Long-term ayahuasca use is associated with preserved global cognitive function and improved memory: a cross-sectional study with ritual users

Ayahuasca is a traditional South American decoction prepared from Banisteriopsis caapi and Psychotria viridis, the former supplying reversible MAO-A–inhibiting ß-carbolines (harmine, harmaline, tetrahydroharmine) and the latter containing the serotonergic hallucinogen N,N-dimethyltryptamine (DMT). Previous preclinical and human work has reported neuroplastic and possible cognitive effects of these alkaloids, and observational studies of ritual ayahuasca users have generally not found cognitive impairment; some studies have even reported better performance on measures of working memory, verbal learning and executive function. However, evidence is mixed, sample sizes are often small, and few studies have examined the effects of very long-term ritual use compared with more recent use within the same religious setting. Martins Fonseca and colleagues set out to evaluate cognitive performance in members of a Santo Daime church (Centro Luz Divina, CLD) by comparing experienced ritual ayahuasca users (>20 years of regular use), beginner users (<3 years), and a matched non-user control group. The primary aim was to determine whether long-term ritualistic ayahuasca consumption is associated with preserved or altered global cognition and specific domains such as verbal and visuospatial working memory, executive function and immediate visual memory; the investigators also compared results to available normative data and confirmed botanical identity and alkaloid presence in the plants and brews used at the site.

This was an observational, descriptive, cross-sectional study carried out between July 2018 and January 2019. Participants were recruited by convenience from the CLD Santo Daime church in Piedade, São Paulo, Brazil. The sample comprised 48 individuals divided into three groups of 16: experienced users (mean duration ~26 years 7 months), beginner users (mean duration ~1 year 3 months), and a non-user control group matched to the study groups by sex, age and education. Inclusion criteria required participants to be over 18 and members of CLD; exclusion criteria included any diagnosed mental disorder. All participants provided informed consent and plant collection and analyses were authorised by relevant biodiversity authorities. Plant material (P. viridis leaves and B. caapi vine) was collected in July 2018 and used to prepare two traditional recipes (sample 1 and sample 2, the latter prepared using sample 1 to obtain higher alkaloid concentration). Taxonomic identification was performed and voucher specimens deposited in a university herbarium. Laboratory extracts were prepared via hot aqueous Soxhlet extraction, lyophilised, and chemical analyses quantified alkaloids; detailed chemical methods were provided in supplemental material. Neuropsychological assessment was administered individually by one of the authors (initials AMF) in a private room at the church for adherents and in the researcher's office for controls, with testing performed on average eight days after the last ayahuasca ingestion. The battery included the Wechsler Abbreviated Scale of Intelligence (WASI; FSIQ, VIQ, PIQ and subtests), Digit Span (DS; forward and backward measures of verbal working memory), Corsi Block-Tapping Task (CBT; visuospatial working memory), Rey–Osterrieth Complex Figure (ROCF; visual perception and immediate memory), Wisconsin Card Sorting Test-64 (WCST-64; executive function) and Five Digit Test (FDT; processing speed and inhibitory control). Sociodemographic, lifetime and recent substance-use data were collected via semi-structured interview. Statistical analyses used SPSS 21.0. Normality was assessed with Kolmogorov–Smirnov and Shapiro–Wilk tests; because distributions were normal, multivariate analysis of variance (MANOVA) was applied with neuropsychological test scores as dependent variables and group, sex and their interactions as independent variables. Homogeneity of variances and covariance matrices were checked with Levene's and Box tests. Post-hoc comparisons used Bonferroni or Tukey’s b adjustments as appropriate. Categorical associations used chi-square tests. Statistical significance was set at p ≤ 0.05.

The sample comprised 48 participants (three groups of 16; each group 8 women and 8 men). Groups were homogeneous on sociodemographic measures except for duration of ayahuasca use (p = 0.001): experienced users averaged 26 years 7 months of regular consumption, beginners averaged 1 year 3 months. Tobacco and alcohol use patterns varied descriptively (the experienced group had the highest proportion of smokers, while the control and beginner groups showed higher or more recent alcohol use), but these differences were not reported as statistically significant beyond the time-of-use contrast. Botanical identification confirmed Banisteriopsis caapi and Psychotria viridis, and alkaloid quantification was performed (numerical results referenced in a table in the paper but not fully reproduced in the extracted text). On global intellectual measures (WASI), the beginner group scored significantly lower than both the experienced and control groups on full-scale and subscale measures (effect sizes n2 = 0.164–0.289), whereas the experienced and control groups performed similarly and were classified within the "average" intellectual range. For executive function, no significant group differences emerged on the WCST-64; however, the FDT reading time was slower in the beginner group compared with controls (large effect size n2 = 0.159). Immediate visual memory (ROCF copy total score) was significantly lower in beginners than in the other two groups (large effect size n2 = 0.314). Percentile classification showed 50% of controls and 31.2% of experienced users at or above the 50th percentile for copy performance, while 87.5% of beginners fell between the 10th and 20th percentiles. Measures of working memory showed the most pronounced differences favouring experienced users. In Digit Span, the experienced group had a higher weighted mean score (13.00) than beginners (9.62) with a large effect size (n2 = 0.227). Raw Forward Digit Span averaged 9.9 for experienced users versus 7.7 for beginners, compared to published Brazilian norms of approximately 6.8–6.9; Backward Digit Span averaged 5.8 in experienced users versus 3.6 in beginners (normative value ~4.8). The difference between forward and backward spans exceeded 3 digits in 87.5% of experienced participants. In the Corsi Block-Tapping Task, experienced users outperformed beginners on the sum of forward and backward trials and on the total score, with performance comparable to published normative samples. The authors reported no evidence of general cognitive decline among ayahuasca users; rather, long-term ritual users displayed superior scores on verbal and visuospatial working-memory tasks compared with beginners, while executive function and immediate memory measures were largely preserved and comparable to controls. The extracted text also notes study limitations including small sample size, absence of detailed illicit drug-use histories, potential sampling biases, and the cross-sectional design which precludes causal inference.

Martins Fonseca and colleagues interpret the findings as indicating global preservation of cognitive function in ritual ayahuasca users and specific enhancements in working memory among long-term users relative to recent initiates. They stress that the experienced group performed comparably to matched controls on most measures of intellectual ability and executive function, and exceeded beginners on both verbal (Digit Span) and visuospatial (Corsi) working-memory tasks. The authors situate these results alongside prior reports that found either preserved cognition or selective improvements among ayahuasca users, noting that previous structural brain findings (for example, changes in cingulate cortex thickness) have not been directly linked to cognitive performance. Possible mechanisms advanced by the investigators include neuroplastic and neuroprotective effects of ayahuasca alkaloids (DMT and ß-carbolines), serotonergic 5-HT2A receptor involvement in spatial and working-memory processes, increased expression of immediate-early genes and brain-derived neurotrophic factor in preclinical models, and ritual-specific cognitive training (for example memorisation of hymns) that might strengthen memory-related skills in long-term adherents. The authors also propose that the ritual setting, set and intention may reduce adverse events and influence cognitive outcomes. Key limitations acknowledged by the study team are the small, convenience sample, lack of detailed information on other illicit drug use, potential sampling and selection biases inherent to studying ritual users within a single church, and the cross-sectional design that prevents causal attribution of cognitive differences to ayahuasca exposure. They recommend future work with larger samples, more thorough substance-use histories, non-ritual cohorts to reduce selection biases, and controlled or experimental designs to test causality and replicate which cognitive domains are most affected. Overall, the discussion concludes that while these data provide preliminary evidence of preserved global cognition and enhanced working memory in long-term ritual ayahuasca users, further research is needed before clinical or mechanistic claims can be firmly established.

The plants collected were consistent with species commonly used to prepare ayahuasca in Brazilian ritual contexts, and chemical analyses confirmed alkaloid presence. Participants demonstrated average intellectual performance overall and no group differences were found in executive functions or immediate visual memory when comparing experienced users with controls. Experienced ritual users outperformed beginners on verbal and visuospatial working-memory tasks (Digit Span and Corsi), a novel observation in this context. The study concludes that ritual ayahuasca users do not appear to exhibit cognitive impairment and that prolonged ritual use may be associated with preserved or enhanced working-memory performance, while emphasising the need for further, more rigorous studies to clarify short- and long-term cognitive effects and potential therapeutic implications.