Acute Effects of Hallucinogens on Functional Connectivity: Psilocybin and Salvinorin-A

Three male rhesus macaques (mean weight 10.6 ± 1.6 kg; mean age 7.6 ± 1.9 years) were imaged under anaesthesia and, where available, underwent repeat imaging sessions. Initial anaesthesia used intramuscular xylazine (0.5–2.0 mg/kg) and ketamine (10 mg/kg), followed by maintenance with isoflurane at approximately 1–1.2%. A 3T Siemens TIM‑Trio with a BrainPET insert and a custom 8‑channel head coil were used. A high‑resolution T1 MPRAGE was acquired, then functional imaging was performed with an EPI sequence during which ferumoxytol (MION; 10 mg/kg) was administered. Functional acquisitions used TE/TR = 22/3000 ms with 1.3 mm isotropic resolution and 100 repetitions; a final fMRI experiment was repeated for 100 min. Baseline data were collected for about 10–15 min prior to drug injection. Psilocybin and salvinorin-A doses arose from two originally separate studies but used the same imaging paradigm. Psilocybin sessions comprised 30 μg/kg (N = 2), 60 μg/kg (N = 3) and 90 μg/kg (N = 2), while salvinorin-A sessions comprised 2 μg/kg (N = 2) and 4 μg/kg (N = 5), yielding seven sessions per drug group when pooled. The authors describe drug application as pseudorandom and note variable inter-scan intervals; three instances of inter-scan intervals shorter than 50 days are specified. Ventilation was used for the salvinorin-A group as a safety precaution, creating a potential physiological confound; this, together with the pooled provenance of data, led the investigators to emphasise within‑group and within‑subject comparisons. Preprocessing included DICOM→NIfTI conversion, slice timing correction, motion correction (FSL MCFLIRT), skull stripping (AFNI 3dSkullStrip), registration to the INIA19 NHP template with ANTs, bias‑field correction and 4 mm FWHM smoothing. Ten to 15 minutes of pre‑ and post‑drug timecourses were extracted, followed by grand‑mean scaling, band‑pass filtering (0.005–0.1 Hz) and removal of linear and quadratic trends. Nuisance regression removed motion parameters and CSF/white matter signals; residuals were demeaned and scaled. Seed‑based voxelwise Pearson correlations were computed between average seed timecourses and voxels, then Fisher Z transformed (Fisher Z is a variance‑stabilising transform used before group statistics). Seeds included claustrum, thalamus, PFC and the DMN and its subcomponents (ACC, PCC, precuneus, angular gyrus); the PFC seed comprised inferior, middle and superior frontal gyri and was treated separately from the DMN. Statistical comparisons used paired t‑tests of pre‑ versus post‑drug Z‑scores with multiple‑comparison correction at p < 0.05 and cluster‑forming threshold Z > 2.3.

Physiological monitoring indicated heart rate, blood pressure and respiration remained within stable ranges under veterinary supervision, no antagonists were required, and all animals recovered after imaging. Because salvinorin-A sessions required controlled ventilation, the authors emphasised within‑group and within‑subject analyses rather than direct between‑drug contrasts. Claustrum-centred analyses showed overlapping and distinct drug effects. Both psilocybin and salvinorin-A increased functional connectivity between the claustrum and regions of the frontal cortex. Psilocybin alone increased connectivity between the claustrum and precuneus/posterior parietal regions, whereas salvinorin-A produced a reduction in connectivity between the claustrum and the right angular/occipital gyrus. Analyses with the angular gyrus as a seed revealed that both drugs tended to reduce connectivity between the angular gyrus and prefrontal cortical areas. When claustral connectivity increased to frontal clusters, the angular gyrus often showed reduced connectivity to those same regions. Psilocybin, additionally, produced strong dissociation between the angular gyrus and thalamic regions; this thalamic angular‑gyrus dissociation was not observed for salvinorin-A. Thalamic seed results differed by drug class. Psilocybin induced significant reductions in bilateral thalamic connectivity to posterior regions including the angular gyrus and portions of the posterior Task‑Positive Network (TPN). In contrast, salvinorin-A produced no significant thalamic connectivity changes in these analyses. The authors note these findings align with prior work implicating thalamo‑cortical coupling specifically with serotonergic hallucinogens. Subcomponent DMN results were informative: psilocybin reduced connectivity from the anterior cingulate cortex (ACC) to frontal cortex and caudate, and increased posterior cingulate cortex (PCC) connectivity to the precuneus. Psilocybin also increased precuneus connectivity to the caudate bilaterally and to unilateral motor clusters. Salvinorin-A appeared to dissociate anterior and posterior cingulate cortices from one another. Both substances produced notable dissociations from the angular gyrus to frontal and ACC areas. The combined pattern supports acute DMN desynchronisation with regionally specific effects; the ACC was highlighted as a potential mediator of DMN desynchrony. Statistical significance for these effects was determined via paired pre‑v‑post drug tests on Fisher Z scores with cluster correction (p < 0.05, Z > 2.3). The authors report some study limitations that affect interpretation: dosages varied across sessions and were not modelled as covariates due to sample size; ventilation for salvinorin-A sessions introduces a physiological confound that limits direct groupwise contrasts; ketamine was used for induction but is not expected to substantially alter DMN FC under the maintenance conditions used; and timing between drug challenges may influence results, although the extracted text does not clearly report the outcome of any timing‑related statistical assessment.

Bagdasarian and colleagues interpret their findings as evidence for both shared and distinct network mechanisms engaged by hallucinogens with differing pharmacology. They propose an extension of the cortico‑claustro‑cortical (CCC) model in which hallucinogenic states involve (1) increased association between subcortical‑claustral neurons and frontal cortical regions, including the anterior cingulate cortex, and (2) dissociation between the angular gyrus and frontal cortex. The observed increases in claustrum–frontal connectivity for both drugs, together with angular gyrus decoupling from frontal regions, led the investigators to suggest a common hallucinogenic axis centred on the claustrum, PFC, ACC and angular gyrus irrespective of receptor target. At the same time, the authors emphasise thalamic and DMN distinctions that appear dependent on pharmacology. Psilocybin produced clear thalamo‑cortical alterations (reductions in thalamus connectivity to posterior/TPN regions) whereas salvinorin-A did not, consistent with previous human work indicating thalamo‑cortical coupling depends on 5‑HT2A R activation. In terms of DMN specificity, both drugs induced desynchronisation but with different subcomponent signatures: psilocybin altered ACC‑to‑frontal and PCC‑precuneus connectivity, while salvinorin-A showed anterior–posterior cingulate dissociation. The authors position the cingulate cortices, particularly the ACC, as a major hub for within‑DMN changes across hallucinogens. Key limitations acknowledged by the study team include potential confounding effects of autonomic changes (for example, ventilation in the salvinorin-A group), heterogeneity of dose regimens pooled from separate studies, small sample size and the constraints of anaesthesia. These factors limit direct between‑drug comparisons and generalisability. The authors recommend future work to probe claustrum–angular gyrus–PFC relationships and to investigate thalamic neuronal signalling to clarify mechanistic differences between serotonergic and non‑serotonergic hallucinogens. They also note that the patterns identified may help define network targets relevant to clinical applications and to psychopathologies characterised by hallucinations.