If glutamate goes up, isn't that excitotoxic?

Excitotoxicity is a chronic, sustained-glutamate problem — the kind seen in stroke or prolonged seizure activity. The transient, localized glutamate surge produced by a sub-anesthetic ketamine dose is a fundamentally different regime. The increase resolves within hours, and the downstream effect researchers care about is what happens after: AMPA-receptor-driven synaptic strengthening, not neuronal damage.

Why not just give a GABA drug like a benzodiazepine?

Benzodiazepines boost GABA broadly across the brain, which is why they sedate, impair memory, and carry dependence risk. The mechanism researchers describe with ketamine is targeted disinhibition of specific GABAergic interneurons in the prefrontal cortex, not blanket GABA enhancement. The two drugs act on opposite ends of the loop, which is one reason chronic benzodiazepine use can blunt ketamine response in some patients.

Does this glutamate-GABA story apply to anxiety too?

There are parallels. Excitatory-inhibitory imbalance has been implicated in anxiety disorders, and glutamate signaling in the prefrontal cortex and amygdala is an active area of research. That said, the evidence base for ketamine in anxiety is thinner than the evidence base in depression. We do not overclaim. Ketamine for anxiety is off-label and should be discussed individually with a clinician.

Are SSRIs working on the same system indirectly?

Possibly. Long-term SSRI treatment does appear to influence glutamate and GABA tone over time, which may be part of why their full effect takes weeks to emerge. The kinetics are slow because the change happens through downstream signaling rather than direct receptor blockade. This is one reason researchers think ketamine's rapid action reflects a different point of entry into the same circuit.

How is this different from the older 'low serotonin' framing?

The serotonin hypothesis was always a working approximation built around how SSRIs were observed to help. The current model in academic psychiatry is closer to this: chronic stress disrupts the balance between glutamate and GABA, weakens synaptic connections in the prefrontal cortex, and ketamine recalibrates that balance and restores synaptic density. It's a circuit-level story rather than a single-neurotransmitter one.

Ketamine, Glutamate, and GABA: The Excitatory-Inhibitory Balance Behind the Effect

Why “low serotonin” was always an oversimplification

For decades, the public version of the depression story has been simple: depression is a chemical imbalance, specifically a shortage of serotonin, and the job of an antidepressant is to top it back up. This framing helped justify the rise of SSRIs in the 1990s, and it stuck because it was easy to explain. It was never quite what the evidence showed.

The serotonin hypothesis was built backward, working from the observation that drugs which raise serotonin sometimes help depression. That is a valid clue, but it is not a mechanism. Direct measurements of serotonin levels in depressed patients have not consistently shown a deficiency. Studies that deliberately deplete serotonin in healthy people do not reliably produce depression. And SSRIs, which raise serotonin within hours, take four to eight weeks to produce mood benefit—a delay that does not match a simple top-up model.

Over the last twenty years, academic psychiatry has moved on. The newer model is a circuit-level story about excitatory-inhibitory balance, synaptic strength, and neuroplasticity in the prefrontal cortex. Glutamate and GABA sit at the center of that story, and ketamine is the drug that made the field take it seriously.

Glutamate as the brain’s main excitatory signal

Roughly 80% of the synapses in your cortex are glutamatergic. Glutamate is the workhorse excitatory neurotransmitter—the “go” signal that lets one neuron tell the next one to fire. It binds to several receptor types, but the two that matter for this discussion are NMDA receptors and AMPA receptors. NMDA receptors act as a slower, voltage-sensitive gate. AMPA receptors handle the fast, moment-to-moment transmission of signals between cells.

When researchers talk about a synapse being “strong,” they often mean it has a high density of AMPA receptors and a robust ability to fire. When a synapse is “weak,” it has fewer AMPA receptors and a smaller dendritic spine. Chronic stress, in animal models and in human postmortem tissue, is associated with synaptic weakening in the prefrontal cortex—fewer spines, fewer connections, less throughput. Ketamine’s antidepressant effect is increasingly understood as a reversal of that weakening.

GABA, interneurons, and the inhibitory side of the balance

Glutamate cannot be the whole story, because a brain made of pure excitation would seize. The counterweight is GABA, the main inhibitory neurotransmitter. GABA is released primarily by interneurons—small, locally projecting cells whose job is to dampen, time, and shape the firing of nearby pyramidal neurons.

One class of these interneurons, the parvalbumin-expressing interneurons, is particularly important for the ketamine story. They tonically inhibit pyramidal neurons and they happen to express NMDA receptors that are unusually sensitive to ketamine’s blockade. That detail—the differential sensitivity of NMDA receptors on interneurons versus pyramidal neurons—is what allows ketamine to do something counterintuitive: a drug that blocks excitation produces, on net, a temporary increase in excitation downstream.

What Sanacora and Krystal’s groups actually proposed

The intellectual scaffolding for the glutamate model of depression was laid down by a small group of researchers at Yale and elsewhere over the last two decades. A landmark paper by Sanacora, Treccani, and Popoli in Neuropharmacology (2012) argued for what they called the glutamate/neuroplasticity hypothesis: chronic stress disrupts glutamate-GABA balance, weakens synaptic structure in prefrontal regions, and rapid-acting antidepressants reverse those changes. They explicitly framed this as an alternative to monoamine-only models, while emphasizing it was not yet a finished story.

A 2016 synthesis in Nature Medicine by Duman, Aghajanian, Sanacora, and Krystal extended that argument with the synaptic-plasticity framing. Stress, in their model, drives glutamate-mediated dendritic remodeling and synaptic loss; rapid-acting antidepressants like ketamine reverse that loss through a burst of activity-dependent plasticity. Studies indicate this is why a single sub-anesthetic infusion can produce mood change within hours rather than weeks. The drug is not adjusting a chemical level—it is triggering a structural shift.

More recently, Krystal, Kavalali, and Monteggia reviewed the mechanistic details in Neuropsychopharmacology (2023), focusing on what is now called the disinhibition model. Their review pulled together the molecular, cellular, and circuit-level evidence and argued the field has converged on a coherent account of how ketamine produces its effects—even if individual details remain contested.

The disinhibition model: NMDA on interneurons explains a lot

Here is the disinhibition story in plain language. At sub-anesthetic doses, ketamine preferentially blocks NMDA receptors on GABAergic interneurons. Those interneurons stop firing as much. Because their job was to inhibit nearby pyramidal neurons, the pyramidal neurons are released—disinhibited—and fire more. That extra firing produces a brief, localized glutamate surge.

That surge then activates AMPA receptors on the pyramidal neurons. AMPA activation drives signaling cascades, including BDNF release and mTOR pathway activation, that promote the formation of new dendritic spines and the strengthening of existing synapses. The visible result, hours later, is a brain whose prefrontal circuits are passing more signal more efficiently than they were before the infusion.

Several things follow from this model. First, the “antidepressant” effect of ketamine is not really a chemical effect at all—it is a structural one, mediated by transient excitation. Second, drugs that block AMPA receptors during ketamine administration block the antidepressant effect, which is exactly what you would predict. Third, the importance of drug interactions with this circuit becomes obvious. Benzodiazepines, which broadly enhance GABA tone, push the system in the opposite direction from where ketamine is trying to push it. Some clinical reports suggest chronic benzodiazepine use can attenuate ketamine response, though the evidence is mixed and individualized clinical judgment matters.

Human MR-spectroscopy evidence for the shift

The disinhibition model started in rodents. The question that mattered for clinical psychiatry was whether anything similar shows up in humans. Magnetic resonance spectroscopy, which can non-invasively estimate glutamate and GABA concentrations in specific brain regions, became the field’s tool of choice.

Reviews of glutamate and GABA imaging in mood disorders, including the broader 2012 *Neuropharmacology* synthesis by Sanacora, Treccani, and Popoli, summarize the human evidence. Across multiple studies, depressed patients showed lower GABA and disturbed glutamate signaling in cortical regions, particularly the prefrontal cortex. After ketamine, prefrontal glutamate measurements shift in a pattern consistent with the rodent disinhibition findings. The human and animal stories are not identical—they never are—but they line up well enough that most reviewers consider the basic mechanism well supported.

For curious readers, our overview of the broader ketamine brain research goes deeper into the imaging and connectivity work. The short version is that ketamine produces measurable, reproducible changes in glutamate-related signals in the human prefrontal cortex, and those changes correlate, imperfectly but meaningfully, with mood improvement.

What this means for next-step antidepressants—and for you

If the glutamate/GABA model is approximately correct, several things follow for clinical care. Patients whose depression has not responded to serotonin-targeted medications are not out of options—they may simply have been treated at the wrong point in the circuit. Treatment-resistant depression is one of the conditions where the glutamate-mechanism rationale is strongest, and where the clinical evidence base for ketamine is most developed.

It also means new drugs in this space—esketamine, and a growing list of compounds that target NMDA, AMPA, or downstream plasticity pathways—will likely keep emerging. Esketamine (Spravato) is FDA-approved for treatment-resistant depression and for major depressive disorder with acute suicidal ideation. Generic racemic ketamine, which is what we use in IV infusions, is FDA-approved as an anesthetic; its use for depression and other psychiatric conditions is off-label.

Honest framing matters here. The glutamate-GABA model is a strong working theory backed by converging evidence from animal work, human imaging, and clinical response data. It is not a final answer, and not every patient responds. Research suggests roughly half to two-thirds of treatment-resistant depression patients show meaningful response to a course of IV ketamine, which is a significant number but not a universal one. We do not promise outcomes. We talk through the evidence honestly and let people decide.

Marla Peterson, CRNA, oversees every infusion at Music City Ketamine and provides anesthesia-level monitoring throughout each session. If you want to read more about the mechanism in plain language, our how it works page is a reasonable next stop.