A compound activates the same receptor as LSD. It triggers the same neuroplasticity cascade that researchers believe underlies the antidepressant effects of psychedelics. It produces no hallucinations at all. By everything scientists thought they understood about how these drugs work, this should not be possible.
Section I
What Was Actually Found?
In early 2026, researchers at the UC Davis Department of Chemistry published a paper in the Journal of the American Chemical Society describing a new library of molecules built from one of life's most ordinary materials: amino acids. The technique is almost alchemical in its simplicity. Take tryptamine, a metabolite your body already makes from the essential amino acid tryptophan. Couple it with various amino acids. Then shine ultraviolet light on the result. The photochemical reaction reshapes the molecules into entirely new structures—structures that happen to be pharmacologically active at the brain's serotonin 5-HT2A receptor.
The receptor matters enormously. It's the same one that LSD, psilocybin, and DMT bind to produce their effects. Activating it powerfully enough is what triggers the cascade of brain rewiring that has made psychedelic-assisted therapy one of the most exciting areas in psychiatry in a generation. Activating it is also what produces visual distortions, ego dissolution, and hours-long altered states that make psychedelic treatment complicated to administer and impossible for a large fraction of patients who need it most.
The UC Davis team, led by Ph.D. students Joseph Beckett and Trey Brasher working with Professor Mark Mascal, screened 100 of these new molecules using computer modeling to see how strongly each one bound to the 5-HT2A receptor. Five candidates emerged for laboratory testing. Their efficacies ranged from 61 percent to 93 percent, which means the top compound—labeled D5—was a full agonist. It could produce the maximum possible biological response at the receptor, the same way LSD does.
So the team gave it to mice and waited for the head twitch response: the rapid, repetitive head-shaking behavior that is the standard animal proxy for hallucinogenic-like effects. It is not a perfect analogue for human hallucination, but it's the best validated behavioral marker available, and it tracks reliably with 5-HT2A activation in compounds that are known to be psychedelic in humans.
D5 did not produce it. A full agonist at the hallucination receptor. No hallucinogenic response.
"The question that we were trying to answer was, 'Is there a whole new class of drugs in this field that hasn't been discovered?'" said Joseph Beckett. "The answer in the end was, 'Yes.'"
This is not an isolated result. A parallel line of research at UC Davis, published in Nature Neuroscience in August 2025, examined a different non-hallucinogenic compound called tabernanthalog (TBG)—a synthetic analogue of ibogaine, a plant-derived psychedelic. TBG also activates the 5-HT2A receptor and also promotes the growth of new dendritic spines in prefrontal cortical neurons. It also does not produce hallucinogenic behavior in animal models. Two different compounds from two different chemical families, built on different molecular scaffolds, arriving at the same paradox: full or partial engagement of the hallucination receptor, with the hallucinations missing.
Section II
Why Did the Old Model Fail?
The received model was clean and intuitive. Psychedelics bind to the 5-HT2A receptor. The 5-HT2A receptor causes hallucinations. The hallucinations and the therapy are a package deal. Some researchers argued the altered state itself was therapeutic—that the ego dissolution and the sense of expanded perspective were not side effects but mechanisms. Others were more agnostic, but nearly everyone assumed that separating the two was at best a distant goal, because they appeared to be caused by the same molecular event.
Key Mechanism
The 5-HT2A Receptor: One Receptor, Two Jobs
Serotonin receptors are proteins embedded in neuron membranes that respond to serotonin, a neurotransmitter the brain uses for signalling between cells. The 5-HT2A subtype is found in high concentrations in the prefrontal cortex, the region most associated with mood regulation, decision-making, and cognitive flexibility. When a molecule binds to it, the receptor can activate multiple downstream signalling pathways simultaneously—different chemical chains that fan out inside the cell and do different things.
One of those pathways runs through a protein called TrkB, the primary receptor for brain-derived neurotrophic factor (BDNF). BDNF is the brain's most potent growth signal. When TrkB activates, it triggers mTOR and AMPA receptor pathways that promote the growth of dendritic spines: the tiny protrusions on neurons where synaptic connections form. More spines mean more connections. More connections are the physical substrate of neuroplasticity—the brain's capacity to rewire itself and form new patterns of thought and response. This is why researchers believe psychedelic-assisted therapy works: it opens a window of heightened neural flexibility during which therapeutic interventions can take deeper root.
The hallucinogenic effect involves a different downstream cascade, one that amplifies glutamate release and disrupts the normal filtering of sensory information in the thalamus and cortex. The key question these new compounds force us to confront is whether the plasticity pathway and the hallucinatory pathway can truly be accessed independently—whether the receptor can be made to do one job without the other.
The model failed for a reason that turns out to be familiar in molecular biology: the receptor was treated as a switch when it is actually a switchboard. Activating the 5-HT2A receptor doesn't produce one response. It initiates a cascade of simultaneous downstream events, and which events dominate depends on how you activate it, from where, and how strongly.
Two factors in particular have emerged as decisive. The first is partial versus full agonism. A full agonist, like LSD, turns on the receptor as completely as possible. A partial agonist activates it only partway. The 2023 Nature Neuroscience study on tabernanthalog found that TBG, as a partial agonist, was sufficient to engage the plasticity pathway—the TrkB/BDNF cascade that drives dendritic spine growth—while apparently not engaging the hallucinatory pathway with enough force to generate behavioral effects. "Full agonists turn on hallucinations and they also turn on plasticity," the UC Davis team noted. "Partial agonists only turn on the receptor part way, and that seems to be sufficient to turn on plasticity."
But D5 is a problem for that clean explanation. D5 is a full agonist. It activates the receptor at 93 percent efficacy. And it still does not produce hallucinogenic behavior. Which means partial agonism cannot be the whole answer.
A second factor may involve where the receptor sits in the neuron. Psychedelics promote neuroplasticity through intracellular 5-HT2A receptors—copies of the receptor located inside the cell, not on its outer membrane. These intracellular receptors appear to be the ones that activate the plasticity cascade, while the membrane-surface receptors are more involved in the hallucinatory response. Serotonin itself, the brain's own molecule, can't reach the intracellular receptors because it can't cross the cell membrane. Psychedelics can, because they're more lipophilic. Some of the new non-hallucinogenic compounds may be hitting the same intracellular targets without adequately activating the surface receptors that drive the hallucinatory response—but the exact mechanism for D5 is not yet characterized.
Figure 1 — Divergent downstream signalling: classical psychedelic vs. new non-hallucinogenic 5-HT2A agonist
There is also a third possibility worth taking seriously. The new scaffold itself may interact with the receptor in a geometrically distinct way, biasing which downstream proteins get recruited. This is called biased agonism, and it's a known phenomenon in receptor pharmacology: two molecules can both bind the same receptor and activate it to similar degrees, yet preferentially engage different internal signalling proteins depending on the exact shape of the binding interaction. The UC Davis researchers found their compound using a photochemical technique that creates molecular geometries that would be very difficult to reach through conventional organic synthesis. The resulting structures may have a distinct binding geometry that favors the plasticity cascade over the hallucinatory one.
The honest answer is that nobody is sure yet. "We determined that the scaffold itself possesses a range of activity," said Trey Brasher. "But now it's about elucidating that activity and understanding why D5 and similar molecules are non-hallucinogenic when they're full agonists."
Section III
What Does the Evidence Actually Show?
Let's be precise about what has and hasn't been established, because the gap between the two matters.
What has been established: these compounds activate the 5-HT2A receptor. Some of them do so at high efficacy levels, up to 93 percent. Tested in mice, they do not produce the head twitch response that tracks with hallucinogenic-like behavior in validated animal models. The Nature Neuroscience study on tabernanthalog went further, demonstrating actual dendritic spine growth in prefrontal cortical neurons of mice—the physical sign of neuroplasticity occurring. When the researchers used genetic techniques to knock out TrkB, the neuroplasticity disappeared, confirming that the TrkB/BDNF pathway is the mechanism. And when they chemically erased the newly grown dendritic spines, the antidepressant-like behavioral effects in stressed mice disappeared too. The chain of causation from receptor activation to spine growth to antidepressant effect is real and has been demonstrated in animal models.
What has not been established: whether any of this translates to humans. The head twitch assay in mice is a proxy, not a proof. No human clinical trials of these specific compounds have been conducted, and they haven't yet gone through the full regulatory pathway that would permit such trials. The TBG work, published in Nature Neuroscience, is the most advanced in terms of mechanistic characterization, but it remains preclinical. The new UC Davis molecules are even earlier in development—the 2026 paper represents the discovery phase, before pharmacokinetics, toxicology, dosing, or efficacy in disease models have been characterized.
"Especially in the psychedelic field, completely new scaffolds are incredibly rare. And this is the discovery of a brand-new therapeutic scaffold."Trey Brasher, Ph.D. candidate, Mascal Lab, UC Davis Institute for Psychedelics and Neurotherapeutics — Journal of the American Chemical Society, 2026
The path from a promising molecule to an approved drug typically runs 10 to 15 years and costs hundreds of millions of dollars. Most candidates fail, often at late stages. The history of drug development for depression in particular is littered with mechanisms that worked brilliantly in rodents and failed to replicate in humans.
That said, the scientific case for pursuing this line of research is unusually solid, for two reasons. First, the mechanism—neuroplasticity through TrkB/BDNF—is well established in humans, not just in animal models. BDNF signalling is known to be impaired in human depression. Ketamine, which is already approved for treatment-resistant depression, appears to work through a related mechanism (it also engages TrkB). The BDNF pathway isn't theoretical; it's the basis of an existing, approved therapy. Second, the evidence that psychedelic-assisted therapy works in humans is now extensive. Multiple Phase 2 trials for psilocybin-assisted therapy have shown remission rates in treatment-resistant depression that exceed conventional antidepressants by a significant margin. The therapeutic signal is real. What remains to be proven is that you can capture it without the hallucinogenic experience.
The psychedelic field has a term for what these compounds might represent: psychoplastogens. The word was coined by David Olson at UC Davis in 2018 to describe compounds that promote neuroplasticity without necessarily producing psychedelic effects. Olson's lab has been building toward this for years. The 2026 compound library represents a new synthetic route to that goal—"a brand-new therapeutic scaffold," in Brasher's words—rather than a modification of an existing one.
That distinction has practical weight. Medicinal chemistry almost always proceeds by modifying known molecules: take psilocybin, change a substituent, see if the pharmacology shifts. The problem with that approach in psychedelics is that you're always one degree of separation from a Schedule I compound, which creates legal, regulatory, and intellectual property complications. A genuinely new scaffold built from amino acids via photochemistry sidesteps that entirely.
Section IV
What Does This Open Up?
The mental health context for this research is not hard to state. Approximately 280 million people worldwide have depression. Of those treated with antidepressants, roughly 30 percent do not respond adequately to two or more medication trials—the threshold that defines treatment-resistant depression. For those patients, the options narrow fast: electroconvulsive therapy, ketamine infusions, and clinical trials. Psychedelic-assisted therapy has shown the most dramatic results yet seen in this population, but it comes with barriers that are genuinely hard to clear.
The barriers are not primarily legal, though the legal status of psilocybin and LSD adds significant regulatory overhead. The deeper barriers are clinical. Psychedelic experiences can last six to eight hours and require trained therapists to supervise them. The experience itself is intensely variable across patients and can be destabilizing. Patients with a history of psychosis or schizophrenia spectrum disorders cannot receive classical psychedelics at all, because 5-HT2A activation can trigger psychotic episodes. That exclusion removes a substantial population of severely ill patients who might benefit from neuroplasticity-based therapy but cannot safely receive the hallucinogenic version of it.
A non-hallucinogenic compound that reliably promotes the same neuroplastic changes would eliminate most of those barriers simultaneously. It could be prescribed like a conventional antidepressant. No supervised six-hour session. No psychosis contraindication if the hallucinatory pathway is genuinely inactive. No special clinical infrastructure. It could reach patients in primary care settings, not just academic medical centers with specialized psychedelic programs.
The question these new results open, and cannot yet answer, is whether the hallucinogenic experience is doing any therapeutic work of its own. This is not a trivial question. Several prominent researchers argue that the psychological experience of ego dissolution—the temporary dissolution of the habitual sense of self—is itself therapeutic, and that trying to strip it out is like trying to separate the surgery from the anesthesia. There is some clinical evidence on both sides. The strongest counter-argument comes from the ketamine data: ketamine does produce an altered state, but it's a dissociative state, not a classic psychedelic experience, and its antidepressant mechanism appears to be primarily neuroplastic. Which suggests the brain rewiring can do substantial work even without the full hallucinogenic encounter.
The tabernanthalog antidepressant data in mice is consistent with that picture. The dendritic spine growth in prefrontal cortex produced measurable antidepressant-like effects in the forced swim test and the sucrose preference test—standard preclinical markers. When TrkB was knocked out, both the spines and the behavioral effects disappeared, leaving the receptor activation unchanged. The plasticity was doing the work. The receptor was the key, but the hallucination was not part of the pathway.
None of this is settled science. The animal data is consistent and coherent, but it is still animal data. The next several years will require the systematic characterization of these new compounds through preclinical toxicology, followed by Phase 1 trials in healthy volunteers, followed by Phase 2 trials in patients. That process will reveal whether D5 and its analogues produce the neuroplastic effects in human neurons that have been observed in rodent models, whether any safety signals emerge, and critically, whether patients who receive these compounds experience measurable relief from depression, PTSD, or addiction without the alterations in perception that currently define the psychedelic class.
The answer, when it comes, will not just be clinically significant. It will settle a question that the field has been circling for years: what exactly is the psychedelic experience for? Is it a side effect of the medicine, or is it the medicine itself? UC Davis may have just built the tool to find out.
Buy me a coffee