Asymmetry in Psychoactive Research: A Bibliometric Study on 15 Psychoactive Drugs

Trading and consuming psychoactive drugs is regulated by national laws and international conventions, which differentiate legal substances such as alcohol and caffeine from controlled substances like LSD or cocaine. Lebrero-Tatay and colleagues describe how these regulatory frameworks, notably post-1961 conventions and national acts, have imposed licences, institutional restrictions, supply barriers and funding difficulties that hinder research on controlled drugs. The authors argue these barriers create an asymmetry in scientific attention across different psychoactive substances, and note prior bibliometric work has identified regional and topic biases but has not fully integrated historical and socio-political context when interpreting publication trends. This study set out to compare publication trends for 15 psychoactive substances from 1960 to 2018, with the aim of testing whether stricter legal control is associated with unstable or declining growth in the research literature. The investigators hypothesised that (a) literature on highly restricted drugs would deviate from an exponential growth curve and show periods of stagnation or decline, and (b) less- or non-restricted drugs would follow a more typical exponential growth. They also anticipated regional differences in productivity related to socio-political factors.

The authors performed a bibliometric analysis using the Web of Science Core Collection in July 2019 to retrieve publications from 1960–2018 that contained topic keywords for each of 15 selected psychoactive drugs. Searches used topic fields with OR-combined keywords chosen after pilot searches; slang terms were avoided. For each drug the investigators extracted annual publication counts and the top five contributing countries using Web of Science analysis tools. Selection aimed to include at least one drug from each legal schedule/class in the US and UK as of July 2019 and at least two substances from each of seven categories in the Drugs Wheel model (cannabinoids, stimulants, empathogens, psychedelics, opioids, dissociatives and depressants). Practical constraints meant some categories had only one representative (for example MDMA, cannabis and ketamine). Datasets with a start year after 1980 were excluded to preserve coverage of the 1970s, and several substances (e.g. mephedrone, 2C-x family, nitrous oxide) were excluded because searches returned many irrelevant records or had insufficient historical coverage. The final list comprised cannabis, alcohol, benzodiazepines, GHB, ketamine, MDMA, codeine, heroin, methadone, LSD, psilocybin, amphetamine, cocaine, khat and methamphetamine. To quantify growth dynamics, annual publication curves for each drug were fitted to an exponential function using non-linear least squares (scipy curve_fit in Python); goodness of fit was assessed with R 2. Initial parameter guesses were stated and the authors note their primary hypothesis concerned the fit (R 2) rather than identical parameter values across drugs. They also calculated a Relative Growth Rate (RGR), a standardised bibliometric measure computed as the change in natural logarithm of counts over fixed periods (decadal intervals, with the last interval 8 years). A positive RGR indicates growth in publications, a negative RGR indicates decline. The authors report that their analysis code was made available in a GitHub repository (the extracted text lists a repository but the URL is not shown).

The search returned 956,703 academic publications (1960–2018) containing keywords for the 15 selected drugs. Temporal patterns varied by substance. During the 1960s most drugs showed a subtle upward trend. In the 1970s trends diverged: publications on heroin, methadone and benzodiazepines rose steeply, whereas LSD and psilocybin declined, with that decline persisting into the 1990s. The 1980s saw steady growth for many drugs. Around 1990 there was an abrupt rise in publications across almost all drugs, consistent with a known surge in Web of Science indexing. The 2000s and 2010s were characterised by rapid increases for most substances, although MDMA and GHB showed stagnation or decreases in these later decades. Exponential-model parameters and fit statistics varied. Psilocybin exhibited the highest estimated growth rate (r = 0.17); most other drugs had growth rates between 0.04 and 0.08. For reference, the overall Web of Science growth rate was 0.04. Goodness-of-fit (R 2) to an exponential curve was highest for alcohol (R 2 = 0.99) and lowest for benzodiazepines (R 2 = 0.66); the Web of Science database curve had R 2 = 0.96. The hypothesised inverse relationship between legal restrictiveness and exponential fit (R 2) was not supported. Decadal Relative Growth Rate (RGR) analyses showed persistent positive RGRs for alcohol, amphetamines, cannabis, ketamine, codeine and methadone. By contrast, khat, benzodiazepines, GHB, MDMA, heroin, LSD, psilocybin, cocaine and methamphetamine each had at least one decade with negative RGR, most commonly during the 1970s and 1980s. When assessing relative share of the combined literature, alcohol accounted for a dominant proportion (reported as between 60% and 80% across years) and literature on legal substances and depressants together represented 60%–80%. LSD’s share dropped from 5–10% in the 1960s to about 1.5% in 1978. Publications classified by schedule showed an increase from the 1960s to the mid‑1970s followed by a drop and stabilisation around 30%; within that, literature on Schedule I drugs has steadily increased since the 1990s while Schedule II literature has decreased. Country-level findings indicated US dominance across the analysed drugs. Example top-five contributors (1960–2018): alcohol — US, China, Japan, India, England; cannabis — US, England, Italy, Germany, Canada; MDMA — US, England, Spain, Germany, Australia; LSD — US, China, England, Canada, Germany; cocaine — US, Canada, England, Spain, Italy. The authors note regional differences in which countries lead research on particular drugs, and they refer readers to an online dataset for the remaining drugs.

Lebrero-Tatay and colleagues interpret three main insights from their bibliometric analysis. First, publication trajectories across the 15 psychoactive substances are heterogeneous, and these trajectories appear to reflect asymmetrical impacts of socio-political events and regulatory changes on research activity. Second, research output is unevenly distributed among substances, with alcohol and other legal depressants and stimulants dominating the literature. Third, there are clear regional differences in research contributions, with the US prevailing but other countries leading for specific drugs in ways that appear linked to national policies and medical regulations. The investigators report that their primary hypothesis — that stricter control would correspond to poorer exponential fits (lower R 2) — was not borne out. Instead, they manually categorised observed growth patterns into three types: monotonic growth (methamphetamine, ketamine, codeine, alcohol), growth with plateaux (khat, cannabis, cocaine), and growth with periods of decay (heroin, LSD, psilocybin, MDMA, methadone, amphetamine, GHB, benzodiazepines). They argue that plateaux or declines deviate from the expected monotonic exponential expansion of scientific literature and may signal external constraints such as prohibitions, licensing hurdles, supply shortages and unstable funding. The authors discuss specific historical examples: the decline in psychedelic research after 1970s policy shifts, the 1985 US ban on MDMA and anti‑MDMA campaigns in the UK, Switzerland’s permissive approach to psilocybin research in the late 1980s, Italy’s earlier legal therapeutic recognition of cannabinoids, and prominent contributions from khat‑producing countries where the substance is legal. The paper highlights a consequence of asymmetric research attention: skewed understanding of neurobiology and addiction mechanisms. The authors note that the predominance of depressant and stimulant research has shaped theories such as dopamine‑centred models of addiction, which may not generalise to other substance classes. They therefore caution that unequal research investment can lead to incomplete theories and suboptimal treatments. The researchers acknowledge limitations: publication counts capture only a narrow dimension of research activity, and their analysis did not include other bibliometric measures such as citation networks or links to social indicators. They recommend future studies apply wider bibliometric tools and suggest policymakers and funders should consider reducing barriers and increasing investment in research on highly regulated substances to improve knowledge of safety and therapeutic potential.