The Functional Neuroimaging of Meditation: A Meta-Analysis (Baten et al., 2026)
Meditation is often talked about as if it were a single activity. In practice, resting attention on the breath, watching whatever arises in the mind, silently repeating a phrase and cultivating goodwill toward others ask the mind to do quite different things. Whether those differences show up in the brain has been hard to settle, because earlier meta-analyses either looked at one style on its own or blended several together.
Caitlin Baten, Arielle Keller, Chris Miller and Matthew Sacchet set out to compare styles side by side. Their preregistered review followed the PRISMA reporting guidelines and pooled 34 functional MRI studies with 700 participants. It used a coordinate-based meta-analysis, a method that combines the brain locations reported as active in each study to find where results converge. The studies covered four practices: focused attention, open monitoring, mantra recitation and loving-kindness. Only five loving-kindness studies were available, too few for the quantitative comparison, so that practice was reviewed in narrative form.
Across practices, activity converged in the rolandic operculum, the insula, the superior temporal gyrus, the supplementary motor area and the hippocampus. When the authors looked at the functions typically associated with those regions, the common thread pointed toward self-monitoring, reappraisal, motivation and awareness.
Each practice also had its own profile. Focused attention was the most distinctive, with more activity in the supplementary motor area, insula and superior frontal regions and less in the middle cingulate cortex. Open monitoring involved the insula and operculum along with middle temporal and frontal areas, while mantra recitation stood out for activity in the putamen, a deep structure involved in movement and habit.
The authors are candid about the limits of the evidence. The studies used very different comparison conditions, some contrasting meditation with rest and others with an active task. Several practices were represented by only a handful of studies, the average quality of the studies was low, and most participants were beginners or intermediate practitioners who meditated in short blocks inside a scanner. Reading a mental process from brain activity, known as reverse inference, is also an uncertain step.
Taken together, the findings suggest that meditation styles lean on shared attention-related networks while also recruiting distinct regions. The authors see this as groundwork for future studies on advanced practice, subjective experience and possible targets for brain stimulation. Larger, higher-quality imaging studies of each practice will be needed to firm up the picture.
Reference and link: https://doi.org/10.1016/j.neubiorev.2026.106709
