The dominant explanation for how psychedelics work has been the serotonin story for two decades. Classic psychedelics bind 5-HT2A receptors. 5-HT2A activation produces altered consciousness. End of story, more or less. Across the integration work I do with founders and senior operators, the people who get the most lasting change are the ones who understand that the 5-HT2A part is the trigger, not the mechanism. The mechanism that actually produces durable shift is glutamatergic, and the integration window depends on protecting it.
Vollenweider and Preller in their 2020 Nature Reviews Neuroscience review consolidated a decade of work showing that 5-HT2A activation on layer 5 pyramidal neurons in the cortex is the upstream trigger for a downstream glutamatergic cascade. That cascade is what engages plasticity machinery, drives BDNF release, and produces measurable dendritic spine growth that persists for weeks. The therapeutic effect is not the trip. The therapeutic effect is the window the trip opens.
This article walks through that mechanism in plain terms. What the glutamate data actually shows. Why AMPA receptor blockade kills the antidepressant signal. Where ketamine's NMDA story converges with the psilocybin AMPA story. And what the model implies for how integration should be timed, not as a metaphor, but as practical protocol. For the broader picture of what shifts in the default mode network during a session, see the DMN explainer.
- Psilocybin increases medial prefrontal cortex glutamate by approximately 9.5 percent within 100 minutes of dosing, measured by proton magnetic resonance spectroscopy across 60 volunteers (Mason et al., Neuropsychopharmacology 2020).
- AMPA receptor antagonism with NBQX abolishes more than 80 percent of psilocybin's antidepressant-like behavioral effect, indicating AMPA is a necessary intermediate, not a parallel effect (Hesselgrave et al., PNAS 2021).
- The Moliner 2023 Nature Neuroscience paper shows psychedelics bind TrkB directly, which positions BDNF signaling as a co-mechanism alongside the glutamatergic cascade.
- Krystal and colleagues in Cell 2024 frame ketamine's mechanism through NMDA antagonism that produces a downstream glutamate burst, converging on the same AMPA and plasticity machinery as classic psychedelics.
- The clinical implication is that the 24 to 72 hour post-session window is when integration practice has measurable biological leverage, and treating it as ordinary recovery time wastes the plasticity that the session paid for.
Why Is the Serotonin Story Incomplete?
The 5-HT2A receptor is necessary but not sufficient to explain what psychedelics do clinically. Vollenweider and Preller in 2020 reviewed two decades of receptor binding, microdialysis, and human imaging data, concluding that 5-HT2A activation is the upstream switch, but the therapeutic phenotype depends on a downstream glutamatergic cascade. The receptor binding finishes within hours. The clinical signal lasts weeks. Something else has to be doing that work.
The mismatch between receptor pharmacokinetics and clinical effect duration is the central puzzle. Psilocin, the active metabolite of psilocybin, clears from blood within roughly 6 to 12 hours. The subjective experience resolves within 4 to 6 hours for most participants. But measurable changes in depression scores, default mode network connectivity, and behavioral flexibility persist for weeks. A receptor that is no longer occupied cannot be doing the work after week two.
The resolution that has emerged across the Vollenweider, Mason, Hesselgrave, Shao, and Moliner papers is that 5-HT2A activation triggers a brief but intense glutamatergic surge in the cortex. That surge then activates AMPA receptors, drives BDNF release through TrkB signaling, and triggers dendritic spine growth. The structural change is what persists, not the receptor occupancy. The serotonin story explains the door. The glutamate story explains the room.
Vollenweider and Preller in their 2020 Nature Reviews Neuroscience review consolidated evidence from receptor pharmacology, rodent microdialysis, and human imaging studies showing that 5-HT2A agonism on layer 5 pyramidal neurons in the prefrontal cortex triggers a downstream glutamate release that engages AMPA and NMDA receptors. The authors framed the therapeutic mechanism as a two-stage process rather than a single-receptor phenomenon. The 5-HT2A binding initiates the cascade within minutes. The glutamatergic and BDNF-mediated downstream events produce structural plasticity that outlasts receptor occupancy by orders of magnitude. The clinical implication is that any model of psychedelic action that stops at the serotonin receptor is incomplete.
What the Pharmacology Cannot Explain
Three observations resist a pure 5-HT2A explanation. First, the antidepressant signal persists weeks after the receptor has cleared. Second, ketamine produces a similar antidepressant signature without binding 5-HT2A at all. Third, blocking AMPA receptors downstream abolishes most of the effect even though 5-HT2A activation is unaffected. Each of these observations points the same direction. The receptor is the trigger. The glutamatergic plasticity is the substance.
What Does the Glutamate Surge Look Like in Human Brain?
Mason and colleagues in 2020 measured a 9.5 percent rise in medial prefrontal cortex glutamate within 100 minutes of oral psilocybin, using proton magnetic resonance spectroscopy in 60 healthy volunteers. The study, published in Neuropsychopharmacology, is the cleanest human evidence that the rodent glutamate story translates upward. The rise was regionally specific and correlated with subjective ego dissolution scores.
The methodology matters. Proton MRS is non-invasive, captures real-time neurochemistry in vivo, and avoids the inferential leaps that come with rodent extrapolation. A 9.5 percent rise sounds modest until you account for the regional concentration. In the mPFC, that increase reflects substantial glutamate efflux from layer 5 pyramidal neurons projecting into the local circuit. The hippocampus in the same study showed a glutamate decrease, suggesting the effect is not a generic surge but a circuit-specific reorganization.
The ego dissolution correlation is the other detail worth holding. Higher mPFC glutamate predicted stronger subjective ego dissolution scores. That correlation does not prove causality, but it links the neurochemical signal to the subjective phenomenology in a way that pure 5-HT2A occupancy does not. The dissolving boundary of self that participants report has a measurable glutamatergic correlate.
Why the mPFC and Not Everywhere
The medial prefrontal cortex is dense with 5-HT2A receptors on the apical dendrites of layer 5 pyramidal neurons. These cells project widely into the local circuit and out to subcortical targets. When 5-HT2A activation fires these cells, the glutamate they release lands on the postsynaptic AMPA and NMDA receptors of nearby neurons. The cascade is regionally concentrated where the receptor density is highest. This is why the mPFC shows up repeatedly in the imaging literature as the locus of psychedelic effect.
How Do AMPA and NMDA Receptors Carry the Signal?
Hesselgrave and colleagues in 2021 demonstrated in PNAS that pre-treatment with the AMPA receptor antagonist NBQX abolishes more than 80 percent of psilocybin's antidepressant-like behavioral effects in rodent forced-swim and tail-suspension models. The interpretation is direct. AMPA engagement is not a parallel pathway. It is the necessary downstream step that converts the 5-HT2A trigger into the lasting behavioral signal.
AMPA receptors are the workhorses of fast excitatory transmission in the cortex. When glutamate binds AMPA, sodium flows in, the postsynaptic neuron depolarizes, and the signal propagates. NMDA receptors, by contrast, are voltage-dependent and require both glutamate binding and prior depolarization to open. NMDA acts as a coincidence detector that gates long-term potentiation. AMPA carries the message. NMDA stamps it as worth keeping.
The clinical translation is that the glutamate surge after psilocybin first activates AMPA, depolarizes the postsynaptic neurons, which then unblocks NMDA receptors and allows calcium influx. The calcium signal triggers the intracellular cascade, including CaMKII activation, AMPA receptor trafficking, and gene expression changes through CREB. This is the same molecular machinery that underlies learning and memory in normal cortex, recruited here in a brief and intense window.
What Blocking AMPA Tells Us
If 5-HT2A activation alone were sufficient for the antidepressant effect, blocking AMPA downstream would not change much. But it does. NBQX pre-treatment eliminates roughly 80 percent of the behavioral signal in the Hesselgrave study, with 5-HT2A activation intact. The trigger fires. The cascade does not. The clinical phenotype disappears. This is the cleanest single experiment that positions AMPA as a necessary intermediate rather than a downstream curiosity.
The NMDA Convergence with Ketamine
Ketamine takes the opposite entry point and lands on the same machinery. Ketamine is an NMDA antagonist at sub-anesthetic doses. The paradoxical antidepressant effect was a puzzle for years until the disinhibition model resolved it. Ketamine blocks NMDA on GABAergic interneurons preferentially, which disinhibits pyramidal neurons and produces a downstream glutamate burst. That burst then activates AMPA, drives BDNF release, and triggers spine plasticity. Different door, same room.
Krystal and colleagues in their 2024 Cell review of ketamine mechanism consolidated evidence that sub-anesthetic ketamine blocks NMDA receptors preferentially on GABAergic interneurons, producing a transient disinhibition of cortical pyramidal cells. The resulting glutamate burst engages AMPA receptors, drives BDNF release through TrkB signaling, and activates the mTOR pathway, which in turn produces dendritic spine remodeling and the rapid antidepressant signal observed within 24 hours of infusion. The convergence with the classic psychedelic mechanism is explicit. Different upstream triggers, including 5-HT2A agonism for psilocybin and NMDA antagonism for ketamine, both produce a downstream glutamate-AMPA-BDNF cascade. The clinical implication is that the plasticity window, rather than the receptor profile, is what makes these molecules therapeutic in mood disorders.
How Do TrkB and BDNF Lock the Change In Place?
Moliner and colleagues in their 2023 Nature Neuroscience paper demonstrated that classic psychedelics, including LSD and psilocin, bind directly to the TrkB neurotrophin receptor at clinically relevant concentrations. The finding repositions BDNF signaling as a co-mechanism rather than only a downstream consequence of glutamate. The plasticity window is reinforced from two directions, the glutamatergic cascade and the direct TrkB binding.
TrkB is the receptor for brain-derived neurotrophic factor, the molecule most consistently associated with antidepressant response, dendritic spine formation, and long-term plasticity. The traditional view was that BDNF released from glutamate-stimulated neurons activated TrkB on neighboring cells. The Moliner finding adds a parallel route. Psychedelic molecules themselves can bind TrkB, dimerize the receptor, and trigger the same downstream signaling that BDNF normally would. Two doors to the same locking mechanism.
The structural plasticity that this produces is measurable. Shao and colleagues in 2021 used in vivo two-photon imaging in mouse mPFC and documented a roughly 10 percent increase in dendritic spine density within 24 hours of a single psilocybin dose. The new spines were stable across at least 30 days of follow-up. This is the structural substrate that explains how a session can produce behavioral change that outlasts the receptor occupancy by weeks or months.
Moliner and colleagues in their 2023 Nature Neuroscience paper reported that LSD and psilocin bind the transmembrane domain of the TrkB neurotrophin receptor at nanomolar affinity, roughly 1000-fold higher than the affinity of standard SSRIs for the same target. The binding promotes TrkB dimerization at the plasma membrane and potentiates BDNF signaling without requiring elevated BDNF release. This finding repositions the plasticity mechanism. The classical view treated BDNF as the downstream consequence of glutamate-driven activity. The Moliner data add a parallel route in which the psychedelic molecule itself acts as a positive allosteric modulator of TrkB. Two doors, the glutamatergic cascade and direct TrkB binding, converge on the same dendritic spine remodeling that defines the plasticity window.
Why the Plasticity Window Is Time-Limited
The glutamate surge resolves within hours. The AMPA and NMDA activation tapers as glutamate clears. The BDNF release and TrkB signaling persist longer, with measurable elevation across roughly 24 to 72 hours. The dendritic spine growth peaks within the first week. After that, the new spines either get pruned through normal turnover or get stabilized through use. Stabilization depends on whether the new circuits get engaged. This is where integration practice has its measurable biological leverage.
"The session is the surgery. The integration window is the rehab. A surgery that goes well still produces a bad outcome if the rehab is skipped, and the same is true here. The plasticity window is the rehab. Treating it as ordinary recovery time wastes the change the session paid for."
Where Does Ketamine Fit in the Glutamate Picture?
Ketamine and classic psychedelics enter through opposite receptor doors but converge on the same downstream glutamate-AMPA-BDNF machinery, which is why both produce rapid antidepressant signals with overlapping plasticity signatures. The Krystal 2024 Cell review framed this convergence as the most important mechanistic insight of the last decade in mood disorder neuroscience. The therapeutic window is the plasticity window, not the receptor profile.
Ketamine blocks NMDA receptors at sub-anesthetic doses. The preferential blockade on GABAergic interneurons produces cortical disinhibition. The disinhibited pyramidal cells release glutamate. The glutamate activates AMPA receptors on neighboring cells. AMPA activation drives BDNF release through TrkB. BDNF triggers mTOR signaling and dendritic spine remodeling. The cascade is the same one that psilocybin reaches through 5-HT2A activation, just entered from a different direction.
The practical implication for someone weighing integration approaches is that the protective window after a ketamine session and after a psilocybin session share more biological substrate than the receptor profiles suggest. Sleep matters in both. Avoiding high-glutamate-load substances such as alcohol matters in both. Engaging the new circuits through deliberate behavior matters in both. The plasticity engine is shared even when the trigger is different.
What Differs Between the Two
The receptor differences produce real phenomenological differences. Ketamine's dissociative profile, shorter duration of subjective effect, and absence of strong 5-HT2A activation make it qualitatively different from psilocybin or LSD. The therapeutic mechanism overlaps. The subjective experience does not. For founders weighing options, the relevant frame is that ketamine and classic psychedelics share the plasticity window, not the experience inside it. The differences in how decisions stabilize after these sessions reflects the experience layer, while the plasticity biology underneath is similar.
What Does the Glutamate Model Mean for Integration?
If the plasticity window peaks within 24 to 72 hours and structural change persists for weeks, integration practice that begins on day seven or day fourteen is starting outside the steepest part of the curve. The biology defines a protocol. Shao and colleagues in 2021 documented dendritic spine density increases of approximately 10 percent within 24 hours of psilocybin in mouse mPFC, with new spines remaining stable across at least 30 days. The first 72 hours are not recovery. They are the rehab window.
The practical structure that uses the biology looks specific. An integration touchpoint within 24 to 48 hours of the session, not a week later. A second touchpoint at one week, when the dendritic spine stabilization phase is active. A third at two weeks, when consolidation is occurring. A fourth at four weeks, when the new pattern either has anchored or has not. Outside that arc, integration becomes ordinary coaching rather than work that uses the plasticity window.
Sleep protection during the first 72 hours is a concrete application. Sleep is when synaptic consolidation occurs, and the literature on sleep-dependent plasticity is mature. Alcohol avoidance during the same window is the other obvious application. Alcohol acutely suppresses glutamatergic transmission and disrupts BDNF signaling, working directly against the cascade the session triggered. For deeper coverage of the sleep side of this, see psychedelics and sleep recovery.
The Pattern Engagement Side
The other application is behavioral. New dendritic spines either get used or get pruned. If the new circuits produced by the session do not get engaged through deliberate behavior, they reduce back to baseline. The participants who get lasting change are the ones who use the plasticity window to install new patterns. The ones who treat the post-session window as recovery time and resume their previous behavior pattern lose most of the change. For more on what a structured session looks like in practice, see what happens in an integration session.
Where the Model Stops
The glutamate model is not a complete account. It explains the molecular mechanism of plasticity well. It does not explain the content of the experience, the specifics of insights that surface, or the meaning-making layer that turns biological change into life change. The biology creates the window. What gets installed in the window is still a human-scale question about meaning, behavior, and identity. The model defines the substrate. It does not write the script. For specific work with high-functioning depression presentations where the model has clear implications, see psychedelics and high-functioning depression.