention training,” it risks explaining too much after the fact.
Cásedas and colleagues draw on cognitive training research, which has long distinguished improvement on a practiced task from transfer to other tasks. Their question is not merely whether mindfulness has a positive average effect. It is which mental operations change and how those changes produce an observed score. A mechanism should generate predictions that can be compared with plausible alternatives. Otherwise a favorable outcome becomes its own explanation.
The authors’ framework is called capacity-efficiency because it separates a system’s underlying resources from how effectively those resources are used during a task. A person may have adequate ability to hold instructions in mind, but worry repeatedly interrupts the work. Reducing that interference could improve performance without increasing total working-memory capacity. That distinction is a hypothesis about process, not something a single before-and-after task score can directly reveal.
What counts as cognitive-affective interference?
Mind wandering can shift attention away from the current task. It is not always a problem; it can support planning, creativity or recovery when the current goal permits it. In a task requiring sustained response, however, spontaneous thought can compete with performance. Negative affect can likewise command attention or change how a person interprets an error. The CEM framework emphasizes these potential sources of interference because mindfulness practice may change a person’s relationship to them.
Consider a student reading a dense passage. The student notices a thought about tomorrow’s exam, follows it into a chain of predictions and loses the thread. A practiced ability to notice the diversion and return might improve reading. But a program could also reduce the student’s exam anxiety, improve scheduling or increase confidence. Those paths may interact. A strong study would measure the proposed intermediate process and test whether its change accounts for the performance difference.
The same observation can support competing stories. Fewer vigilance lapses could mean greater sustained-attention capacity, less intrusive thought, less fatigue or altered motivation during the test. A theory becomes useful when it says what pattern of results would favor its account over the others. The framework’s value lies partly in asking for these more discriminating designs.
Why the evidence base is suggestive, not uniform
Zainal and Newman’s meta-analysis of 111 randomized trials found small to moderate cognitive benefits for mindfulness-based interventions across several measured domains. That broad result makes a mechanism question worth asking. Yet the interventions, comparators, populations and tests varied. An average effect in a synthesis does not identify which component was active in a particular program.
Whitfield and colleagues’ systematic review and meta-analysis of adult mindfulness-based programs found a small overall effect across objective cognitive outcomes, with clearer subgroup effects for executive function and working memory. Across studies, programs outperformed inactive controls but not active comparators. The contrast between reviews is instructive. Different inclusion rules, dates, comparison conditions and cognitive domains can produce different pooled estimates. A framework should explain such variation rather than flatten it into “mindfulness works” or “mindfulness does not work.”
Verhaeghen’s meta-analyses also separate intervention comparisons from cross-sectional differences between long-term meditators and non-meditators, and from correlations involving trait mindfulness. These are not interchangeable forms of evidence. A difference between experienced practitioners and non-practitioners could reflect prior characteristics, practice, other lifestyle factors or a mixture. A randomized intervention offers stronger causal leverage for a specified course, but still may not identify its mechanism.
The primary 2024 paper proposes an integrative explanation in this mixed landscape. It does not report a new randomized trial of the CEM model in the article itself. Calling it a Perspective keeps the level of evidence visible. Its predictions can be fruitful without already being confirmed.
From a framework to a test
One experiment could compare a well-specified mindfulness program with another credible activity matched for teacher contact, time and expectancy. It would assess cognition before and after, measure task-unrelated thought and negative affect during performance, and test whether changes in those measures track the cognitive result. It could include tasks that differ in how vulnerable they are to distraction. This design is an Institute illustration of how one might test the framework, not a study performed by Cásedas and colleagues.
If the CEM account is right in a particular setting, improvement might be largest when interference would otherwise be high, while a low-interference task shows little change. A simple capacity-growth account might predict a broader gain across tasks drawing on the same resource. These predictions are not perfectly exclusive: both mechanisms could operate together. The point is to make the mechanisms measurable and discriminable enough to learn from an experiment.
A study should also avoid treating a self-report item about mind wandering as a transparent window into all distraction. People who practice noticing thoughts may report more wandering because they now detect it, even if their behavior improves. Experience-sampling probes during a task, performance measures and physiological data can add perspectives, but each has limitations. A change in questionnaire score must not be assumed to equal a change in the underlying process.
The design should assess transfer. If reduced interference appears only during a familiar exercise, it may still be valuable to the participant, but it cannot support a broad claim about school performance or daily work. A separate task, real-world outcome and later follow-up are needed for those claims. The CEM paper’s relation to cognitive-training literature makes this boundary especially important.
Researchers could also examine when the proposed mechanism fails. A task that already occupies attention fully may leave little room for reduced wandering to matter. A task with strong emotional cues may be more sensitive to affective interference. A course may change self-reported calm without changing either task. These patterns would help define the framework’s boundary rather than count as embarrassing exceptions to conceal. A mechanistic account earns its place by predicting both where a benefit should appear and where it should be small or absent.
Repeated measurement has costs of its own. Frequent probes asking a participant whether their mind has wandered can interrupt the very process under study. Physiological measures may add information but are not uniquely specific to worry or attention. A credible test can combine methods, report the uncertainties of each and resist treating one convenient marker as a direct meter of efficiency. The strongest evidence would connect a changed intermediate process to improved behavior across more than one carefully chosen task.
Capacity is not a simple tank
The term “capacity” can invite a false picture of a fixed container. Cognitive performance is shaped by strategy, effort, emotion, task features and context. A person may do better because the task becomes more meaningful or because interruptions are removed. Those changes can be real without representing a larger internal reservoir. The CEM distinction is useful if it directs attention to mechanisms, not if it merely replaces one metaphor with another.
Efficiency can also be misused. An employer might hear “reduce interference” and place the burden on workers to meditate through constant notifications. Yet the most direct route to less interference may be fewer interruptions, reasonable workloads or better access to rest. A cognitive theory of practice cannot determine who is responsible for changing the environment. Institutions should not borrow laboratory language to excuse avoidable demands.
Conversely, environmental reform and personal practice need not be rivals. A student may benefit from a quiet room and from learning to notice when a worry takes over. The relative contribution of each support should be tested. The fact that a technique helps one person does not prove that the condition producing the difficulty was fair.
What this paper contributes to contemplative science
The central contribution is a more specific causal story. Instead of promising a generalized upgrade, the authors suggest a route through mind wandering and negative affect that can be assessed during a task. They also connect meditation research to broader cognitive-training questions about mechanisms and transfer. That is valuable because a field can accumulate positive outcomes without learning which component to preserve or adapt.
The paper may also help organize seemingly inconsistent findings. A group with high task-related worry might show improvement if practice reduces interference; a group with little interference may show less change on the same task. That is a plausible interpretation, not a result established across populations by the Perspective. It should motivate preregistered tests rather than retrospective explanations for every outcome.
It is also possible that different practices work through different routes. Focused attention, open monitoring and compassion practice need not share one active ingredient. A model that claims to describe mindfulness broadly should state which instructions and populations it covers. Detailed intervention descriptions are not an administrative nuisance; they are part of testing the theory.
Limits of the source and adjacent research
The publisher makes the abstract, figures and bibliographic material visible, but the full Perspective text requires subscription access in this review. This article therefore relies on the authors’ abstract-level claims for the CEM mechanism and uses accessible reviews for the wider evidence context. It does not attribute detailed predictions or quotations that could not be checked in the full paper. Final publication should include a full-text source audit if access becomes available.
Van Dam and colleagues’ methodological critique of mindfulness research remains relevant. Definitions, comparison conditions, measures, expectations and reporting quality can all alter conclusions. A mechanism framework is only as credible as the studies used to test it. Even an elegant diagram cannot correct a weak measure or missing comparator.
The safety question should remain visible as well. Farias and colleagues’ adverse-event review documents unwanted experiences associated with meditation and large variation in frequency estimates by method. The CEM framework is about cognitive change, not a safety protocol. A program that measures attention while ignoring distress has an incomplete account of its effects. Benefits and costs can occur in different participants or even in the same participant over time.
How to read a future claim
If a future study says a mindfulness course improved executive attention, ask how it defined that domain, what comparison it used and whether the finding survived a later test. If it says reduced mind wandering caused the gain, ask how wandering was measured, whether change preceded the performance change and which alternatives were ruled out. If it says the gain transferred to life, ask for a life outcome measured independently of the practiced task.
These questions do not express hostility to contemplative practice. They respect the difference between a potentially useful exercise and an explanatory claim. The aim is to learn which practice helps whom, for which task, in which context, and at what cost. That information is more useful than a general promise of “better cognition.”
For a person considering practice, the paper is not a prescription. It offers a way to interpret why a particular exercise might change a particular kind of performance. A person may value the experience for reasons unrelated to cognitive testing. Scientific measurement should clarify one question without claiming to own every reason people meditate.
What the evidence does not show
The 2024 Perspective does not prove that mindfulness increases a general pool of mental resources, and it does not prove that reduced mind wandering is the only explanation for cognitive improvements. Its capacity-efficiency account is a proposed mechanism. Evidence from meta-analyses supports studying cognitive effects, but pooled outcomes do not validate the mechanism automatically.
The framework does not show that every meditation tradition is reducible to attention training. It does not establish long-term transfer from a laboratory task to school, work, clinical outcomes or social relationships. It also does not demonstrate safety for all participants. Each stronger claim would require its own design, population and outcome measures.
There is a further conceptual limit: feeling less distracted and performing better are different outcomes. One may matter without the other. A test that recognizes this difference can make both contemplative practice and cognitive science more honest.
Return to the reader
The reader who notices worry and returns to a page may have acquired a useful way to manage interference. They may also be helped by a quieter room, an easier task or more rest. The CEM framework asks which path explains an observed change, rather than awarding all credit to one attractive story.
Cásedas and colleagues provide a research agenda for testing that story. The contribution is not a final verdict that mindfulness makes minds bigger or more efficient in every context. It is a sharper question about how practice, affect and task demands meet in observable performance. That question is worth carrying into the next generation of careful trials.
Research references
- Cásedas, L., Schooler, J. W., Vadillo, M. A., & Lupiáñez, J. (2024). An integrative framework for the mechanisms underlying mindfulness-induced cognitive change. Nature Reviews Psychology, 3, 821–834. View source
- Zainal, N. H., & Newman, M. G. (2024). Mindfulness enhances cognitive functioning: A meta-analysis of 111 randomized controlled trials. Health Psychology Review, 18(2), 369–395. View source
- Whitfield, T., Barnhofer, T., Acabchuk, R., et al. (2022). The effect of mindfulness-based programs on cognitive function in adults: A systematic review and meta-analysis. Neuropsychology Review, 32, 677–702. View source
- Verhaeghen, P. (2021). Mindfulness as attention training: Meta-analyses on the links between attention performance and mindfulness interventions, long-term meditation practice, and trait mindfulness. Mindfulness, 12, 564–581. View source
- Van Dam, N. T., et al. (2018). Mind the hype: A critical evaluation and prescriptive agenda for research on mindfulness and meditation. Perspectives on Psychological Science, 13(1), 36–61. View source
- Farias, M., Maraldi, E., Wallenkampf, K. C., & Lucchetti, G. (2020). Adverse events in meditation practices and meditation-based therapies: A systematic review. Acta Psychiatrica Scandinavica, 142(5), 374–393. View source