Why depression research moved to plasticity
The monoamine model struggled to explain the treatment lag: if depression were simply low serotonin, SSRIs would work in hours, not weeks. The plasticity model explains it better. SSRIs raise BDNF over weeks, and the antidepressant effect tracks that timeline more closely than it tracks the serotonin change.
Ketamine sharpened the argument further. It produces rapid antidepressant effects and rapid synaptogenesis, and blocking BDNF signalling in animals abolishes the antidepressant effect entirely. The plasticity is not a side effect — it appears to be the mechanism.
Chronic stress does the reverse: it reduces hippocampal BDNF and produces measurable dendritic atrophy. Depression, on this reading, is partly a disorder of a brain that has lost the ability to reorganise.
The window, and what fills it
Here is the part that gets skipped. Plasticity is permissive, not instructive. Raising BDNF makes circuits more changeable; it does not specify which circuits or in what direction.
Animal work makes this uncomfortably clear. New hippocampal neurons produced by neurogenesis survive only if they are recruited by learning. Unused, they die. The molecular intervention creates the raw material; activity decides what is kept.
There is a clinical version of the same finding. Trials that pair Cerebrolysin with structured rehabilitation consistently outperform trials that administer the compound alone. Same compound, different outcome, depending entirely on what the person did during the window.
This is why we do not treat meditation and rehearsal as a soft add-on. If you open a plasticity window and spend it doomscrolling, you have made yourself more efficient at doomscrolling.
What raises BDNF without a syringe
Aerobic exercise is the most reliably demonstrated BDNF intervention in humans, with acute post-exercise elevations shown repeatedly.
Sleep governs whether the day's plasticity is consolidated. Deep sleep is when synaptic changes are stabilised and pruned.
Intermittent fasting raises BDNF in animal work through ketone signalling, with weaker but suggestive human data.
Learning something genuinely difficult drives BDNF in the circuits doing the learning — which is the most targeted version of all.