New Vipassana Meditation Research, April 2026

Why a roundup, and why now

The most-cited resource for “what does science say about Vipassana” is a body of work that mostly predates 2025: the Holzel and Britton brain-imaging meta-analyses, the Bowen et al. Vipassana-and-recidivism papers from the early 2010s, the slow-wave sleep work from Sleep and Biological Rhythms, the Tihar Jail reports. The peer-reviewed literature did not stop in 2024. Between August 2025 and April 2026 four papers landed that change at least two things about the public picture: there is now a current PRISMA systematic review you can cite without apologizing for its age, and there is a magnetoencephalography paper with twelve highly experienced monks that distinguishes Vipassana from Samatha at the level of neural-state-space dynamics.

The Giridharan systematic review searched the literature only through April 2025, by its own methods section. So even the most-current review on the public web cannot include the Wankhade observational study from the same journal four months later, the Pascarella MEG study from two months after that, or the Oswal ānāpāna-in-schools paper that came out the day before this page was written. That is the gap a roundup actually fills.

Paper 1 of 4

Wankhade, Lavangare, and Dani: a 152-participant observational mindfulness study

Journal

Cureus, article 383716

Published

August 12, 2025

Sample

152 participants (4 of 156 did not complete)

Design

Pre/post observational, no control group

Instrument

Philadelphia Mindfulness Scale (PHLMS)

Open access

PMC12424088

152 people who attended a 10-day Goenka-tradition Vipassana course completed the Philadelphia Mindfulness Scale before and after. Total mindfulness rose from 61.47 to 81.09 (p < 0.001). Both subscales moved: awareness from 30.76 to 41.11, acceptance from 30.70 to 39.99. The authors report 74.3 percent of participants reaching a high level of mindfulness post-retreat, up from a much lower baseline rate.

The paper is honest about its design. Pre-post with no control group cannot tell you what fraction of the change is the course itself versus regression to the mean, demand characteristics, or social-desirability effects. Self-report scales taken in the immediate afterglow of a 10-day silent retreat are exactly the condition where every one of those biases is most active. The number worth holding is the magnitude of the shift; the explanation of the shift is what later work, with controls, will need to disentangle.

Paper 2 of 4

Giridharan, Soumian, Kumar, and Godbole: a PRISMA systematic review

Journal

Cureus 17(9): e93355

Published

September 27, 2025

Search range

January 2010 to April 2025

Studies included

11 (3 RCTs, 2 single-arm, 1 pilot, 5 observational)

Total participants

792, median 40 per study

Open access

PMC12558566

A PRISMA-registered systematic review across PubMed, Cochrane, Google Scholar, PsycINFO, and Scopus, restricted to studies of the Goenka-tradition 10-day Vipassana course or its direct continuations. Eleven studies met inclusion criteria. The headline numbers are a stress effect size of d = 0.79, mindfulness improvements at d = 0.68 to 0.75, monthly migraine frequency reduced by approximately 2.7 days, and hippocampal topology alterations significant at p = 0.009. Effects were sustained at 6 to 12 month follow-ups in studies that measured that long.

The review explicitly does not pool the studies into a meta-analysis. The authors cite methodological heterogeneity, small sample sizes, and a moderate risk of bias across the body of work as the reasons. Their phrase for the overall picture is “moderate evidence supports the benefits” for stress, anxiety, mindfulness, pain (especially migraine), and selected neurobiological markers, particularly with intensive 10-day-plus retreats. The dose-response pattern they note is intuitive: longer or more intensive retreats produced larger and longer-lasting effects in experienced practitioners.

The April 2025 search cutoff is the practical limit of the review. It cannot include the Wankhade observational study from August 2025 in the same journal, the Pascarella MEG paper from November 2025, or the Oswal Frontiers paper from April 2026. The next systematic review that includes those three will be a meaningfully different object.

Paper 3 of 4

Pascarella et al.: twelve monks, MEG, and the criticality finding

Journal

Neuroscience of Consciousness

Published

November 23, 2025

PubMed

41287816

Sample

12 Thai Forest Tradition monks, 15,000+ hours each

Recruitment site

Santacittarāma monastery, near Rome

Method

Magnetoencephalography (MEG), Chieti-Pescara MEG lab

Conditions

Rest, Samatha (focused), Vipassana (open-monitoring)

Lead authors

Annalisa Pascarella (CNR, Italy), Karim Jerbi (Université de Montréal)

Twelve Buddhist monks from the Santacittarāma monastery outside Rome, each with more than 15,000 hours of meditation experience, were brought to the magnetoencephalography lab at Gabriele d'Annunzio University in Chieti-Pescara. Each monk was scanned at rest and during two meditation conditions: a focused-attention practice (Samatha) and an open-monitoring practice (Vipassana). The analysis pipeline used machine-learning classifiers and three quantitative measures: spectral power across canonical frequency bands, signal complexity (Lempel-Ziv compression), and deviation-from-criticality.

The headline finding is at the criticality layer. Both meditation conditions increased neural-signal complexity relative to rest. Samatha and Vipassana looked statistically distinguishable from each other on these measures. Vipassana was associated with a smaller deviation-from-criticality, which the authors interpret as the brain operating closer to the boundary between order and chaos. Samatha did not produce the same shift; the authors describe it as a more stable, more focused state in their discussion. They are careful to call the criticality finding a state-space description, not a mechanistic claim.

A second finding worth flagging: the paper reports widespread reductions in gamma-band long-range temporal correlations during meditation, which contrasts with several earlier papers that reported gamma increases. The authors attribute the contrast to improved 1/f-slope correction methods that were not standard in earlier work. Whichever direction the next replication goes, this is a useful reminder that “meditation increases gamma” is not as settled as it has sometimes been reported.

Paper 4 of 4

Oswal, Jagtap, and Dasarwar: ānāpāna in two eighth-grade classrooms

Journal

Frontiers in Public Health, Volume 14

Published

April 24, 2026

DOI

10.3389/fpubh.2026.1772248

Design

Quasi-experimental, two intact classes, mixed methods

Sample

89 eighth graders (n = 44 experimental, n = 45 waitlisted control)

Duration

One academic year, three measurement points

The intervention was ānāpāna, the preliminary practice the Goenka tradition offers to children separately from the 10-day adult course. Two eighth-grade classes were assigned to experimental (n = 44) or waitlisted control (n = 45). The experimental group received ānāpāna training and practiced daily over one academic year. Quantitative measures included sustained-attention tasks and psychological well-being scales at three time points; qualitative data came from focus group discussions with students and interviews with parents and teachers.

The headline numbers: Cohen's d = 0.61 for sustained attention by end of program, Cohen's d = 0.43 for self-adjustment at the interim measurement. Parents in the experimental group described better focus on studies, better emotional regulation, and improved relations with peers. Qualitative themes from the focus groups clustered around attention, self-acceptance, and emotional regulation. Both quantitative and qualitative streams pointed in the same direction.

The design is quasi-experimental rather than randomized: the authors used two intact classes and could not randomize at the student level. Parents and teachers reporting on outcomes were not blinded to which class was the experimental one. The paper says all of this directly. As school-based mindfulness research goes, the effect sizes are well above the noise floor for the field, and the mixed-methods triangulation is more stringent than most school-program evaluations attempt.

Reading the four papers together

The four papers do not converge on a single mechanism, and they are not asking the same question. They cover different populations (course graduates, expert monks, eighth graders), different time horizons (immediate post-course, lifetime expert practice, one academic year), different instruments (self-report scales, MEG, school attention tasks), and different practices (full Vipassana, Samatha vs Vipassana, ānāpāna alone). What they share is a willingness to name their own design limits in the abstract or discussion, which is more than most coverage of meditation research does.

The point at which the four start to talk to each other is the attention layer. Wankhade reports a large jump in self-reported mindfulness, with both awareness and acceptance subscales moving. Giridharan reports d = 0.68 to 0.75 for mindfulness across the included studies. Pascarella reports a state-space shift toward criticality that is consistent with theoretical accounts of why attention might feel different during open-monitoring practice. Oswal reports a behaviorally measurable improvement in sustained attention in students who practiced ānāpāna daily for an academic year. Different measurements, partly overlapping construct, mostly consistent direction. Whether the construct itself is one thing or four is the open question the next round of research will have to deal with.

What this roundup does not, and cannot, do

None of these papers tells anyone how to practice. The Goenka tradition reserves technique transmission for an authorized assistant teacher inside a 10-day residential course. That line stands. Pascarella et al. did not measure novices learning the method, they measured monks who had already spent more than 15,000 hours in it. Oswal et al. did not teach ānāpāna from a webpage, they brought trained instructors into a school. Wankhade and Giridharan are evaluating outcomes in people who already took the course. The research is downstream of the transmission, not a substitute for it. For anything operational, the redirect is to dhamma.org and an authorized assistant teacher.

For the longer-form picture of older Vipassana research, the existing scientific evidence guide on this site covers the pre-2025 work in more detail: the slow-wave-sleep findings, the Northwest Rehabilitation Facility recidivism numbers, the Tihar Jail studies, the meta-analyses of eight brain regions consistently altered by meditation. This page is the time-boxed update on top of that, not a replacement for it. The two are meant to be read together.

A peer-practitioner caveat

I write as an old student of six 10-day courses in the Goenka tradition (Dhammamanda in Northern California, CYO in the Bay Area, North Fork in Central California), with about forty days of dhamma service and a daily sit measured in years rather than months. I am not a teacher. I am not a neuroscientist. I read these papers because friends ask me about them, and because the practice has mattered enough in my own life that I want to know what the field is actually claiming and what it is not.

The Pascarella paper is the one I find genuinely surprising. The interpretive frame, that Vipassana pulls neural dynamics toward a chaos-order critical boundary while Samatha settles into a more stable basin, lines up better with the phenomenology of open-monitoring practice than I expected a single small MEG study to manage. It is twelve subjects. It needs replication. It is also the most interesting Vipassana neuroimaging result I have seen since the cortical-thickness papers from the early 2010s, and the public coverage of it (one Medical Xpress write-up, one DNYUZ syndication, one ScienceAlert version) does not mention any of the three other 2025 to 2026 papers it sits next to. That is the gap this roundup tries to close.