Wait — ketamine blocks NMDA, but the effect runs through AMPA?

Yes. The working model is that ketamine preferentially blocks NMDA receptors on inhibitory interneurons, which removes a brake on glutamate release. The resulting glutamate surge then activates AMPA receptors on pyramidal neurons. AMPA throughput is the proximate driver of the antidepressant effect; NMDA blockade is the upstream event that lets it happen.

Are there AMPA-only antidepressants in development?

Several AMPA-receptor potentiators have been studied in depression, with mixed clinical results. The mechanism is interesting and the rationale is sound, but clinical translation has been harder than it looked on paper. Ketamine remains the best-studied rapid-acting antidepressant that engages this system, and Spravato (esketamine) is the only FDA-approved member of the family.

Does Spravato work the same way?

Yes. Esketamine is the S-isomer of racemic ketamine and acts on the same NMDA and AMPA receptor system. The mechanism described here, including the glutamate surge, AMPA activation, BDNF release, and mTOR-driven synaptogenesis, applies to both racemic IV ketamine and intranasal esketamine.

Why does this matter to me as a patient?

Most of this is mechanism, not bedside-relevant. But it does explain something patients notice directly: response can show up within hours rather than the weeks required for traditional antidepressants. The synaptic switch flips fast because AMPA throughput and the downstream plasticity cascade are quick events at the cellular level.

Is hydroxynorketamine the real active molecule?

Honest answer: this remains contested. Some animal work suggested that hydroxynorketamine, a metabolite of ketamine, produces antidepressant-like effects on its own through AMPA activation, without NMDA blockade. Subsequent studies have been more nuanced, and replication has been mixed. We do not yet know the full answer, and we do not oversell the hydroxynorketamine story to patients.

Ketamine and AMPA Receptors: The Proximate Switch Behind the Antidepressant Effect

Glutamate, NMDA, and AMPA in 90 seconds

Glutamate is the brain’s main excitatory neurotransmitter. When one neuron wants to tell another to fire, it almost always uses glutamate to do it. The receiving neuron has several different receptor types waiting to catch that glutamate, and the two most relevant for understanding ketamine are NMDA receptors and AMPA receptors. They sit side by side in the same synapses, but they do different jobs.

AMPA receptors are the fast lane. When glutamate hits an AMPA receptor, sodium rushes in within milliseconds and the receiving neuron depolarizes. AMPA does the moment-to-moment work of carrying signals from one neuron to the next. NMDA receptors are slower and more conditional. They only open fully when the cell is already depolarized and glutamate is present, which makes them coincidence detectors that gate longer-term changes in synaptic strength.

For decades, ketamine has been described as an NMDA-receptor antagonist. That description is correct. But it has also led to a tempting shortcut: if ketamine blocks NMDA, then NMDA blockade must be what produces the antidepressant effect. The data has not been kind to that shortcut. The cascade after the NMDA block is where the action is, and AMPA is the proximate switch.

Why blocking NMDA does not directly cause the antidepressant effect

Several other NMDA-receptor antagonists exist. Memantine, used for Alzheimer’s disease, is an NMDA antagonist. So is dextromethorphan in cough medicine. Higher-affinity research compounds have been tested in depression. None of them produce the rapid, robust antidepressant response that ketamine does. If NMDA blockade by itself were sufficient, those drugs should work as antidepressants. They largely do not.

That paradox is what pushed researchers to look downstream. The current working model, supported by more than 15 years of converging evidence, is that ketamine’s NMDA blockade is preferential for inhibitory interneurons. Those interneurons normally hold pyramidal neurons in check by releasing GABA. Take the brakes off the brakes, and pyramidal neurons release a brief surge of glutamate. That glutamate has to go somewhere. It goes to AMPA receptors.

If you would like the broader tour of how ketamine works at a system level before diving deeper, our overview at how does ketamine work covers the same territory in plainer language. The page at how it works is the patient-facing version.

Maeng 2008 and the NBQX experiment that fingered AMPA

The cleanest single experiment pointing at AMPA came from Maeng and colleagues, published in Biological Psychiatry in 2008. The team used the standard rodent depression-model paradigms, including the forced-swim test, where ketamine produces a reliable antidepressant-like behavioral signature. Then they pretreated the animals with NBQX, a selective AMPA-receptor antagonist. NBQX itself did not change baseline behavior much. But when NBQX was on board, ketamine’s antidepressant-like effect disappeared.

This is the kind of pharmacological logic that makes mechanism people happy. If blocking AMPA blocks ketamine’s effect, then AMPA throughput is required for the effect. NMDA blockade alone is not sufficient, because the NMDA block is still happening when NBQX is on board. The chain has to go through AMPA. Maeng 2008 is the paper most people cite as the moment AMPA went from interesting hypothesis to load-bearing piece of the model.

Li 2010, mTOR, and the connection to spine growth

The next year, Li and colleagues published a landmark study in Science in 2010 that connected the AMPA story to the structural changes ketamine produces in the prefrontal cortex. Using rats, the team showed that a single dose of ketamine rapidly activated the mTOR signaling pathway, increased synaptic protein levels, and grew new dendritic spines on prefrontal pyramidal neurons within hours.

The pharmacological controls in that paper are what make it persuasive. Blocking mTOR prevented both the synaptic protein changes and the antidepressant-like behavior. And, critically, blocking AMPA upstream prevented mTOR activation in the first place. The order of operations was clear: AMPA throughput drives mTOR signaling, mTOR signaling drives the synthesis of new synaptic machinery, and the new synaptic machinery shows up as visible new spines on dendrites. The behavior tracks the spines.

Autry and colleagues, publishing in Nature in 2011, filled in another required step. They showed that ketamine’s antidepressant-like effects depend on rapid release of brain-derived neurotrophic factor (BDNF) and signaling through its receptor, TrkB. Mice engineered to lack BDNF in the relevant circuits did not respond to ketamine. The BDNF release sits between AMPA activation and mTOR—another link in the chain that you cannot remove without breaking the whole thing.

Where Krystal and colleagues land in the current synthesis

By 2023, the field had enough converging data to consolidate a working model. Krystal, Kavalali, and Monteggia’s review in Neuropsychopharmacology laid out the synthesis explicitly: ketamine causes a glutamate surge, AMPA carries that surge, AMPA activation triggers BDNF release, BDNF activates TrkB and mTOR, and mTOR drives the synaptogenesis that correlates with sustained antidepressant effect. AMPA is the gas pedal. NMDA blockade just takes the foot off the brake.

This synthesis is consistent with what clinicians have been seeing for two decades at the bedside. The antidepressant response often appears within hours, far faster than monoamine-based antidepressants. That timeline only makes sense if the proximate driver is a fast event at the synaptic level, not a slow change in receptor density or gene expression. AMPA throughput plus same-day plasticity fits the timeline. Slow changes do not.

For readers who want more on the plasticity side, our piece on the neuroplastic window covers the period after an infusion when the new synaptic capacity is most useful, and our ketamine brain research overview tracks the broader literature.

What this means for next-generation antidepressants

If AMPA throughput is the proximate driver, then in principle a drug that boosts AMPA without touching NMDA at all might produce a rapid antidepressant effect. AMPA-receptor positive allosteric modulators, sometimes called AMPAkines, have been studied in depression. The clinical results have been mixed. Some compounds showed signal in early trials and then failed to confirm in larger ones. Translating mechanism into a clean clinical drug has been harder than the model suggests it should be.

Hydroxynorketamine, a metabolite of ketamine, has been proposed as a possible AMPA-driven active molecule that could deliver the benefit without the dissociative experience. The original animal work was provocative. Subsequent replication has been mixed. We do not yet know whether hydroxynorketamine alone is sufficient, and we do not oversell the story to patients.

For now, racemic IV ketamine and intranasal esketamine remain the practical clinical tools. Esketamine (Spravato) is FDA-approved for treatment-resistant depression. Racemic ketamine itself is FDA-approved as an anesthetic; its use for depression and other psychiatric and pain conditions is off-label. The mechanism described here applies to both, because they engage the same NMDA and AMPA system.

What we tell patients without overselling

Most patients do not need to know the receptor pharmacology to benefit from treatment. But many ask, and a clear, honest version of the mechanism is more useful than the marketing-grade version. Here is what we say.

Ketamine engages a glutamate-based plasticity system that other antidepressants do not. The proximate event is increased AMPA-receptor throughput in cortical circuits, triggered by an upstream NMDA blockade that disinhibits glutamate release. That AMPA activation drives BDNF release and mTOR-mediated synaptogenesis, and the new synaptic capacity is what plausibly underlies the rapid mood improvement many patients report. Research suggests this is a real mechanism, not a marketing story.

We hedge appropriately. Studies indicate that roughly 50 to 70 percent of patients with treatment-resistant depression respond to a series of infusions, depending on the trial and the response criterion. That leaves a meaningful minority who do not respond. Mechanism does not guarantee response in any individual. We talk through that honestly during consultation. If you would like the depression-specific overview, see ketamine for treatment-resistant depression or our top-level depression page.

At Music City Ketamine, every infusion is delivered with anesthesia-level monitoring. Marla Peterson, CRNA, oversees every infusion and is on-site throughout the session, with continuous pulse oximetry, blood pressure, and heart rate tracking. Sessions are $475 each, and we are transparent about cost from the first conversation. Insurance generally does not cover ketamine for depression, because the use is off-label. Spravato, which is FDA-approved, has different coverage rules and we can discuss those if relevant.

The mechanism is interesting. The clinical question, though, is always the same: is this likely to help you? That answer depends on your history, your prior treatments, your current medications, and a careful conversation. Mechanism gets you in the door. The clinical fit is what determines whether to walk through it.