Acute effects of LSD on amygdala activity during processing of fearful stimuli in healthy subjects

Mueller and colleagues situate their work in a context of renewed scientific interest in classical psychedelics, noting that LSD produces profound changes in perception, self-awareness and emotion that are largely mediated by agonism at serotonin 5-HT2A receptors. Previous modern neuroimaging studies have focused mainly on psilocybin, and clinical research has explored psychedelics for conditions such as depression, anxiety and addiction; however, data on the neural correlates of LSD in humans—particularly during emotional processing—were limited or absent at the time of this study. To address this gap, the investigators tested the acute effects of a single oral 100 μg dose of LSD on brain responses to fearful versus neutral faces using functional magnetic resonance imaging (fMRI). They preregistered an a priori hypothesis that LSD would reduce amygdala reactivity to fearful stimuli and that this reduction would be related to the subjective psychedelic experience; the study used a double-blind, randomised, cross-over design in healthy, mostly psychedelic‑naive volunteers to examine this question.

The study employed a randomised, placebo-controlled, double-blind, cross-over design. Each participant completed two sessions (LSD and placebo) separated by a washout of at least 7 days. Capsules contained either 100 μg D-lysergic acid diethylamide hydrate or mannitol; only the person dispensing the treatment knew allocation, while participants and study personnel remained blind to order. The trial was approved by the relevant ethics committee and registered prospectively (ClinicalTrials.gov NCT02308969). Twenty healthy volunteers were included in the final analysis (originally 24 recruited; four were excluded for head-motion artefacts). Inclusion and exclusion criteria were comprehensive and aimed to enrol physically and mentally healthy adults with minimal lifetime exposure to illicit drugs; the extracted text reports that only two participants had prior single-occasion psychedelic experience. Participants attended a screening visit, two 25 h test sessions and an end‑of‑study visit; they were instructed to abstain from illicit drugs during the study and from caffeine, chocolate and alcohol for at least 8 h before sessions. Plasma samples for LSD concentration were taken approximately 2 and 3 h after administration. fMRI acquisition used a 3 Tesla scanner with standard echo-planar imaging (3 × 3 × 3 mm voxels, repetition time 2.5 s), collecting 152 volumes per run. The emotional task was a 6 min event-related paradigm presenting 30 unique face stimuli (Ekman faces) at 50% or 100% fear intensity or neutral expression, each shown twice for 2 s (60 trials total). Interstimulus intervals varied 3–8 s (mean 5.9 s). To maintain attention, participants performed a gender discrimination button-press task; accuracy and reaction time were recorded. Preprocessing and statistical analysis were conducted in SPM12: slice-timing correction, realignment, coregistration to a T1 structural image, normalisation to MNI space and smoothing with a 6 mm FWHM kernel. The model included motion parameters and temporal derivatives; contrasts compared 50% and 100% fearful faces against neutral faces. Whole-brain testing used an initial threshold of P = 0.001 uncorrected with k = 10 voxels, and a priori regions of interest (amygdala, fusiform gyrus, medial frontal gyrus) were examined using small volume correction (SVC) based on the Harvard-Oxford atlas with a 5 mm sphere; SVC results were reported at P < 0.05 familywise error (FWE) corrected. The extracted text contains an unclear numerical value for the head-motion exclusion threshold (the motion cut‑off as printed is not clearly reported), and the authors excluded subjects with excessive motion accordingly. Correlation analyses between subjective drug effects (visual analogue scale rated 2 h after administration) and extracted amygdala beta values used SPSS.

Twenty data sets entered the analyses (9 men, 11 women; mean age 32 ± 10.2 years, range 25–58), all right‑handed and all but one with an academic background. Four participants were excluded because of head-motion artefacts. No participant tested positive for drugs at screening or sessions, and no serious adverse events occurred. Measured plasma LSD concentrations immediately before and after the scan averaged 1.3 ± 0.6 ng ml-1 and 1.1 ± 0.5 ng ml-1, respectively (mean ± s.d.). Behavioural performance on the gender discrimination task did not differ significantly between conditions. Mean reaction times were 964 ± 128 ms under LSD versus 910 ± 289 ms under placebo (t21 = 2.0, P = 0.06). Correct responses were 93.1 ± 10.8% under LSD and 97.3 ± 3.3% under placebo (t21 = -1.8, P = 0.08); omission rates also did not differ significantly. Across both treatments, viewing neutral faces versus baseline activated a network including the fusiform and occipital gyri and cingulate regions (FWE-corrected P < 0.05). Under placebo, fearful versus neutral faces elicited significant activation of the left amygdala (MNI x = -20, y = -12, z = -12; cluster size 22; Z = 3.59; SVC P < 0.05 FWE at cluster level). Comparing LSD with placebo, LSD reduced neural responses to fearful versus neutral faces in amygdala regions and the medial frontal gyrus at the uncorrected whole-brain threshold (P < 0.001, k = 10). After small-volume correction, significantly reduced activity remained in the left amygdala (MNI x = -15, y = 9, z = -14; cluster size 24; Z = 3.12; SVC P < 0.05 FWE) and in the right medial frontal gyrus (MNI x = 15, y = 42, z = 16; cluster size 12; Z = 3.78; SVC P < 0.05 FWE). No brain regions showed greater activity under LSD than placebo for the fearful versus neutral contrast. Importantly, amygdala blood oxygen-level dependent response to fearful stimuli under LSD correlated negatively with the subjective drug effect on the visual analogue scale (r = -0.46, P < 0.05), indicating that stronger subjective effects were associated with greater amygdala attenuation.

Mueller and colleagues interpret their findings as evidence that acute administration of 100 μg LSD attenuates amygdala reactivity to fearful faces in healthy volunteers, and that this attenuation is related to the intensity of the subjective psychedelic experience. They note that this is, to their knowledge, the first fMRI study to examine LSD effects on emotional processing and that the chosen dose reliably produces typical psychedelic effects while participants were largely psychedelic‑naive—conditions the authors argue are relevant to clinical applications of LSD-assisted psychotherapy. The authors place their results in the context of previous work with psilocybin, which similarly showed reduced recognition of negative facial expressions and decreased amygdala responses to fearful faces; they also draw parallels with findings that selective serotonin reuptake inhibitors (SSRIs) can reduce amygdala reactivity to negative stimuli. Mechanistically, they suggest that LSD’s agonism at 5-HT2A receptors—receptors expressed in the amygdala—provides a plausible pharmacological basis for the observed effects. The observed decrease in right medial prefrontal cortex (mPFC) activity is discussed in terms of mPFC–amygdala circuitry: ventral mPFC regions have been implicated in inhibition of amygdala responses, while dorsal regions may participate in aversive amplification. In clinical terms, the investigators propose that reducing processing bias toward negative stimuli could be therapeutically relevant for disorders marked by such biases (for example, depression and social anxiety disorder). They also link the amygdala attenuation to previously reported 'empathogenic' effects of LSD (increased openness and trust) and suggest this neural change might facilitate psychotherapeutic processes such as alliance building and reduced perception of negative emotions. The authors emphasise, however, that psychedelic-induced longer-term benefits reported in other studies outlast acute pharmacology and that further work is needed to understand contributions of psychological and biological mechanisms such as neuroplasticity. The paper acknowledges key limitations: blinding was imperfect because the subjective effects of the dose made treatment assignment effectively apparent, measures of negative affect were not collected and only a single moderate dose was tested. The authors also note practical constraints of higher doses in fMRI settings due to potential anxiety, and they caution that the observed anxiolytic-like effect may be context- and population-dependent and therefore could differ in clinical or uncontrolled environments.