Psychedelics produce enduring behavioral effects and functional plasticity through mechanisms independent of structural plasticity

Psychedelics such as psilocybin are reported to produce durable antidepressant effects in humans after only one or two doses, but the cellular and molecular mechanisms that sustain effects for months remain unclear. Previous preclinical work has shown that psychedelics can trigger changes in gene expression and promote synaptic growth, including transient increases in dendritic spine formation in cortex within hours or days after dosing. However, most animal studies examine relatively short post‑treatment intervals (hours to weeks), leaving open whether structural changes persist on the time scale of human clinical effects and whether 5‑HT2A receptor activation alone is necessary or sufficient for long‑lasting action. M. and colleagues set out to address these gaps using the Wistar Kyoto (WKY) rat, a stress‑vulnerable strain used to model depressive‑like behaviour. The study compared a single intraperitoneal dose of psilocybin and the selective 5‑HT2A agonist 25CN‑NBOH, assessing behavioural persistence in the forced swim test (FST) at 5 and 12 weeks post‑dose, functional cellular plasticity in infralimbic medial prefrontal cortex (mPFC) Layer 5 neurons using whole‑cell electrophysiology ~100 days after dosing, and indicators of structural plasticity via spine counting and qRT‑PCR for synaptic transcripts. The design aimed to determine whether enduring behavioural effects are accompanied by sustained functional or structural changes and whether 5‑HT2A activation alone can account for long‑term effects.

Animals and drugs: Male Wistar Kyoto rats (postnatal days 50–55) were randomly assigned to three treatment groups (saline, psilocybin, 25CN‑NBOH) with n = 8 animals per group. Female rats were excluded. Rats were pair‑housed under a 12:12 light–dark cycle with ad libitum food and water. Psilocybin (1.0 mg/kg) and 25CN‑NBOH (1.5 mg/kg) were dissolved in sterile saline and administered intraperitoneally in a volume of 1 mL/kg on day 0. Dose selection referenced prior literature and the authors’ previous work. Behavioural testing: The primary behavioural endpoint was the forced swim test (FST), used here as a measure of passive coping (immobility) that the authors interpret as an antidepressant‑like behavioural readout. Rats were pre‑exposed to a 15‑minute swim session one day before testing. FST testing consisted of a 5‑minute recorded swim scored for immobility, swimming and climbing by blinded scorers using a modified sampling technique. Locomotor activity (LCA) in an open field (61 cm square arena) was recorded for 5 minutes immediately prior to each FST to control for sedative or stimulant effects; distance travelled was analysed with EthoVision tracking software. Behavioural assessments occurred at 35 days (5 weeks) and 84 days (12 weeks) after a single injection. Immobility was analysed with repeated measures ANOVA (treatment × time) with Holm–Sidak post hoc tests; LCA was analysed with mixed‑effects REML due to some missing data caused by equipment failure. Electrophysiology and tissue processing: Approximately 100 days after drug administration, animals were perfused and brains hemisected; one hemisphere was flash‑frozen for molecular assays and the other used for acute slice electrophysiology. Coronal slices containing the infralimbic mPFC were prepared using NMDG‑based solutions and subsequently held in protective solution before recording. Whole‑cell recordings were obtained from Layer 5 pyramidal neurons in the infralimbic mPFC at 30–32 °C using K‑gluconate internal solution with 0.2% biocytin for post hoc visualisation. Measures included resting membrane potential (RMP), input resistance, rheobase (minimum current to elicit spikes), suprathreshold firing rate, and spontaneous excitatory postsynaptic current (sEPSC) rate and amplitude. Neurons were classified by firing pattern into adapting and bursting types. Electrophysiological analysis: The authors performed principal component analysis (PCA) on a normalised set of intrinsic and synaptic parameters to detect coordinated shifts across measures. For individual metrics, one‑way ANOVAs were conducted separately for adapting and bursting neurons across treatment groups, with Tukey’s HSD post hoc comparisons following significant main effects. Parameters were normalised to unit variance prior to PCA so variables on different scales contributed equally. Imaging, spine counting and gene expression: After recordings, slices were fixed, stained with streptavidin‑568 to reveal biocytin‑filled neurons, and imaged on a confocal microscope (100×, z‑step 0.06 μm). Dendritic spine analyses were performed on three‑dimensional projections using ImageJ; spines were classified as mushroom, stubby or thin using explicit size/shape criteria. Spine counts were averaged from duplicate counts; a total of n = 13, 17 and 8 neurons were scored for saline, psilocybin and 25CN‑NBOH groups respectively, with an average of 115 spines scored per neuron across groups. For gene expression, the left hemisphere infralimbic mPFC was dissected from flash‑frozen tissue, RNA extracted, cDNA synthesised and qRT‑PCR performed in duplicate. ΔΔCt normalisation used ywhaz as reference, and one‑way ANOVA with Tukey post hoc tests compared expression across groups (n = 8 per treatment). Statistical analyses were performed in GraphPad Prism and MATLAB for electrophysiology.

Behavioural outcomes: Following a single administration of drug, both psilocybin and 25CN‑NBOH produced persistent reductions in immobility in the FST at 5 weeks (35 days) and 12 weeks (84 days) post‑dose. Repeated measures ANOVA identified treatment (F(2,21) = 20.28, p < 0.0001) and subject (F(21,21) = 3.068, p = 0.0066) as significant sources of variation for immobility; time was not significant and no treatment×time interaction was reported, indicating effects were maintained between 5 and 12 weeks. Psilocybin reduced immobility versus saline at 5 weeks (p = 0.0077) and 12 weeks (p = 0.0032). 25CN‑NBOH reduced immobility versus saline at 5 weeks and 12 weeks (both p < 0.0001) and produced significantly lower immobility than psilocybin at both time points (5 weeks p = 0.0290; 12 weeks p = 0.0257). Swimming behaviour paralleled immobility results: treatment (F(2,21) = 22.92, p < 0.0001) and subject (F(21,21) = 2.753, p = 0.0123) were significant, with both drugs increasing swimming versus saline (psilocybin p ≈ 0.002; 25CN‑NBOH p < 0.0001) and 25CN‑NBOH exceeding psilocybin (p values 0.0408 and 0.0249 at 5 and 12 weeks respectively). No significant differences were observed in climbing behaviour or in overall locomotor activity (LCA), and no within‑group changes occurred between 5 and 12 weeks, indicating preserved effect size. Four LCA videos from Week 5 were excluded due to equipment failure. Electrophysiology and functional plasticity: Approximately 100 days after dosing, Layer 5 infralimbic‑mPFC neurons were recorded and classified as adapting or bursting. In adapting neurons, psilocybin produced a significant depolarisation of RMP (main effect F(1,32) = 6.97, p = 0.0106), with post hoc tests showing psilocybin versus saline p = 0.0126. No treatment effects were observed on input resistance, rheobase, firing rate, or sEPSC amplitude/frequency in adapting neurons. The authors note only four regular spiking adapting neurons were recorded from the 25CN‑NBOH group, limiting conclusions for that subclass. Bursting neurons showed broader changes: RMP differed across treatments (F(1,73) = 3.41, p = 0.0384) with a significant difference between psilocybin and 25CN‑NBOH (p = 0.0298). Rheobase was reduced by treatment (F(1,74) = 4.73, p = 0.0117), with psilocybin showing lower rheobase than control (p = 0.0101). Firing rate increased with psilocybin (F(1,74) = 5.74, p = 0.0048; psilocybin vs control p = 0.0033). Synaptic measures in bursting neurons revealed treatment effects on sEPSC amplitude (F(1,67) = 4.58, p = 0.0137) and sEPSC frequency (F(1,67) = 3.6, p = 0.0329), with post hoc comparisons indicating significant differences between psilocybin and 25CN‑NBOH (amplitude p = 0.0098; frequency p = 0.0455). These univariate findings indicate that psilocybin enhanced excitability in bursting neurons relative to controls and relative to 25CN‑NBOH in some measures. Multivariate electrophysiology (PCA): PCA on normalized electrophysiological parameters revealed coordinated, cell‑type specific shifts. In adapting neurons, PC1 (driven by RMP and input resistance) was significantly reduced in both psilocybin and 25CN‑NBOH groups versus controls (F(1,28) = 7.47, p = 0.0027; psilocybin vs control p = 0.0216; 25CN‑NBOH vs control p = 0.0054). In bursting neurons, PC2 (driven by rheobase and firing rate) was significantly increased in the psilocybin group (F(1,66) = 8.39, p = 0.0006), with psilocybin differing from saline (p = 0.0422) and from 25CN‑NBOH (p = 0.0004). The PCA results are presented by the authors as evidence of coordinated functional plasticity that differs by cell type and by drug. Structural markers and gene expression: Confocal imaging of biocytin‑filled neurons and subsequent spine classification found no differences in spine density or in counts of spine subtypes (mushroom, stubby, thin) between treatment groups at 12 weeks post‑dose. qRT‑PCR of several synaptic protein transcripts in the infralimbic mPFC also showed no significant differences across groups (one‑way ANOVA with Tukey post hoc). For spine analysis, the authors report scoring 13, 17 and 8 neurons in the saline, psilocybin and 25CN‑NBOH groups respectively, with ~115 spines averaged per neuron.

The authors interpret their results as showing that a single administration of either psilocybin or the selective 5‑HT2A agonist 25CN‑NBOH produces antidepressant‑like behavioural effects in WKY rats that persist for at least 12 weeks, and that these persistent behavioural changes are accompanied by enduring functional plasticity in infralimbic mPFC Layer 5 neurons measured approximately 100 days after dosing. They emphasise that 25CN‑NBOH produced long‑lasting behavioural and functional effects despite being selective for 5‑HT2A receptors, supporting the idea that 5‑HT2A activation is sufficient to elicit persistent changes; at the same time, the authors note this does not exclude contributions from other serotonin receptors engaged by psilocybin and suggest further work using receptor‑selective antagonists. M. and colleagues highlight a dissociation between functional and structural measures at the long timepoint: whereas electrophysiology revealed cell‑type specific increases in excitability (notably psilocybin‑related depolarisation in adapting neurons and enhanced excitability in bursting neurons), spine density and the expression of several synaptic transcripts in the infralimbic mPFC were indistinguishable from controls at 12 weeks. They situate these findings within prior literature showing rapid, short‑term increases in spine formation after psychedelics and other rapid‑acting antidepressants, and propose a model in which an early transient structural growth phase (hours to weeks) is followed by synaptic pruning that returns spine numbers toward baseline by months, while functional changes in neuronal excitability persist. The authors cite analogous observations in ketamine studies where early spine increases are followed by elevated spine elimination and a return toward baseline. Acknowledged uncertainties and limitations reported by the authors include the exclusion of female rats (they state insufficient behavioural literature to justify including females for this model), limited sampling of certain neuron subclasses (only four regular‑spiking adapting neurons recorded from the 25CN‑NBOH group), and the absence of direct mechanistic data linking the observed functional plasticity to particular molecular mechanisms. The researchers note they found no correlations between spine density, FST performance, and electrophysiological properties in their data. They speculate that enduring functional plasticity may arise from long‑lived molecular changes such as epigenetic modulation of ion channel expression that increase the efficacy of remaining synapses, but state this remains to be tested. The authors recommend further experiments to define precise mechanisms, validate receptor contributions using antagonists, and clarify the causal relationship between identified neuronal cell‑type changes and behavioural outcomes.