Serotonergic hallucinogens and recognition of facial emotion expressions: a systematic review of the literature.

Recognition of emotions in facial expressions (REFE) is a core component of social cognition and is impaired in several psychiatric conditions, notably anxiety and mood disorders. Earlier research has shown that deficits in REFE are linked to clinical features—social anxiety disorder involves hypervigilance to threatening expressions, for example—and that some anxiolytic and antidepressant treatments produce changes in emotional processing that may precede clinical improvement. MDMA has been robustly shown to reduce identification of negative emotions and to modulate amygdala and prefrontal activity, and recent preclinical and clinical work has suggested that classical serotonergic hallucinogens (LSD, psilocybin, ayahuasca/DMT, mescaline) have anxiolytic and antidepressant properties mediated via 5-HT2A receptor agonism in brain regions implicated in emotion processing. Rocha and colleagues set out to systematically review human experimental studies that directly examined the effects of serotonergic hallucinogens on REFE. The stated goal was to determine whether these compounds alter recognition of facial emotions and to summarise neurophysiological and clinical correlates, with attention to potential relevance for treating disorders characterised by REFE deficits such as depression and anxiety.

The researchers conducted a systematic review according to PRISMA guidance. Electronic searches of PubMed, PsycINFO, and Web of Science were performed using terms combining candidate psychedelics (ayahuasca OR DMT OR psilocybin OR LSD OR mescaline) with face/emotion recognition search strings. The search included all publications up to 19 October 2018 without language restriction. Inclusion criteria required human experimental studies administering a serotonergic hallucinogen and assessing recognition of facial expressions; exclusion criteria included animal studies, non-hallucinogen interventions, studies that did not assess facial expression recognition, reviews, observational studies, case reports, and letters. Two reviewers independently screened titles/abstracts and full texts, resolving disagreements by discussion or a third reviewer when necessary. Extracted variables included authors, year, country, study design (open-label or placebo-controlled), sample size, drug and dosing, facial recognition task and measured variables (accuracy, reaction time, neurophysiology, neuroimaging), and main subjective/neuroimaging findings. The initial search yielded 62 records; after removing 22 duplicates 40 articles were screened and seven met the inclusion criteria (one further article was excluded on full-text review because it used a different social cognition test). Quantitative synthesis (meta-analysis) was not attempted due to heterogeneity in drugs, doses, sample characteristics, and REFE tasks, so the authors conducted a qualitative synthesis. The seven included studies comprised five double-blind, randomized, crossover, placebo-controlled trials and two open-label studies. All identified experimental work involved either psilocybin or LSD; no included experimental REFE studies were found for ayahuasca/DMT or mescaline. Reported samples ranged from small healthy volunteer cohorts (roughly 20–40) to two open-label samples of patients with treatment-resistant depression (n≈17–20).

Overall finding across the included studies was a tendency for single or a few doses of psilocybin or LSD to reduce recognition of negative facial emotions and to modulate amygdala responses; both drugs were reported as well tolerated. Psilocybin studies: Schmidt and colleagues (2013) performed a randomized, double-blind, placebo-controlled crossover study in healthy volunteers (two groups of n=21) comparing oral psilocybin (115 μg/kg) and intravenous ketamine with placebo while measuring event-related potentials during a static facial affect discrimination task. Both psilocybin and ketamine significantly reduced fear recognition; ketamine additionally reduced recognition of happiness. Both drugs reduced the N170 evoked potential associated with face/emotion processing. Bernasconi and colleagues used a randomized, double-blind, placebo-controlled crossover design (n=30) with oral psilocybin (170 μg/kg) and EEG to measure evoked responses during a static facial emotion recognition task; psilocybin reduced neural responses to fear and neutral faces in amygdala and parahippocampal gyrus at 168–189 ms and reduced responses to happy faces in limbic and temporo‑occipital areas at 211–242 ms. Roseman and colleagues conducted an open-label study in 20 patients with treatment‑resistant depression who received two doses of psilocybin (10 mg then 25 mg one week apart) with fMRI during a passive facial task; they observed increased right amygdala response to fearful and happy faces and reported that altered recognition of fear and neutral faces predicted clinical improvement. In a subsequent comparison (same patient sample versus 16 healthy controls) using a dynamic emotion task, depressed patients showed slower reaction times than controls at baseline, but reaction times in patients significantly decreased after psilocybin; the magnitude of reaction‑time reduction correlated positively with reduction in anhedonia, and depressive symptoms were reduced one week after the second dose. Grimm and colleagues (n=18 healthy) used oral psilocybin (0.16 mg/kg) in a randomized, double‑blind, placebo‑controlled crossover fMRI study with a static task and found increased reaction time for recognition of fear, happiness and anger without changes in accuracy, reduced functional connectivity between right amygdala and left striatum for anger, and reduced connectivity between right amygdala and right medial frontal cortex for happy faces; reduced amygdala–medial frontal connectivity correlated with reductions in anxiety and depression scores. LSD studies: Dolder and colleagues reported pooled data from two randomized, double‑blind, placebo‑controlled crossover studies in healthy volunteers (100 μg, n=24; 200 μg, n=16). Using a static FERT, both doses significantly reduced recognition of fearful expressions and increased emotional empathy and prosociality without affecting cognitive empathy; both doses were well tolerated. Mueller and colleagues analysed a subsample (n=20) and, using fMRI during a static fear task, found LSD reduced activation of the left amygdala and right medial prefrontal cortex during fear presentation; the reduction in amygdala activation correlated negatively with subjective effects of LSD. Safety and tolerability: Across studies, single or few doses were generally well tolerated with few adverse reactions reported. Heterogeneity: differences across studies in task type (static, dynamic, passive), stimuli databases, emotion sets presented, outcome measures (accuracy, reaction time, ERP, BOLD, connectivity), doses and populations precluded meta-analysis and complicated direct comparison. Authors emphasise that modulation of amygdala activity was a recurring neurobiological correlate of the behavioural effects; however, the association between amygdala changes and clinical improvement was observed in only one open‑label depression study.

Rocha and colleagues interpreted the body of evidence as indicating that classical serotonergic hallucinogens—psilocybin and LSD in the available literature—tend to reduce recognition of negative facial emotions and modulate amygdala activity during emotional processing. The investigators noted this pattern was observed across different paradigms, tasks and measures despite methodological heterogeneity. Several studies linked the observed changes in emotional processing with short‑term improvements in mood or anxiety measures, but the clinical relevance of these associations is uncertain because the key patient finding arose from a single open‑label study. The authors situate these results in the context of earlier findings with traditional antidepressants, anxiolytics and MDMA: for example, brief SSRI treatment has been reported to reduce recognition of negative expressions, benzodiazepines reduce identification of negative emotions, and MDMA consistently reduces identification of negative emotions alongside amygdala attenuation. They suggest the shared outcome—altered REFE—may represent a convergent mechanism contributing to therapeutic effects, even though pharmacology differs. Classic hallucinogens are partial agonists at cortical 5‑HT2A receptors and also act at 5‑HT1A and 5‑HT2C receptors; the authors propose receptor effects and increased neuroplasticity as candidate mechanisms. Important limitations acknowledged include small sample sizes, variable use of placebo or adequate control groups, and heterogeneity across tasks, stimuli and dosing regimens. The reviewers also highlight the need to clarify whether neural effects differ between healthy volunteers and patients with mood/anxiety disorders, noting some evidence for differential activation patterns. To advance mechanistic understanding they recommend experimental studies combining psychedelics or SSRIs with a selective 5‑HT2A antagonist in an emotional processing paradigm, and further functional neuroimaging. Finally, the authors stress that while tolerability in acute experimental contexts was generally good, larger and longer controlled trials in clinical populations are needed to determine therapeutic potential and durability.

The systematic review concludes that available experimental evidence suggests serotonergic hallucinogens (psilocybin and LSD) modify recognition of facial emotions by reducing recognition of negative expressions and modulating amygdala function, changes that could be therapeutically useful in disorders characterised by REFE deficits such as anxiety and depression. The authors emphasise that these findings require replication in larger, placebo‑controlled trials with clinical populations, and that mechanistic studies—particularly those using 5‑HT2A antagonists and functional neuroimaging—are needed to clarify causality and neural substrates. Major limitations of the review include the small number of studies, small sample sizes, and heterogeneity of tasks and controls; despite these caveats the evidence supports further controlled research into the effects of serotonergic hallucinogens on social‑cognitive processes and their relevance to mood and anxiety disorders.