Mechanistic insights toward dissociating therapeutic from psychedelic effects: bridging the gap between psychedelic research and mental health care

Abstract

Clinical trials investigating psychedelic compounds for depression and anxiety-related disorders are yielding promising preliminary results. Psychedelics produce profound alterations in brain function—such as suppression of the default mode network and thalamocortical dysregulation—leading to intense subjective experiences including ego dissolution and mystical-type states.

While often described as meaningful or therapeutic, these effects can also be unpredictable or distressing. Because of these effects, which require labor-intensive, potentially cost-challenging preparation, administration, integration sessions, and exclude individuals with psychiatric vulnerabilities (e.g., bipolar disorder, borderline personality disorder, schizophrenia), questions remain about their real-world scalability. Consequently, there is growing interest in the possibility of dissociating the therapeutic benefits of psychedelics from their consciousness-altering effects.

In this perspective, we examine emerging clinical and preclinical evidence investigating this scientifically timely and pressing question. Pharmacological strategies such as serotonin 2A receptor (5-HT_2AR) antagonism and the development of “biased” psychedelic analogues or psychedelic-inspired compounds with reduced or null psychedelic potential may represent potential routes through which therapeutic efficacy may be retained even in the absence of psychedelic experiences. Current preclinical data suggest that downstream molecular and network-level mechanisms may mediate therapeutic effects independently of 5-HT_2AR-driven subjective states.

Whether these mechanisms can fully substitute for the experiential component of psychedelic therapy remains to be determined. Nonetheless, confirming this dissociation could mark a turning point in psychopharmacology, enabling the development of next-generation psychedelic-inspired treatments that may be more scalable, accessible, and appropriate for a broader range of psychiatric populations.

Introduction

Psychedelic compounds like psilocybin, lysergic acid diethylamide (LSD), and Ayahuasca have demonstrated beneficial effects on depressive symptoms across numerous trials in the past two decades [1–3]. Psychedelics elicit profound changes in brain activity, which result in alterations of consciousness, mostly mediated by the serotonin 2A receptor (5-HT_2AR), and which can culminate in phenomena known as ego dissolution and mystical experiences. Because adequate psychological support is required prior to, during, and after treatment—and such support is labor- and cost-intensive—scalability of these interventions in real-world settings appears challenging.

Despite the lack of unequivocal causal links, several clinical studies have suggested that the emergence of altered states of consciousness is centrally implicated in the therapeutic effects of psychedelics [4–9]. Preclinical research, however, is increasingly providing evidence for biological correlates and signaling pathways engaged by psychedelics that may contribute to therapeutic-relevant effects and that can be studied independently of pathways associated with phenomena resembling altered states of consciousness. This emerging body of evidence has led to the hypothesis that certain biological effects engaged by psychedelics may be separable from consciousness-altering phenomena; however, whether such effects translate into preserved therapeutic efficacy in the absence of altered states of consciousness remains unproven.

In this Perspective, we examine emerging translational and clinical findings supporting this hypothesis and consider how such a novel framework might enhance the scalability, safety, and accessibility of psychedelic-based treatments for psychiatric disorders.

The 5-HT_2AR mediates the consciousness-altering effects of psychedelics

Supporting a crucial role for the 5-HT_2AR, the consciousness-altering effects of psilocybin, LSD, and Ayahuasca can be prevented, or reversed, by 5-HT_2AR antagonists such as ketanserin and risperidone. Importantly, the extent of 5-HT_2AR blockade and the resulting attenuation of psychedelic effects are dose-dependent for both the psychedelic compound and the antagonist. For example, 20 mg of ketanserin reduced the subjective effects of psilocybin (0.25 mg/kg) by 50–70%, while 40 mg completely inhibited them [10]. Ketanserin (50 mg) also prevented the psilocybin (0.215 mg/kg)-induced electrophysiological alterations and visual-evoked potentials associated with visual hallucinations [11], and the deficits in sensorimotor gating, and controlled inhibition processes (40 mg ketanserin, 0.26 mg/kg psilocybin) [12]. Similarly, ketanserin (40 mg) prevented the LSD (25–200 µg)-induced alterations in consciousness, dream-like imagery, assignment of personal significance to insignificant stimuli, ‘anxious ego dissolution’, and neurobiological correlates such as alterations in neural responses to music, and changes in thalamic and global brain connectivity [13–19]. Additionally, ketanserin (40 mg) before Ayahuasca (0.75 mg/kg DMT dose), considerably (60–80%) attenuated the intensity of the subjective experience, and prevented neurophysiological modifications such as alpha, delta, and theta rhythms decreases [20].

Similarly, 0.5 mg of risperidone reduced the subjective effects of psilocybin (0.25 mg/kg) by 70–80%, and 1 mg suppressed them almost entirely [10]. In contrast, haloperidol (0.021 mg/kg) a dopamine D₂ receptor antagonist) had no impact on the subjective effects and even increased ratings of “oceanic boundlessness” [10, 21]. Importantly, while ketanserin (40 mg) almost fully blocks the consciousness-altering effects of psychedelics, it only partially attenuated the empathogenic effects of LSD (25–200 µg) [14] and did not block the LSD (100 µg)-induced increase in circulating Brain-Derived Neurotrophic Factor (BDNF) levels [22]. This suggests that some of the therapeutic effects of psychedelics might be retained following the removal of the consciousness-altering effects. Supporting that the consciousness-altering effects of psychedelics might not be required for the therapeutic effects, several studies found no correlation between the intensity of the psychedelic effects and therapeutic improvement. For example, in two studies, the short-term and long improvements in individuals with depression following psilocybin treatment were not correlated with the intensity of the subjective or mystical-type experiences [23], although the increase in general well-being was [24]. Similarly, Davis and colleagues found that a complete mystical experience was not necessary for therapeutic benefit [25].

Activation of the 5-HT_1AR may also attenuate the consciousness-altering effects of psychedelics, without interfering with the therapeutic ones. Employing the 5-HT_1AR agonist buspirone reduced the psilocybin-induced hallucinations in humans [26]. Therefore, targeting serotonin receptors beyond the 5-HT_2AR may also influence the ‘psychedelic gradient’—the spectrum of effects elicited by psychedelics depending on dose and receptor interaction profiles.

5-HT_2AR antagonists can also be employed to “reverse” the subjective effects elicited by psychedelics. This could prove valuable, for example, in the rare event of particularly challenging experiences (such as extreme panic or risk of harm to the self or others). In healthy volunteers, 40 mg of ketanserin 1 h after 100 µg of LSD resulted in the reversal of the LSD-induced ego dissolution, altered state of mind, and visual and auditory alterations. Importantly, the duration of the subjective effects decreased by approximately 60%, from 8.5 h to 3.5 h. Therefore, 5-HT_2AR blockade could be useful as an emergency treatment [22].

Clinical evidence

To date, no human studies have determined whether the therapeutic effects of psychedelics require 5-HT_2AR activation and the ensuing psychedelic effects. A case report described a patient affected by TRD who experienced rapid and sustained antidepressant effects from a single 25 mg dose of psilocybin, despite experiencing minimal (if any) consciousness-altering effects. The individual had discontinued all oral medications except trazodone, which was continued at 200 mg nightly for insomnia [27]. Trazodone, an antagonist of 5-HT_2A/2BR, as well as of adrenergic receptors ɑ1 and ɑ2, was previously reported to antagonize the serotonergic stimulation elicited by LSD [28], enhance 5HT neurotransmission [29] and increase 5-HT levels in the rat PFC [30]. This observation suggests that, in specific clinical contexts, antidepressant effects of psilocybin may co-occur with minimal subjective psychedelic experiences; however, given the anecdotal nature of a single case report, these findings should be interpreted cautiously and as hypothesis-generating.

A phase II proof-of-concept RCT aims to determine in Canadian individuals with TRD the feasibility and acceptability of combining the atypical antipsychotic risperidone (a 5-HT_2A/2B/2CR inverse agonist, with additional serotonergic, dopaminergic, and histaminergic targets) with psilocybin (ClinicalTrials.gov ID NCT05710237) [31]. Despite the low 5-HT_2AR selectivity of risperidone, this study provides clinical proof of concept for whether the therapeutic effects of psilocybin require 5-HT_2AR activation. Another phase II RCT aims to determine in patients with MDD whether the 5-HT_2AR antagonist primavanserin blocks the subjective effects of psilocybin, and whether the antidepressant effects are retained after 7 days (ClinicalTrials.gov ID NCT06592833). Lastly, an exploratory phase IIa trial design has been suggested in individuals with moderate-to-severe TRD, involving the administration of psilocybin under general anesthesia with the GABAA receptor positive allosteric modulator propofol to determine if waking consciousness is required for the antidepressant effects of psilocybin [32].

Studies concerning other rapid-acting antidepressants have begun to directly test the necessity of conscious subjective experience for therapeutic efficacy. Converging evidence supporting the dissociation between consciousness-altering effects and therapeutic outcomes comes from recent work by Lii and colleagues investigating ketamine administration under general anesthesia in patients with MDD [33]. In this study, ketamine administered to fully unconscious patients retained its rapid antidepressant efficacy, providing direct clinical evidence that pharmacologically-induced antidepressant effects can emerge independently of conscious subjective experience. However, the control (non-ketamine) group also exhibited comparable improvement, suggesting that nonspecific factors—such as expectancy, perioperative context, or placebo effects—may have substantially contributed to the observed outcomes. As such, the study does not allow definitive attribution of antidepressant effects to ketamine per se. Nevertheless, the similar magnitude of improvement in the ketamine group remains consistent with the possibility that antidepressant effects can occur independently of conscious subjective experience. Although ketamine differs pharmacologically from classic psychedelics, these findings are consistent with the possibility that antidepressant mechanisms may not require conscious awareness of drug effects, thereby strengthening the rationale for similar investigations in the psychedelic domain.

While clinical evidence remains limited, emerging experimental approaches are beginning to directly try to address this question. A recently published study protocol investigating lower doses of psilocybin (1 and 10 mg, as opposed to the widely adopted 25 mg, which elicits full-blown psychedelic effects) for obsessive-compulsive disorder outlines a mechanistically informed human model aiming to disentangle 5-HT_2AR–mediated neurocognitive effects from subjective psychedelic experiences [34].

In addition, CYB-003, a deuterated psilocin derivative, has undergone Phase I/IIa clinical evaluation to assess safety, tolerability, pharmacokinetics, and pharmacodynamics in alcohol use disorder and MDD (ClinicalTrials.gov ID NCT05385783). While CYB-003 retains psychedelic effects, its pharmacokinetic profile may allow for a shorter and more predictable treatment experience. Following promising interim Phase II data [35], CYB-003 received breakthrough designation by the FDA [36]. A Phase III pivotal trial is planned to further assess the therapeutic value of this compound.

Another potential avenue to decouple the consciousness-altering from therapeutic effects of psychedelics is to employ close structural analogues with reduced psychedelic properties, for example due to lower signaling efficacy at the 5-HT_2AR, rather than altered receptor bias [37] or different potency. One study reported that repeated administration of the non-psychedelic LSD analogue lisuride led to improvements in depressive and anxiety symptoms in 80% of patients with post-stroke depression who received treatment [38].

In sum, current clinical evidence remains limited, and no RCTs have yet demonstrated that psychedelics exert antidepressant effects independent of 5-HT_2AR–mediated psychedelic experiences. At present, this possibility is supported by isolated clinical observations, and therefore definitive conclusions cannot be drawn. Case reports and proof-of-concept studies involving agents such as trazodone, risperidone, primavanserin, and propofol point toward plausible dissociation between therapeutic and psychedelic mechanisms clinically. However, these findings remain suggestive and limited by sample size, methodological constraints, and confounding pharmacodynamic effects. Until results from well-powered RCTs become available, the necessity of 5-HT_2AR activation—and the accompanying subjective experience—for psychedelic-induced antidepressant effects remains an open question. Future RCTs, particularly those incorporating 5-HT_2AR antagonists, anesthesia, or functionally selective psychedelic analogues, are essential to unequivocally address this fundamental issue.

Preclinical evidence

Preclinical research allows for in-depth investigation of the cellular and molecular mechanisms through which psychedelic compounds exert their effects, although important differences between rodent models and human experience remain. In particular, disease phenotypes and the subjective or affective dimensions of psychedelic states cannot be faithfully reproduced in animals. Nevertheless, animal models are valuable for examining the functional and biological consequences of psychedelic exposure and for dissecting the underlying signaling pathways.

Evidence from preclinical studies suggests that therapeutic benefits may be separable from the hallucinogenic properties of these compounds. A major limitation in this field is the absence of reliable animal correlates of the subjective psychedelic experience. Consequently, research has relied heavily on indirect behavioral measures, most notably the head-twitch response (HTR), a stereotyped motor behavior elicited in rodents by serotonergic psychedelics, which is commonly used as a proxy for 5-HT_2AR activation. The HTR is widely regarded as a reliable proxy of 5-HT_2AR activation, with its magnitude correlating with the compound’s potency at this receptor, rather than as a model of subjective or phenomenological effects [39]. The HTR is blocked by pretreatment with 5-HT_2A receptor antagonists, suggesting agonism at the 5-HT_2A receptor is responsible for this behavior. This section explores preclinical evidence involving non-psychedelic 5-HT_2A receptor agonists or the coadministration of psychedelic 5-HT_2A agonist with a 5-HT_2AR antagonist (Table 1).

Table 1.

Summary of preclinical studies investigating the behavioral and neurobiological effects of psychedelic and non-psychedelic 5-HT_2AR agonists in animal models of depression and anxiety, with or without blockade of hallucinogenic-like activity.

Study	Animal model characteristics		Intervention characteristics				Behavioral outcome (depression/anxiety)		Neurobiological correlates
	Species (strain), sex, weight/age	Phenotype induction	Investigated substance	Hallucino-genic activity Y/N	Dose, schedule, route	Block of hallucinogenic effect (5HT2AR blockade) (timing, dose, route)	Test(s) employed	Results	Evaluated parameter(s)	Results
Cameron et al., [57]	Mice (C57/BL6J) males and females, 8 weeks	Unpredictable mild stress (7 days)	Tabernanthalog (ibogaine analog)	N	10 or 50 mg/kg, single i.p. injection	Not applied	Forced swim test (1 day after administration)	Tabernanthalog induced an antidepressant-like effect	Spine dynamics in sensory cortex (1 day after administration)	Tabernanthalog induced an increase in spine formation
	Mice (C57/BL6J) males, 9–10 weeks	Healthy			50 mg/kg, single i.p. injection	Pretreatment (10 min prior) with ketanserin (4 mg/kg, i.p. injection)	Forced swim test (1 day after administration)	Tabernanthalog-induced antidepressant-like effect blocked by pretreatment	Dendritic arbor complexity in rat embryonic cortical neurons	Tabernanthalog-induced effect on structural plasticity blocked by pretreatment
Arias et al., [58]	Mice (C57BL/6J), males, 30–35 g	Healthy	DM506 (ibogamine derivative)	N	5 or 15 mg/kg, single i.p. injection	Pretreatment (15 min prior) with volinanserin (0.05 mg/kg, s.c. injection)	Novelty-suppressed feeding (in stressed mice) and elevated zero maze (in naïve mice) tests (1 h and 1 day after administration)	DM506-induced (15 mg/kg) anxiolytic-like effect inhibited by pretreatment	–	–
Kaplan et al., [59]	Mice (C57BL/6J), females and males, 3–5 months	Vesicular monoamine transporter 2 (VMAT2) heterozygous (HET) mice (genetic model)	(R)-69, (R)-70 (5HT2AR agonists)	N	0.5 or 1.0 mg/kg, single i.p. injection	Not applied	Tail suspension test	(R)-69 and (R)-70 normalized the enhanced immobility of VMAT2 HET mice	–	–
	Mice (C57BL/6J), females and males, 4 months	Mice assigned to foot-shock (FS) and non-FS (NFS) conditions (learned helplessness model)	(R)-70 (5HT2AR agonist)	N	1.0 mg/kg, single i.p. injection	Not applied	Sucrose preference, tail suspension, and elevated zero maze tests	(R)-70 induced antidepressant- and anxiolytic-like effects in FS mice	–	–
Cao et al., [51]	Mice (C57/BL6J, B6D2F1), not available, 8 weeks	Acute restraint stress (ARS) for 5 h	IHCH-7079, IHCH-7086 (β-arrestin-biased 5HT2AR agonists)	N	10.0 mg/kg, single i.p injection	Combined treatment with MDL100907 (1 mg/kg, i.p. injection)	Forced swim and tail suspension tests (30 min after administration)	IHCH-7079 and IHCH-7086 attenuated ARS-induced immobility; their effect was abolished by pretreatment	–	–
	Mice (C57BL/6), males, 8 weeks	Chronic corticosterone administration (25 μg/ml, 21 days)		N				IHCH-7079 and IHCH-7086 reduced corticosterone-induced immobility; their effect was abolished by pretreatment	–	–
Dong et al., [52]	Mice (C57BL/6J), females and males, 9–10 weeks	Healthy	AAZ-A-154 (5HT2AR agonist)	N	20 mg/kg, single i.p injection	Not applied	Forced swim test (30 min and 1 week after administration)	AAZ-A-154 induced antidepressant-like effects in both models	–	–
	Mice (C57BL/6J), females and males, adult	Vesicular monoamine transporter 2 (VMAT2) heterozygous (HET) mice (genetic model)		N	15 mg/kg, single i.p injection		Sucrose preference test (1 day and up to 12 days after administration)		–	–
Lewis et al., [53]	Mice (C57BL/6), females and males, 7–8 weeks	Healthy	2-Br-LSD (LSD analog)	N	0.3, 1.0 or 3.0 mg/kg, single i.p. injection	Pretreatment (15 min prior) with volinanserin (M100907) (0.125 mg/kg, i.p. injection) (only with 1 mg/kg 2-Br-LSD)	Forced swim and open field tests (1 day after administration)	2-Br-LSD induced anxiolytic- (females only) and antidepressant-like effects; the latter were blocked by pretreatment	Synaptic spine analysis in the prefrontal cortex (1 day after administration)	2-Br-LSD-induced an increase in average spine density
	Mice (C57BL/6), females, 8 weeks	Chronic variable stress (CVS) regimen for 5 weeks (2 stressors per day)			3.0 mg/kg, single i.p. injection; 1 mg/kg, 4 i.p. injections every 48 h	Not applied	Splash and open field tests (2 and 4 h after last administration)	2-Br-LSD-induced a partial reversal of the maladaptive effects of CVS	–	–
Qu et al., [60]	Mice (C57BL/6), males, 8 weeks	Lipopolysaccharide (LPS) administration	Lisuride (ergoline derivative)	N	1.0 mg/kg, single i.p. injection	Not applied	Forced swim test (3 h after administration)	Lisuride ameliorated LPS-induced increase in depressive-like behavior	Dendritic spine density	Lisuride ameliorated LPS-induced decrease in dendritic spine density
Cao et al., [51]	Mice (C57/BL6J, B6D2F1), not available, 8 weeks	Acute restraint stress (ARS) for 5 h	Lisuride (ergoline derivative)	N	1.0, 0.25, 0.0625 mg/kg, single i.p. injection	Not applied	Forced swim and tail suspension tests (30 min after administration)	Lisuride attenuated ARS-induced immobility	–	–
Pogorelov et al., [61]	Mice (C57BL/6J), females and males, 3–8 months	Vesicular monoamine transporter 2 (VMAT2) heterozygous (HET) mice (genetic model)	Lisuride (ergoline derivative)	N	0.5 or 1.0 mg/kg, single i.p. injection	Not applied	Sucrose preference and tail suspension tests	Lisuride exerted antidepressant-like effects	–	–
Hesselgrave et al. [42]	Mice (C57BL/6J, males, 8 weeks	Chronic multimodal stress (CMMS)	Psilocybin	Y	1 mg/kg, single i.p. injection	Pretreatment (60 min prior) with ketanserin (2 mg/kg, i.p. injection)	Sucrose preference and female urine preference tests	Psilocybin-induced antianhedonic-like effect not prevented by pretreatment	Strengthening of excitatory synapses in the hippocampus	Psilocybin-induced restoration of synaptic strength not prevented by pretreatment
Shao et al., [41]	Mice (Thy1GFP), females and males, 8–10 weeks	Healthy	Psilocybin	Y	1 mg/kg, single i.p. injection	Pretreatment (10 min prior) with ketanserin (1 mg/kg, i.p. injection)	–	–	Dendritic remodeling in the medial frontal cortex	Psilocybin-induced structural plasticity not blocked by pretreatment
Morales-Garcia et al., [45]	Mice (C57/BL6), males, 3-months	Healthy	DMT	Y	2 mg/kg, i.p. injections on alternating days for 21 days	Combined treatment with ritanserin (0.2 μg/animal, i.p. injection)	Morris water maze and novel object recognition tests (hippocampal neurogenesis-related tests)	DMT-induced improved performance not nullified by combined treatment	Activation, stem cell differentiation into neuronal cells, and neurogenesis of the subgranular neurogenic niche	DMT-induced enhancement of neurogenesis not prevented by combined treatment
Moliner et al., [46]	Mice (C57BL/6NTac), females and males, 17–19 weeks	Chronic stress consisting of five consecutive pretest sessions (5 min) of the forced swim test	LSD	Y	0.1 mg/kg, single i.p. injection	Pretreatment (30 min prior) with M100907 (0.5 mg/kg, i.p. injection)	Forced swim test (7 days after administration)	LSD-induced antidepressant-like effect not prevented by pretreatment	Binding with the receptor for BDNF (TrkB) and promotion of endogenous BDNF signaling, spinogenesis, and dendritogenesis	LSD-induced effect on neurotrophic signaling and plasticity not prevented (similar results with ketanserin pretreatment with LSD/psilocin)
Sekssaoui et al., [43]	Mice (C57BL/6J, females and males, 8-9 weeks	Chronic stress consisting of five consecutive pretest sessions (10 min) of the forced swim test (chronic behavioral despair model, CDM) in 5-HT2AR -deficient (5-HT2A−/−) and wild type (WT) mice	Psilocybin	Y	1 mg/kg, single i.p. injection	5-HT2AR-deficient (5-HT2A−/−) mice	Novelty-suppressed feeding, sucrose preference, and forced swim tests (2 days after administration)	Psilocybin induced in 5-HT2A−/− and WT CDM mice the same antidepressant-like effects	–	–
			DOI	Y	1 mg/kg, single i.p. injection			DOI did not produce in 5-HT2A−/− CDM mice the antidepressant-like effects that were observed in WT CDM mice	–	–
de la Fuente Revenga et al., [49]	Mice (C57BL/6), males, 10-20 weeks	5-HT2AR-deficient (5-HT2A−/−) and wild type (WT) mice	DOI	Y	2 mg/kg, single i.p. injection	5-HT2AR-deficient (5-HT2A−/−) mice	Fear conditioning test (1 day after administration)	DOI did not produce in 5-HT2A−/− mice the effects on contextual fear extinction that were observed in WT mice	Dendritic spine density in cortical pyramidal neurons	DOI did not produce in 5-HT2A−/− mice the effects on dendritic spine density that were observed in WT mice
Aarrestad et al., [44]	Mice (C57BL/6J, age 5–10 weeks; CD-1, age 10–13 weeks; 5-HT2AR KO) males and females, 5–20 weeks, Rats (Sprague Dawley) males and females, 8–12 weeks; (Wistar rats) males and females, 7–12 weeks Pigs (Danish domestic pigs), females, 8–10 weeks	Chronic IFN-α administration, chronic social defeat stress	Tabernanthalog, BOL-148, AAZ-A-154, 6-F-DET LSD, DMT, 5-MeO-DMT, ketamine, psilocybin	N Y	Tabernanthalog (1, 3, 10, 15, 30 and 50 mg/kg), 5-MeO-DMT (10 and 50 mg/kg), ketamine (3, 5 and 10 mg/kg), Psilocybin (0.08, 0.8, 3 mg/kg), AAZ-A-154, (20 mg/kg) single i.p. injection	Not applied	Tail suspension test Forced swim test Social preference test Head-twitch response	Non-hallucinogenic analogues-induced antidepressant-like effects require 5-HT2ARs	Voltage-clamp electrophysiology of mPFC layer 5 pyramidal neurons Fiber photometry Two-photon imaging Dendritic spine density Radioligand binding assay Aequorin assay IP1 assay BRET assay TrkB functional assay MAO inhibition assays NET and SERT binding assays Microdialysis snRNA-seq Immediate early gene expression	Non-hallucinogenic analogues-induced dendritic spine plasticity requires 5-HT2Ars Non-hallucinogenic analogues do not induce immediate early gene expression
Shao et al., [56]	Mice (C57BL/6J, Fezf2-2A-creER (B6;129S4-Fezf2tm1.1(cre/ERT2)Zjh/J, 036296), Plxnd1-2A-creER (B6;129S4-Plxnd1tm1.1(cre/ERT2)Zjh/J, 036296), Thy1GFP line M (Tg(Thy1-eGFP)MJrs/J, 007788), and Camk2acre (B6.Cg-Tg(Camk2a-cre)T29-1Stl/J, 005359) males and females, 5–8 weeks	Chronic restraint stress	Psilocybin	N	1 mg/kg, single i.p. injection	Localized mPFC KO of 5-HT2AR-expressing neurons (through AAV-hSyn-Cre-P2A-dTomato) Camk2acreHtr2af/f mice (mPFC-specific 5-HT2AR KO)	Learned helplessness Tail suspension test Stress-induced resistance to fear extinction	Antidepressant- and/or anxiolytic-like effects	In vivo two-photon structural imaging of dendritic spines in the mPFC Calcium imaging of dendrites Confocal imaging of dendrites In vivo multi-unit electrophysiology of mPFC neurons Ex vivo electrophysiology Single-cell RNAseq qRT-PCR Immunohistochemistry	mPFC-specific 5‑HT2AR knockout abolishes the psilocybin-induced dendritic remodeling in mPFC pyramidal tract neurons and psilocybin-induced decrease of learned helplessness 5-HT2AR activation on PT neurons of the mPFC are required for psilocybin’s therapeutic effects
Puigseslloses et al., [55]	Mice (Swiss CD‑1) males and females, 6–8 weeks	Repeated tail suspension stress (6 min/day ×4 days)	5-Br-DMT DMT 5-F-DMT, 5-Cl-DMT	N Y	0, 0.3, 1, 3, 10 and 30 mg/kg, single i.p. injection	Pretreatment with 5‑Br‑DMT (10 mg/kg i.p. injection) prior to 5‑F‑DMT (1 mg/kg, i.p. injection)	Tail suspension test	Antidepressant-like effects	–	–

The table outlines animal characteristics, experimental interventions, behavioral outcomes, and neuroplasticity-related findings, highlighting efforts to decouple therapeutic from consciousness-altering effects.

Recent work has expanded this framework by demonstrating that psychedelic-induced functional and structural plasticity in cortical neurons lacking the 5-HT_2AR can emerge from circuit- and compartment-specific mechanisms. Notably, the long-lasting increased frequency of spontaneous excitatory postsynaptic currents in low rheobase neurons (that lack expression of postsynaptic 5-HT_2AR) of the retrosplenial cortex elicited by 25CN-NBOH is mediated by presynaptic 5-HT_2AR on thalamocortical afferents originating in the anterior thalamus [40].

Suppression of psychedelic-like effects does not necessarily abolish therapeutic or plasticity-related outcomes. For instance, ketanserin pretreatment does not prevent dendritic remodeling in the medial frontal cortex following exposure to several serotonergic psychedelic compounds [41], nor does it block the antianhedonic and electrophysiological effects of psilocybin in chronically stressed mice [42], while psilocybin retains antidepressant-like efficacy in chronically stressed 5-HT_2AR null mice [43]. Consistently, the non-hallucinogenic psychoplastogen tabernanthalog induces robust antidepressant-like and spinogenesis-related effects through 5-HT_2A, TrkB, mTOR and AMPA receptor activation, despite minimal psychedelic-like activity and no glutamate burst or proximate immediate early gene activation [44].

Mechanistic studies have also highlighted that therapeutic-like outcomes can be supported by signaling routes and circuit adaptations beyond canonical hallucinogenic pathways. DMT-induced enhancement of neurogenesis and learning performance depends on sigma-1 receptor rather than 5-HT_2AR activation [45]. Moreover, LSD and psilocin exert antidepressant-like and plasticity-related effects through direct TrkB engagement and activation of endogenous BDNF signaling [46]. In parallel, recent mechanistic evidence suggests dissociable neural representations of psychedelic-like versus therapeutic-like endpoints in opioid addiction-based models [47, 48].

Nevertheless, intact 5-HT_2AR signaling remains necessary for some of the behavioral and electrophysiological outcomes of serotonergic psychedelics. For example, DOI produces antidepressant-like effects and increases dendritic spine density in chronically stressed wild-type mice but not in 5-HT_2AR null animals [43, 49]. Additionally, the psilocybin- and NBOH-2C-CN-induced increase in firing activity in ex vivo 5-HT_2A--expressing pyramidal neurons of the PrL/ACC mPFC is blocked following the administration of a 5-HT_2AR blocker [50]. These findings emphasize that dissociation between hallucinogenic and therapeutic effects is compound- and context-specific.

A complementary strategy involves the development of non-hallucinogenic or weakly psychedelic analogs. Non-psychedelic 5-HT_2A agonists such as IHCH-7079 and IHCH-7086 reduce stress-induced immobility [51], while AAZ-A-154 decreases immobility and increases sucrose preference in anhedonia-prone models [52]. Weakly serotonergic psychedelic analogs, including 2-Br-LSD, 5-Br-DMT, and JRT, promote immediate early gene expression, dendritogenesis, antidepressant-like effects, and cognitive flexibility with reduced psychedelic-like activity [53–55].

Cell-type and circuit specificity also appear critical for understanding when dissociation occurs. For example, recent work indicates that psilocybin-induced therapeutic effects and dendritic plasticity are mediated by activation of the 5-HT_2AR expressed by defined neuronal subpopulations in discrete brain regions, such as mPFC pyramidal tract neuron, which represent the subcortical output pathway [56].

Ibogaine-derived compounds provide particularly compelling evidence for dissociation. Tabernanthalog preserves robust psychoplastogenic effect (defined as the ability to rapidly induce structural and functional neural plasticity, including dendritic growth, spine formation and synaptogenesis), antidepressant-like, and anti-addictive effects despite lacking psychedelic-like activity, although these effects remain 5-HT_2A dependent [44, 57]. Similarly, the ibogamine derivative DM506 induces anxiolytic-like effects that are abolished by 5-HT_2A antagonism [58]. Additional support comes from low-psychedelic 5-HT_2A agonists such as (R)-69 and (R)-70, which normalize depressive- and anxiety-like behaviors in genetic and learned helplessness models [59].

Finally, the non-psychedelic 5-HT_2A agonists lisuride, tabernanthanalog, and IHCH-7079 represent an instructive example of context-dependent dissociation, modulated by reduced 5-HT_2AR signaling efficiency, rather than ligand bias properties [37]. In inflammatory models of depression, lisuride ameliorates depressive-like behavior and modulates dendritic spine density independently of 5-HT_2AR activation [60], while in other paradigms it attenuates stress-induced immobility [51], exerts antidepressant effects in VMAT2 heterozygous mice [61], and produces long-lasting anxiolytic- and antidepressant-like effects in wild-type but not 5-HT_2AR null mice [43]. Earlier studies also demonstrated dose-dependent reductions in immobility elicited by lisuride to be antagonized by haloperidol [62].

Mechanisms by which 5-HT_2AR activation could lead to psychedelic versus therapeutic effects

Biased agonism at the 5-HT2AR

The 5-HT_2AR, a Class A G protein-coupled receptor (GPCR), is central to the psychedelic effects of psychedelics. The observation that some agonists of this receptor induce profound psychedelic effects while others do not, yet produce therapeutic benefits, has led to the speculation that distinct intracellular signaling pathways downstream of 5-HT_2A activation underlie these divergent outcomes. One key concept that has emerged to explain this phenomenon is biased agonism, or functional selectivity—the ability of a ligand to preferentially activate specific signaling cascades, despite binding to the same receptor [63–65].

Canonical signaling through the 5-HT_2AR primarily involves coupling to Gq proteins, which leads to phospholipase C (PLC) activation, generation of inositol triphosphate (IP3), calcium release from the endoplasmic reticulum, and activation of protein kinase C (PKC) [66, 67]. In parallel, the activated 5-HT_2AR can recruit β-arrestin2, which facilitates receptor desensitization and internalization, and initiate non-canonical signaling pathways such as the phosphoinositide 3-kinase (PI3K)/Src/protein kinase B (Akt) cascade [66, 67]. Although no consensus exists toward a specific pathway that would be responsible for the psychedelic effects, both kinases PKC and AKT are important promoters of neurite elongation [68–70], implying both Gq and β-arrestin 5-HT_2A mediated pathways could be theoretically implicated in the therapeutic actions of psychedelics.

It has been proposed that the consciousness-altering properties of psychedelics may be specifically mediated by β-arrestin2 [71], given that LSD exhibits a bias toward β-arrestin2 recruitment relative to serotonin [72] and that β-arrestin2 null mice are less sensitive to LSD-induced disruptions in prepulse inhibition and show reduced HTR [71]. However, this interpretation has been challenged: the extent of β-arrestin recruitment varies among psychedelics, and β-arrestin-biased ligands failed to induce the HTR in rodents, while the magnitude of the HTR correlates with Gq activation [73–76]. Chronic treatment with β-arrestin-biased ligands attenuates the DOI-induced HTR, implicating β-arrestin2, likely through the desensitization of 5-HT_2A, as responsible for psychedelic tachyphylaxis [76]. Additionally, β-arrestin2 KO mice responded normally to DOI-induced HTR [74]. This suggests that Gq but not β-arrestin2 may be the primary driver of psychedelic signaling. Therefore, developing β-arrestin2-biased ligands might uncover novel psychedelic-inspired therapeutics with decreased consciousness-altering potential.

Investigations into G protein coupling have further complicated the binary Gq vs. β-arrestin model. GPCRs are known for their coupling promiscuity [77, 78], and 5-HT_2A is no exception. In addition to its canonical interaction with Gq, the 5-HT_2AR can also couple with other Gq family proteins, including G11, G14, and G15 [73, 79], as well as members of the Gi family (Gi1, Gi2, Gi3), with evidence for these interactions observed both in recombinant systems and in postmortem human and rodent cortex [80]. Genetic inhibition of Gi1 or Gi3 via antisense oligodeoxynucleotides abolished the HTR in rodents, while pharmacological inhibition of Gq using YM-254890 did not [80]. Conversely, Gq inhibition significantly reduced 5-HT_2A-mediated memory deficits, whereas Gi inhibition had no effect [80]. These findings suggest a possible dissociation between psychedelic and other cognitive effects, where Gi signaling would be implicated in the psychedelic effects and Gq signaling to memory and cognition improvements [80].

Together, while the possibility that the therapeutic and psychedelic effects of psychedelics are mediated by distinct pathway biases exists, current evidence does not support a simple dichotomy. Contextual variables, such as receptor coupling promiscuity, ligand structure, and experimental model systems, significantly influence signaling outcomes and must be considered when interpreting these data. To reduce these uncertainties, future research should prioritize physiologically relevant experimental models that better capture the complexity of neural systems, such as native neuronal preparations, patient-derived induced pluripotent stem cell–based models, and in vivo systems preserving intact circuitry and neuromodulatory dynamics. Moreover, advanced signaling assays should be employed that can discriminate pathway-specific activations and systematically investigate how ligand structures affect receptor signaling in diverse contexts.

Receptor dimerization and crosstalk with other receptors

Beyond biased agonism, the signaling landscape is further shaped by receptor–receptor interactions through heterodimerization [81–83]. These lateral allosteric interactions can profoundly alter receptor pharmacodynamics, modulate ligand binding, and reroute downstream signaling pathways.

The 5-HT_2AR forms functional heteromers with several receptors, including the dopamine D2 receptor [84, 85], the 5-HT_1AR [86], the cannabinoid CB1 receptor [87], and the metabotropic glutamate receptors mGlu2 and mGlu3 [88, 89]. Among these, the 5-HT_2A/mGlu2 heteromer has received particular attention due to its putative implications in psychedelic drug action and schizophrenia. This heteromeric complex enables mGlu2, which normally couples to Gi/o proteins, to signal through Gq-like calcium pathways—a response that is blocked by the 5-HT_2A antagonist volinanserin [90]. Importantly, the behavioral and transcriptional effects of psychedelics, such as DOI-induced expression of the psychedelic-specific 5-HT_2A-mediated immediate early gene egr-2 [91], are abolished in mGlu2 knockout mice, while expression of c-fos, a general marker of 5-HT_2A activation, remains intact [92]. These findings suggest that the 5-HT_2A/mGlu2 interaction might be necessary for the psychedelic response but not for other 5-HT_2AR-mediated functions.

The 5-HT_2AR is also capable of homodimerization [84, 93, 94], altering agonist binding in a ligand-dependent manner and resulting in pathway-specific antagonism [93]. Therefore, the dimeric state of 5-HT_2A introduces an additional layer of signaling specificity that could influence the clinical profiles of different ligands. Receptor dimerization and heteromerization might thus “texturize” the functional output of 5-HT_2A activation.

Cellular context and subcellular localization as determinants of 5-HT_2A signaling

Another critical layer of regulation is the cellular context in which 5-HT_2ARs are expressed. The subcellular localization of receptors, their interactome, and the cell-type-specific expression of scaffolding proteins collectively shape GPCR signaling outcomes [95]. The postsynaptic density protein 95 (PSD-95), for instance, binds to the C-terminal 5-HT_2A tail and anchors the receptor at excitatory synapses, promoting Gq-coupled membrane signaling and reducing its β-arrestin₂-mediated internalization [96, 97]. Genetic deletion of PSD-95 in mice decreases the HTR, suggesting a crucial role of this scaffolding protein in the psychedelic effects [98].

Recent work has proposed that the therapeutic effects of psychedelics—particularly their neuroplastic effects—may depend on the activation of intracellular receptor pools, rather than the surface-localized ones [99]. Internalized 5-HT_2AR, localized to endosomes or the Golgi apparatus, may engage distinct signaling pathways from those at the plasma membrane, including cascades related to cellular growth and plasticity, such as mTOR and TrkB pathways.

Considering this evidence, the spatial segregation of receptor populations and the associated signaling outcomes provides a potential framework to explain how structurally similar 5-HT_2A agonists can produce divergent behavioral and therapeutic profiles. Ligands that preferentially engage surface-localized, PSD-95–bound receptors may drive psychedelic responses through G-protein signaling, whereas those that access and activate internal receptor pools may induce therapeutic plasticity without perceptual distortions. In summary, the functional outcomes of 5-HT_2A activation are determined not only by the pharmacological properties of ligands but also by the conformational states, receptor dimerization, intracellular localization, and cellular context. These multiple levels of regulation underline a putative complex balance between hallucinogenic and therapeutic signaling through the 5-HT_2AR.

Mechanistic strategies to dissociate therapeutic effects from psychedelic-like phenomena

Several key areas warrant further exploration. One concerns the dose-dependent modulation of subjective effects via 5-HT_2AR engagement, supporting the concept of a “psychedelic gradient,” whereby psychedelic effects vary as a function of dosage and receptor activation. This framework is consistent with evidence showing that increasing psychedelic doses are associated with graded increases in brain entropy, a measure of neural signal diversity and dynamical complexity derived from neuroimaging studies, which has been linked to the intensity of subjective psychedelic experiences [100, 101].

Randomized controlled trials will be crucial to determine whether decoupling the consciousness-altering from therapeutic effects of psychedelics can be achieved. Multimodal imaging studies will be instrumental in determining if the brain connectivity changes and increased brain entropy elicited by psychedelics are also elicited by psychedelic-inspired non-psychedelic drugs.

Mechanistic evidence supports the feasibility of dissociating psychedelic from therapeutic effects. Psychedelics might elicit therapeutic effects through the TrkB receptor without requiring 5-HT_2A activation [46]. Additionally, psychedelics engage glutamatergic signaling and downstream pathways including AMPA receptors, mTOR, EGFR, ERK, Taar1, and c-fos, forming a BDNF–TrkB–mTOR feedback loop that promotes plasticity and therapeutic effects [1–3]. This suggests that downstream neuroplasticity-related signaling may be engaged directly by psychedelic-inspired non-psychedelic drugs. If confirmed, the significance of this capacity to rapidly enhance structural and functional plasticity [102] would underscore the need to move beyond approaches focused on correcting ‘neurochemical imbalances’, toward strategies designed to open windows for the reorganization of neural circuits.

The anti-inflammatory and immunomodulatory properties of psychedelics may also play a role in this dissociation. Psychedelics such as psilocybin, Ayahuasca, and ketamine acutely and persistently reduce inflammatory cytokines in humans, potentially via biased 5-HT_2A and vagal pathways [1, 103, 104]. Sigma-1 receptor signaling may also mediate some of the therapeutic effects independently from 5-HT_2AR signaling. DMT and 5-MeO-DMT are weak agonists at sigma-1 receptors, yet DMT-induced neurogenesis and anti-inflammatory effects depend on interactions with this receptor, rather than the 5-HT_2AR [45, 105, 106]. The context of administration remains a critical factor. Indeed, the setting in which psychedelics are delivered is understood to profoundly influence therapeutic outcomes [107]. Future studies manipulating environmental variables in psychedelic therapy are therefore required to assess causality. In parallel, expanding preclinical research using refined behavioral assays and ethologically relevant animal models will be essential to further interrogate the dissociation between psychedelic-like and therapeutic effects at molecular, cellular, electrophysiological, and systemic levels. In particular, the adoption of promising translational paradigms, such as recently developed assays focusing on reward processing and motivational behavior [108], as well as the rodent version of the Probabilistic Reward Task [109], may allow for a more precise dissection of anhedonia- and reward-related dimensions. Combined with established stress-based models, such as chronic unpredictable stress-exposed rodents or Wistar–Kyoto rats, these approaches could substantially enhance the translational relevance of preclinical findings and improve the alignment between animal models and clinically meaningful symptom domains.

The ability to dissociate the consciousness-altering effects from the therapeutic benefits of psychedelics could redefine psychiatric treatment paradigms, paving the way for more scalable, accessible, and personalized interventions for depression and related conditions. Although clinical evidence is still limited, artificial intelligence (AI)–driven approaches are increasingly being investigated to aid the discovery of psychedelic-inspired compounds and optimize clinical trials with improved safety and efficacy profiles. These strategies encompass multiple facets, including pharmacodynamics, study design, predicting and understanding subjective psychedelic experiences, personalization of treatments, clinical screening, delivery, and follow-up—which may be facilitated via digital tools such as chatbots or apps. AI can also support the application of psychological preparation, integration, and broader mental health support, as well as enhance treatment outcomes through brain-modulatory devices, including virtual reality and haptic interfaces, all while ensuring ethical and security safeguards [110–114].

In parallel, generative deep learning techniques—such as variational autoencoders and generative adversarial networks—have been successfully applied to design de novo bioactive molecules with optimized receptor profiles and pharmacological characteristics, an approach now extending to neuropsychiatric drug discovery [115]. Taken together, these AI-enabled frameworks offer a structured approach for the rational development of next-generation psychedelics or psychedelic-inspired therapeutics, that may enable enhanced precision, reduced subjective burden, and ultimately, personalized therapeutic interventions [116].

Clinical studies remain very limited, and the quest for new psychedelic-inspired compounds with minimal or null consciousness-altering effects mandates the employment of RCTs. If validated, this approach could pave the way for next-generation psychedelic-inspired therapies that retain therapeutic efficacy while eliminating the need for subjective psychedelic experiences. Such a paradigm shift could broaden the therapeutic reach of psychedelics and potentially transform the future of mental health care. To achieve this, a deeper understanding of 5-HT_2AR signaling—including biased agonism, receptor localization, and downstream pathway engagement—will be critical in guiding rational drug design

Until results from well-powered, rigorously controlled clinical trials become available, the necessity of psychedelic 5-HT_2AR activation—and the accompanying subjective experience—for psychedelic-induced antidepressant effects remains an open question. Future investigations, particularly those incorporating receptor antagonists, anesthesia, or functionally selective psychedelic analogues, are essential to unequivocally address this fundamental issue in psychedelic science and psychiatric therapeutics.

Author contributions

MP: Conceptualization, Methodology, Validation, Resources, Writing - Original Draft, Writing - Review & Editing, Supervision, Project administration; GDA: Conceptualization, Formal analysis, Writing - Review & Editing, Supervision; AI: Conceptualization, Methodology, Validation, Formal analysis, Investigation, Data Curation, Writing - Original Draft, Writing - Review & Editing, Visualization; CC: Data Curation, Writing - Review & Editing; SC: Formal analysis, Investigation, Data Curation, Writing - Original Draft, Writing - Review & Editing, Visualization; EB: Formal analysis, Investigation, Writing - Review & Editing, Visualization; ADS: Data Curation, Writing - Original Draft; LDR: Data Curation, Writing - Original Draft, Supervision, Funding acquisition; MGT: Data Curation, Writing - Original Draft; CDA: Conceptualization, Resources, Writing - Review & Editing, Supervision, Funding acquisition; GDL: Conceptualization, Resources, Writing - Review & Editing, Supervision, Project administration, Funding acquisition; RGW: Conceptualization, Resources, Writing - Review & Editing, Supervision, Project administration; SLS: Conceptualization, Resources, Writing - Review & Editing, Supervision, Project administration, Funding acquisition; FZ: Conceptualization, Methodology, Validation, Formal analysis, Investigation, Resources, Data Curation, Writing - Original Draft, Writing - Review & Editing, Visualization, Supervision, Project administration, Funding acquisition; GM: Conceptualization, Resources, Writing - Review & Editing, Supervision, Project administration, Funding acquisition.

Funding

This work was funded by the European Union – NextGenerationEU – within the Italian National Recovery and Resilience Plan – PNRR M6C2 – Investment 2.1 “Valorisation and strengthening of biomedical research in the National Health Service” (project title: Advancing PRecise Interventions for resistant DEpression: rebalancing brain networks and investigating the trajectories of antidepressant effect with non-psychedelic psilocybin and personalized neuromodulation (PRIDE); project code PNRR-MCNT2-2023-12377068 to LDR, RGW, GM, and FZ). This work was funded by the European Union – NextGenerationEU – within the Italian National Recovery and Resilience Plan, (project title: “Bipolar Disorder Integrative Staging: Incorporating the Role of Biomarkers into Progression Across Stages (BOARDING-PASS)”; project code PNRR-MAD-2022-12376693 to CDA and GM). GDL was supported by #NEXTGENERATIONEU (NGEU) and funded by the Ministry of University and Research (MUR) National Recovery and Resilience Plan (PNRR) project MNelevant findESYS (PE0000006) – (DN. 1553 11.10.2022). MP, GDA, AI, CC, SC, EB, ADS, MGT, and SLS declare no relevant funding.

Competing interests

The authors declare no competing interest.