Meditation, BDNF, and the Hippocampus: What One Trial Found

Meditation may affect a biological pathway that connects chronic stress to brain plasticity. In a 2025 randomized trial, nine months of structured contemplative training increased serum BDNF in healthy adults; reduced hair glucocorticoid concentration accounted for 15.1% of that increase, and participants also showed a concurrent increase in dentate gyrus volume. That is unusually specific evidence for contemplative science. It is also one unblinded trial with a passive control group, several measurement questions, and no independent replication.

The study does not show that a meditation app will grow your hippocampus, nor that BDNF changes translate into better memory, less depression, or a longer life. It does offer a testable hypothesis: sustained mental training may reduce some measures of cortisol output, and that change may explain part of a rise in a neurotrophin linked to hippocampal plasticity.

A pathway with a number attached

BDNF, or brain-derived neurotrophic factor, is involved in neuronal survival and plasticity. Cortisol is a stress hormone with several distinct measurement methods, each reflecting somewhat different aspects of hypothalamic-pituitary-adrenal activity. The appeal of the new finding is not simply that meditation changed two biomarkers. The researchers tested whether change in cortisol statistically mediated change in BDNF.

In the 2025 Puhlmann et al. randomized clinical trial, 332 meditation-naive, healthy German adults, average age 40.5 years, completed a progressive ReSource training program. The program moved through three sequential three-month modules targeting attention-based mindfulness, socio-affective skills, and socio-cognitive skills. A passive retest group included 90 participants. Researchers collected data at four time points over nine months.

Serum BDNF increased significantly and cumulatively after three, six, and nine months in the combined training cohorts. The trial also found a concurrent increase in dentate gyrus volume, a hippocampal subregion involved in learning and memory. Reduced hair glucocorticoid concentration, or HCC, statistically mediated 15.1% of the training-related BDNF increase.

Fifteen percent is meaningful because it converts a plausible story into a measured contribution. It is not a complete explanation. About 85% of the BDNF change fell outside that HCC-mediated path. The study cannot apportion that remainder among other possibilities, such as changes in sleep, physical activity, diet, or the social effects of long-term partner exercises.

An earlier report from the same ReSource trial established the upstream result: contemplative training reduced hair glucocorticoid levels in healthy adults. Puhlmann et al. in 2021 also makes clear that this was an open-label study with a passive retest control, a design choice that shapes how much causal certainty the later mediation analysis can carry.

Why the result is interesting, but not settled

Randomization helps, but it does not make every downstream explanation disappear. Participants knew whether they were devoting nine months to a demanding training program. Expectation, group belonging, and the behavior of motivated participants can affect sleep, alcohol use, exercise, and what happens before a blood draw. A passive control group cannot separate those effects from the specific content of contemplative practice.

The mediation model has another constraint. The trial assessed stress-induced cortisol only once per participant, at one of three time points in separate participant subsets, because repeating the stress induction could alter later responses. That means this part of the proposed pathway was estimated across different people at different times rather than repeatedly within the same people. Statistical mediation supports consistency with a pathway; it does not prove that cortisol reduction caused the BDNF increase.

Measurement adds further uncertainty. HCC is widely used as a marker of longer-term cortisol exposure, but results vary substantially across samples and measurement methods. Bürgin et al. describe considerable heterogeneity in hair-cortisol findings, including differences between immunoassays and liquid chromatography-tandem mass spectrometry. The 2025 meditation report did not provide assay-specific coefficients of variation for its HCC measurement, so readers cannot judge how its group-level change relates to the precision of that specific assay.

Different cortisol measures also do not necessarily report one unified biological state. A systematic longitudinal comparison by Singh Solorzano et al. found imperfect correspondence between saliva and hair cortisol over time. The meditation trial examined HCC, cortisol awakening response, and acute stress-induced cortisol. Those measures may capture partially distinct processes rather than three interchangeable readings of chronic stress suppression.

Serum BDNF has its own interpretive limits. It is a peripheral biomarker, not a direct reading of brain BDNF activity, and its pre-analytical handling matters. The meditation study described a correlation between peripheral and central BDNF expression, but a single serum measurement cannot establish that the observed change occurred in the dentate gyrus or caused the MRI finding. The concurrent structural and biomarker results are compatible with a shared process; they do not establish a complete biological chain.

What other research can and cannot add

A 2025 cross-sectional study by Okur et al. found relationships among metacognitive beliefs, salivary cortisol, BDNF, and NDNF. That supports the idea that cortisol and neurotrophin measures can covary in people living ordinary lives. Because it was cross-sectional and used saliva rather than serum, it cannot establish direction of effect or independently confirm the meditation trial's mechanism.

Exercise research provides useful context, though it cannot be relabeled as mindfulness evidence. A 2025 systematic review and pairwise, network, and dose-response meta-analysis found type- and dose-dependent BDNF elevation from exercise in people with depression. The Yuping et al. review describes several concurrent neurobiological mechanisms, not a single cortisol route. Those findings apply to exercise in a clinical population, not meditation in healthy middle-aged adults.

Other exercise trials show why both plausibility and restraint are warranted. In older adults at varying risk of mild cognitive impairment, a 2024 resistance-exercise randomized trial measured changes in hippocampal subfields alongside BDNF-related biomarkers. In patients with mild-to-moderate Alzheimer's disease, exercise increased proBDNF, BDNF, and humanin in neuron-derived extracellular vesicles in an analysis of 95 trial participants; that signal has a closer connection to neurons than ordinary serum measurements. Delgado-Peraza et al. therefore adds biological context, not confirmation of meditation's effect.

Behavioral interventions do not reliably produce MRI-detectable hippocampal enlargement. In a six-month randomized trial of aerobic exercise among breast cancer patients, Koevoets et al. found no significant increase in hippocampal or global gray-matter volume. Differences in population, intervention, duration, and MRI endpoint prevent a direct head-to-head comparison with the ReSource study. Still, the null result makes the single dentate gyrus finding a replication target, not a settled fact.

What to do with this if you care about BDNF

The trial tested a specific intervention: nine months of structured, progressive, group-based contemplative training. It did not test a 48-hour retreat, occasional unguided meditation, or a few minutes between meetings. No evidence here supports claims that shorter or less structured formats produce the same biomarker or brain-structure results.

If chronic stress is a real problem for you, sustained contemplative practice may be a reasonable tool for stress regulation. This trial gives a biologically coherent reason to study that approach further. It does not identify baseline cortisol levels at which the program works best, prove a clinical benefit, or show that the effect extends to depression, bipolar disorder, Parkinson's disease, or other conditions. BDNF is state-sensitive in psychiatric illness and can change with treatment, which makes simple mechanistic claims especially risky in clinical populations.

For someone choosing a behavioral strategy specifically to raise BDNF, exercise has the broader and more replicated evidence base, including dose- and type-dependent findings in depression. Meditation can complement exercise for stress management, attention, or emotional skills. The 2025 trial makes it scientifically interesting as a possible contributor to BDNF regulation. Its 15.1% mediation estimate is a precise starting point, not a finished mechanism.

Verdict: Nine months of structured contemplative training raised serum BDNF and coincided with lower HCC and larger dentate gyrus volume in one randomized trial of healthy adults. The cortisol pathway explained only part of the BDNF change, the control was passive and unblinded, and the structural result needs independent replication.