Insights on Psychedelics: A Systematic Review of Therapeutic Effects

Psychedelic-assisted experiences at psychoactive doses commonly produce a range of acute subjective effects that have been linked to therapeutic benefit, including connectedness, emotional breakthrough, ego-dissolution and mystical-type experiences. Among these subjective changes, many users and clinicians report sudden, confidence‑laden changes in understanding—termed "insight"—which the authors define here as the sudden emergence of a change in perspective accompanied by clarity or certainty. Earlier studies have reported correlations between acute insight ratings and improvements in wellbeing, depression, anxiety and substance use, but insight itself has not been the subject of a comprehensive, systematic synthesis. Kugel and colleagues set out to fill that gap by conducting the first systematic review focused on psychedelic‑catalysed insight. The review aimed to characterise the prevalence of insight after psychoactive doses of classic serotonergic psychedelics, its relationship to dose, its time-course, and its association with therapeutic outcomes, and to compare insight with mystical‑type experiences as predictors of clinical benefit. The investigators also assessed risk of bias across included studies and made their protocol and materials openly available.

The review followed a pre-registered protocol (PROSPERO CRD42023405854) and PRISMA guidance. Inclusion criteria required peer‑reviewed human studies in English reporting primary quantitative data on insight‑related constructs following psychoactive doses of classic serotonergic psychedelics (LSD, psilocybin, DMT, ayahuasca, 5‑MeO‑DMT or mescaline). All study designs were eligible, including randomised controlled trials, other interventions and survey studies. The search iteratively combined psychedelic substance terms with insight‑related terms and instrument names; the authors reran searches as new psychometric instruments were identified. Seven instruments providing insight data were identified and are documented in the supplementary materials. Screening was performed in Covidence with dual review at key stages: the primary reviewer screened all records and a secondary reviewer screened random subsamples, with consensus and arbitration procedures described. Data extraction captured study design and sample characteristics, drug and dose, timing and scores on insight measures, data sufficient to compute effect sizes for active versus placebo conditions where available, and correlations between insight and mental‑health outcomes; 20% of included studies were double‑checked for extraction accuracy. Where necessary, scores were standardised to conventional units for each scale. Because the included studies were heterogeneous in measures, designs and populations, the protocol did not specify a meta‑analysis. Instead, synthesis combined vote‑counting for prevalence and outcome associations, condition‑level correlations for dose–response relationships, graphical displays of time‑course where multiple timepoints existed, and tabular summaries of associations between acute insight and therapeutic outcomes. For comparisons between insight and mystical‑type experience as predictors, the authors used vote‑counting and, where data permitted, statistical tests (Steiger z‑tests) comparing dependent correlations. Risk of bias was assessed using an adapted ten‑domain instrument tailored to the heterogeneous literature; studies were rated low, medium or high risk on each domain. The primary reviewer completed the RoB assessments and a secondary reviewer checked 20% of studies. Agreement thresholds and procedures for resolving discrepancies were prespecified and reported.

Searches and screening produced 98 included studies from an initial set of 741 unique records after deduplication. The final corpus comprised a mixture of designs: in Discussion the authors report 26 retrospective surveys, 14 prospective surveys, 44 randomised controlled trials and 14 non‑randomised intervention studies. Substances studied included psilocybin (the most frequently represented) and LSD among others. Measures of insight varied across studies. Eighty‑five studies used standardised instruments while 13 used unconventional measures. The 5D/11D‑ASC Insightfulness subscale was the most commonly reported measure (60 studies), followed by the single‑item JHU Psychological Insight (JHU‑PI; 15 studies), the Psychological Insight Questionnaire (PIQ; 14 studies) and the Noetic Quality subscale of the MEQ43 (11 studies). These instruments differ in whether they emphasise phenomenology (certainty, profundity) or content (psychological themes). On prevalence, psychedelic administration at psychoactive doses was consistently associated with insight experiences. Of 46 studies that directly compared classic psychedelics (without pretreatment) to placebo, 43 studies (93.5%) found significantly higher insight scores in the psychedelic condition. A large study cited by the authors reported that 99% of 886 participants indicated they gained insight immediately after a psychedelic experience. The three null comparator studies used a 26 μg LSD dose, which the authors note sits near the microdose threshold and may explain the lack of difference. Insight was positively related to dose. Fifty studies reported dose‑specific data without pretreatment, and across drug–measure combinations higher dose conditions tended to have higher insight scores. For psilocybin in RCTs, the authors report a moderate to strong positive correlation between dose and the 11D‑ASC Insightfulness factor (r = 0.64, 95% CI [0.37, 0.81]). Thirteen studies reported multiple psychoactive doses within a study; 12 of these (92%) observed higher insight ratings at higher doses. Exploratory condition‑level regressions suggested that clinical populations may report higher insight ratings across doses than non‑clinical populations, although the authors attribute this to potential systematic differences such as greater psychotherapeutic support in clinical trials. Longitudinal data on insight ratings were sparse. Most interventional and prospective survey studies measured insight only once: 92% measured insight at a single timepoint, and 88% of those collected that measurement within 24 hours of the experience. Only eight studies reported insight at more than one timepoint. Available evidence indicates that many acute insight ratings persist over time rather than being rapidly devalued; examples include short‑term increases in insight ratings after psilocybin and higher Noetic Quality ratings in a 25‑year follow‑up of the Good Friday Experiment compared with six months post‑session. Regarding therapeutic associations, the majority of studies found that acute psychedelic‑catalysed insight related to later clinical benefit. Across study types, 25 of 29 studies (86%) that examined the relationship reported at least some significant associations between insight and therapeutic outcomes. In randomised trials, 6 of 8 RCTs (75%) that tested the link found a significant relationship. The associations covered a range of outcomes: reductions in psychopathology symptoms (depression, substance use, trauma symptoms), improvements in processes such as reduced thought suppression and experiential avoidance, and increases in positive mental health measures (wellbeing, life satisfaction). The authors report pooled correlations calculated using random‑effects models for psychopathology outcomes (r = -0.33, 95% CI, I2 = 73%) and for positive mental health outcomes (r = 0.43, 95% CI [0.24, 0.58], I2 = 76%), noting considerable heterogeneity. When directly compared with mystical‑type experiences as predictors, insight was more strongly associated with therapeutic outcomes in a majority of studies. Of 22 studies that measured both constructs, 12 (55%) favoured insight, 5 (23%) favoured mystical‑type experiences and 5 (23%) had mixed results. Across 45 outcome comparisons, 31 (69%) indicated stronger correlations for insight. In a subset of nine studies where Steiger z‑tests could be computed for 15 comparisons, seven comparisons favoured insight, eight showed no significant difference and none favoured mystical‑type experiences. The authors caution that overlap in items (for example, Noetic Quality items appearing in mystical experience scales) complicates disentangling the constructs. Risk of bias assessments revealed variability across studies. Interventional studies had lower mean RoB scores than survey studies (M = 0.51 vs M = 0.69), with RCTs tending to score lowest (M = 0.49). RoB scores showed a modest negative correlation with year of publication (r = -0.14), indicating improvements over time, and transparency scores improved in more recent work (RoB transparency correlated with year r = -0.35). Selection bias due to unrepresentative or convenience sampling was the most common risk.

Kugel and colleagues interpret their findings as evidence that psychedelic‑catalysed insight is a common and clinically relevant feature of classic psychedelic experiences at psychoactive doses. The authors emphasise four principal conclusions: insight is highly prevalent after macrodoses; insight intensity tends to increase with dose; many people continue to regard acute insight experiences as meaningful over time; and acute insight is frequently associated with therapeutic improvement, often more reliably than mystical‑type experiences. The review situates these conclusions within prior literature by noting that the prominence of insight in psychedelic reports aligns with cultural and historical characterisations of psychedelics as epistemically revealing substances. The authors also relate their operational definition of insight to problem‑solving and psychotherapeutic literatures, distinguishing phenomenological aspects (suddenness, certainty, emotionality) from content‑focused measures (psychological themes). Key limitations acknowledged include heterogeneous operationalisations and timing of insight measures, the predominance of studies that did not make insight a primary outcome, and the resulting inability to perform a quantitative meta‑analysis on many questions. The authors note the risk of selective reporting and publication bias, and caution that strong correlations between acute experience and outcomes in trial data do not establish causality. They also highlight methodological limitations in many primary studies (lack of placebo control, inadequate blinding, small samples) that constrain causal inference. Potential harms and complexities are explicitly discussed. The investigators warn that insight experiences can be misleading, epistemically biased and, in some contexts, maladaptive; they draw attention to case reports of distress arising from compelling but erroneous revelations and to theoretical links between repeated insight‑like shifts and psychosis. To address these risks, the authors recommend future work to characterise harmful trajectories and early indicators of epistemic harms and to adapt clinical practice accordingly. For future research and practice, the authors call for improved measurement of psychedelic‑catalysed insight, including instruments that capture both phenomenological qualities (suddenness, certainty, surprise, emotionality) and a broader range of content themes. They advocate studies that link the specific content of insights to subsequent behavioural change, designs that can probe causality (for example, manipulations that minimise subjective effects), and longitudinal work to map how insight ratings and content evolve and relate to durable outcomes. Finally, the authors suggest that clinicians and researchers should broaden the focus beyond mystical‑type experiences when assessing mechanisms of therapeutic efficacy in psychedelic interventions.

This systematic review synthesised 98 studies reporting quantitative data on psychedelic‑catalysed insight. The authors conclude that insight is common at psychoactive doses, its intensity is positively correlated with dose, and many people continue to regard these acute insights as meaningful over time. Importantly, acute insight is frequently associated with therapeutic benefit across substances, study designs and outcomes, and often shows stronger associations with improvement than do measures of mystical‑type experience. The review highlights the need for better measurement, longitudinal and causal research, and clinical attention to both the potential benefits and epistemic risks of insight experiences.