A large randomized trial involving 5,049 adults in the United States and United Kingdom has found that brief mindfulness, self-compassion and gratitude practices can modestly reduce negative feelings, while raising a more complicated question: could people benefit from being matched to the type of meditation that suits them? The research, published online August 10, 2026 and appearing in the October volume of Behaviour Research and Therapy, examined both immediate responses and the possibility of personalized recommendations.
The study was led by Otto Simonsson and colleagues from research institutions including Karolinska Institutet and the University of Wisconsin–Madison. Participants were randomly assigned to five- or ten-minute versions of the three practices or to a comparison condition in which they used the internet as they ordinarily would. The investigators measured immediate changes in subjective experience rather than long-term health outcomes.
Across the trial, the meditation conditions produced greater reductions in negative affect than the comparison activity. They did not show a significant advantage on positive affect. That difference is important: the results suggest a small short-term shift in unpleasant feelings, not a broad improvement in wellbeing or evidence that a few minutes of meditation changes someone’s mental health over time.
The more unusual part of the research involved prediction. The team tested whether information collected before practice could help statistical models identify which intervention might produce a better response for a particular participant. One analytical approach, known as Causal Forest, showed evidence that matching could matter for reductions in negative affect. Another approach, Super Learner, did not produce the same consistent pattern.
Those mixed results are a reason for caution rather than a reason to dismiss personalization. A meditation teacher may already notice that some students respond more readily to breath awareness, while others prefer a compassion or gratitude exercise. The study attempts to move that observation into a testable framework, using information gathered before an intervention instead of relying entirely on broad population averages.
Researchers emphasized that their findings do not establish a ready-to-use algorithm for choosing someone’s meditation practice. The measures were self-reported, the sessions were brief, and the study did not test whether algorithm-guided assignment improves outcomes in a prospective trial. Its participants were recruited from the United States and United Kingdom, which also limits how confidently results can be generalized internationally.
For mindfulness teachers, coaches and developers of digital practices, the professional implication is specific. Personalization may become more evidence-informed, but it is not yet a substitute for listening to a person’s preferences, circumstances and response to practice. The next research step is to test recommendations directly, rather than assume that a promising statistical match already translates into better care.
Source: Simonsson and colleagues, Behaviour Research and Therapy (original research)
Web: https://pmc.ncbi.nlm.nih.gov/articles/PMC13544526/
