The kappa opioid receptor and the sleep of reason: Cortico-subcortical imbalance following salvinorin-A

Opioid receptors (µ, δ, and κ) were identified in the 1970s and differ in distribution, ligand selectivity, and behavioural effects. Whereas µ opioid receptor agonists produce potent analgesia accompanied by euphoria and addiction risk, kappa opioid receptor (KOR) agonists have shown analgesic, antidepressant, and neuroprotective properties but are often associated with dysphoria and psychotomimetic effects. The mechanisms that underlie KOR-induced dysphoria and the hallucinatory or psychotomimetic phenomena remain unclear. Early human investigations of KOR agonists have reported vivid alterations of perception, depersonalisation, and disturbances of space and time; together these observations suggest that KOR agonism may produce a pattern of subjective and neurophysiological effects that differs from classical serotonergic psychedelics mediated by 5-HT2A receptors. Ona and colleagues set out to characterise the functional brain changes produced by full KOR agonism using salvinorin-A, a highly selective non‑nitrogenous KOR agonist. Because salvinorin-A has a very rapid onset and brief peak effect, the investigators combined two complementary sub-studies: an EEG study (sub-study 1) to capture rapid electrophysiological dynamics, and a SPECT study (sub-study 2) to capture regional cerebral blood flow (rCBF) at the drug's peak while also collecting subjective measures. The overarching aim was to link subjective phenomenology with electrophysiological and perfusion signatures of KOR activation in healthy volunteers with prior hallucinogen experience.

Both sub-studies used a double-blind, randomized, placebo-controlled within-subjects design with two experimental sessions one week apart. Volunteers were required to have previous experience with hallucinogens (at least 10 prior uses and no history of adverse effects); exclusion criteria included psychiatric disorder history, pregnancy, illness, or substance dependence. Participants underwent clinical screening including interview, physical examination, and laboratory tests. A fully psychoactive dose of 1 mg vaporised pure (>99%) salvinorin-A or placebo was administered by inhalation; volunteers practised a 30 s uninterrupted inhalation to standardise delivery. The psychoactive effects peak within approximately 1 minute and last 10–15 minutes, so procedures were tailored to capture this brief window. Sub-study 1 (EEG) recruited 24 participants (11 male, 13 female; mean age 33 years); one participant was excluded after artifact rejection. Vigilance-controlled EEG was recorded at baseline (3 min eyes closed, immediately prior to vaporisation), and at 0, +3, and +6 minutes after administration. Recordings used a 19-channel 10/20 montage referenced to averaged mastoids, sampled at 100 Hz and bandpass filtered 0.1–45 Hz; EOG channels were recorded for ocular artefact correction. A two-step artifact-removal pipeline was applied (blind source separation for ocular contamination, then automatic rejection for saturation, muscle and movement artefacts). Source estimation used standardised LORETA (sLORETA) to map frequency-band power to cortical regions; a minimum of six 5-s artifact-free epochs per time point was required. Sub-study 2 (SPECT and subjective measures) enrolled 20 participants (16 male, 4 female; mean age 35.1 years); one subject was excluded due to a technical failure that prevented psychoactive exposure. Subjective effects were measured using the Hallucinogen Rating Scale (HRS) with six subscales (somaesthesia, affect, cognition, perception, volition, intensity) and retrospective visual analogue scales (VAS) covering global intensity, liking, desire to repeat, and specific experiential domains. For perfusion imaging, 99mTc-HMPAO SPECT scans were acquired before and immediately after salvinorin-A inhalation. To capture rCBF at peak effect, the radiotracer was injected at +25 s after the end of the 30 s inhalation (i.e. +55 s from inhalation start); acquisitions lasted ~50 minutes. Images were reconstructed, normalised to MNI space, smoothed (12 mm FWHM), and analysed using SPM8. Statistical analysis used within-subject paired t-tests. Subjective measures were tested with paired t-tests and Bonferroni correction (reported thresholds: HRS .05/6 = .008; VAS .05/12 = .004). EEG LORETA power values were baseline-corrected, log-transformed, and compared voxel-wise with paired t-tests; multiple comparisons were controlled using a nonparametric permutation test estimating a maximal t distribution. SPECT rCBF differences were tested voxel-wise with proportional scaling for global uptake; only voxel-wise p < .001 and cluster-level family-wise error (FWE) corrected p < .05 (random field theory as implemented in SPM) were considered significant.

Sub-study 1 (EEG): Of 24 recruited, one participant was excluded for insufficient artifact-free data. EEG alterations were largest in the first 3 minutes after inhalation. Source-estimated changes during the first 3 minutes after salvinorin-A versus placebo showed significant increases in delta-band power localised to temporal regions and in gamma-band power localised to occipital (visual) areas. Alpha-band power decreased significantly and was localised to the cingulate gyrus, precuneus, and superior parietal lobe. Sub-study 2 (subjective measures and SPECT): Of 20 recruited, one was excluded due to a failed administration. On the HRS, all six subscales showed significantly higher scores after salvinorin-A than placebo; the largest increases were observed for intensity and volition, with notable rises also for cognition and perception. VAS ratings were also significantly elevated across all items after salvinorin-A; the highest VAS scores were for "I liked the experience," "I would like to take the substance again," "good effects," and "changes in dimensionality." SPECT revealed a pronounced, widespread decrease in rCBF across cortical regions, most prominently in frontal, temporal and parietal lobes, with smaller decreases in parts of the occipital lobe including the calcarine sulcus. In contrast, significant post-salvinorin-A increases in rCBF were observed in the cerebellum and medial temporal structures bilaterally, especially the amygdala, hippocampus, and parahippocampal gyrus. These SPECT findings met the voxel-wise and cluster-level statistical thresholds described in the Methods. The extracted text does not provide numeric effect sizes, confidence intervals, or exact cluster coordinates in the prose.

Ona and colleagues interpret their findings as evidence that full KOR agonism via salvinorin-A produces a distinctive cortico‑subcortical imbalance characterised by widespread cortical hypoperfusion together with limbic and cerebellar hyperperfusion, and concomitant electrophysiological changes. Subjectively, the drug produced intense experiences with marked loss of contact with self and surroundings (high intensity and volition scores), alongside relatively low affective content; VAS ratings suggested that participants overall found the experience positive and would repeat it. The investigators note that this subjective pattern differs from typical profiles reported for psilocybin or ketamine but shows some similarities to reports with DMT and ayahuasca. Electrophysiologically, the observed increase in delta power (temporal, parietal, parahippocampal regions) is linked by the authors to a dream-like state and loss of bodily awareness, whereas the alpha decrease in parietal and temporal cortices may reflect reduced inhibitory control during the potent acute effects. Enhanced gamma activity in visual areas is interpreted as reflecting engagement of visual processing during eyes-closed imagery. The authors highlight that these EEG signatures differ from some reports of serotonergic hallucinogens, supporting the view that distinct neuropharmacological mechanisms underlie KOR-induced psychotomimetic phenomena. Regarding perfusion, the authors emphasise the contrast with many serotonergic psychedelic studies that report increased frontal perfusion or DMN alterations; here, salvinorin-A produced dramatic cortical decreases, notably in left prefrontal, superior frontal and premotor cortices—regions with relatively high KOR density—and concomitant increases in medial temporal structures including the amygdala. The amygdala hyperperfusion is suggested to reflect heightened emotional salience or differential limbic dynamics compared with serotonergic psychedelics, which often show amygdala attenuation. The pattern of precuneus hypoperfusion is noted alongside its recognised role in consciousness and the default mode network, and may relate to parallels with changes seen in different states of consciousness such as sleep. The authors acknowledge several limitations: small sample sizes in both sub-studies, recruitment of participants with prior hallucinogen experience which may have reduced dysphoric responses and limits generalisability, and psychometric constraints of the HRS (authors advise treating its scales as approximations). They further note that SPECT was chosen for its ability to 'fix' rCBF at the drug peak, but that complementary neuroimaging modalities could help clarify mechanisms. Finally, the investigators suggest broader implications: KOR signalling may be relevant to disorders of perception and cognition (for example, schizophrenia), and KOR agonists appear to act as a switch interrupting cortical processing—an effect the authors metaphorically liken to Goya's "the sleep of reason produces monsters."