Do changes in blood BDNF levels consistently track clinical improvement across brain stimulation trials for depression?

Why BDNF looked like a plausible gauge of recovery

BDNF is a signaling protein that supports neuroplasticity—the brain's ability to remodel connections—and it is found not only in the brain but also in blood [5]. Because impaired plasticity is thought to be part of depression, researchers proposed that measuring BDNF in blood might serve as a window onto both depression severity and treatment response [5]. The underlying logic is intuitive: if a treatment restores plasticity, BDNF should rise, and if BDNF rises, symptoms should ease.

Some observations fit that story. People with untreated major depression tend to have lower blood BDNF than healthy controls [5], and carriers of a common BDNF gene variant (Val66Met) show blunted BDNF expression [1]. Antidepressant treatments increase blood BDNF, more clearly with drugs than with non-drug approaches [5]. These patterns made BDNF an attractive candidate biomarker—a measurable signal that might tell clinicians whether a treatment is working.

Stimulation trials: BDNF rises, but does improvement follow?

The strongest support for a link comes from a meta-analysis of 16 randomized controlled trials (1,271 patients) comparing non-invasive brain stimulation with sham stimulation in major depression [4]. Stimulation significantly raised peripheral BDNF and reduced depressive symptoms, and moderator analyses found that greater BDNF elevation was associated with greater antidepressant effects [4]. The effect was not uniform across techniques: repetitive transcranial magnetic stimulation (rTMS) significantly increased BDNF, whereas transcranial direct current stimulation (tDCS) did not, and rTMS protocols lasting at least four weeks produced larger BDNF increases than shorter ones [4].

Smaller studies complicate the picture. In one trial of 25 drug-resistant depressed patients receiving rTMS, BDNF concentration increased after treatment, but the marker that predicted treatment success was not BDNF—it was a lower initial level of the inflammatory signal TNFα [3]. That same study found no significant BDNF differences between healthy, depressed, and schizophrenic groups before treatment [3], which sits awkwardly with the idea that low BDNF reliably marks depression.

A review of the broader treatment literature reaches a more skeptical conclusion. Across 30 included studies, antidepressant treatments raised blood BDNF, yet neither baseline BDNF nor the change in BDNF correlated with depression severity or treatment outcome [5]. The authors concluded this undermines BDNF's use as a biomarker in depression [5]. An earlier review similarly noted mixed trial results across treatment options and measurement techniques, and stated that blood BDNF cannot be recommended for use as a biomarker in clinical practice [1].

Why the signal is noisy: measurement and biology

Part of the inconsistency is technical. Blood BDNF can be measured in serum or plasma, and these are not equivalent; the earlier review specifically recommends serum, or a serum-to-whole-blood ratio, as a more accurate peripheral measure [1]. The 2024 systematic review likewise stresses factors that affect the accuracy and reproducibility of BDNF measurements and offers methodological recommendations to streamline future assessment [5]. When studies differ in what they measure and how, their results are harder to compare.

Sample size and publication bias add further uncertainty: the earlier review cites heterogeneous results, low sample sizes, and publication bias as challenges to interpreting the relationship between peripheral BDNF and depression [1]. There is also a timing issue. The rapid-acting antidepressant ketamine produces an early increase in blood BDNF [1], suggesting BDNF dynamics may differ by treatment and by when you look—so a single measurement schedule may miss or misrepresent the relationship.

It is also worth being precise about what the stimulation evidence covers. The meta-analysis examined non-invasive techniques (rTMS and tDCS) [4], and the rTMS/ECT trial measured BDNF in depressed patients receiving rTMS [3]. Deep brain stimulation research has instead focused on brain-electrical signals: in ten treatment-resistant patients, recordings from the subcallosal cingulate tracked clinical state and recovery trajectories, with 90% responding and 70% reaching remission at 24 weeks [2]. That work did not measure blood BDNF, so it supports the general idea that objective biomarkers can track depression recovery without telling us whether BDNF is one of them.

So should BDNF change be read as a sign of improvement?

The honest answer is: sometimes on average, not reliably in an individual. At the group level, stimulation trials and meta-analysis show BDNF rising alongside symptom improvement, and the meta-analysis found that larger BDNF increases were associated with larger antidepressant effects [4]. That is a statistical association across trials, not proof that a given patient's BDNF rise causes or predicts their recovery.

At the individual level, the evidence is weaker. The systematic review found no correlation between BDNF change and treatment outcome [5], and the earlier review concluded blood BDNF is not ready for clinical use as a biomarker [1]. A reasonable reading is that BDNF is a useful research signal—especially when measured consistently in serum and tracked over adequate treatment durations [1][4][5]—but not yet a dependable clinical yardstick for judging whether a particular patient is getting better.

Sources used in this answer

Related research questions