Abstract
Background:
Psychedelic-assisted therapy (PAT) has demonstrated substantial efficacy across a range of mental disorders. However, heterogeneity between patients confers differential responsiveness. This systematic review aims to explore factors which may predict therapeutic responses to PAT.
Methods:
A systematic search was performed from inception through to March 2024 and studies that assessed predictors of response related to the use of classic psychedelics for mental disorders were included.
Results:
A total of 54 studies investigating potential predictors of treatment response to psychedelic-assisted therapy were included in the review. These studies encompassed adult populations diagnosed with substance-use disorders, major depressive disorder, anxiety disorders, obsessive-compulsive disorder, post-traumatic stress disorder and existential distress related to life-threatening illness as well as naturalistic samples reporting psychopathological symptoms without a formally confirmed diagnosis. The most frequently reported predictor of therapeutic response was the intensity of the acute psychedelic experience, particularly mystical-type experiences (MTEs), though this was not consistent across all disorders or time points. Factors related to set, setting and dose were frequently associated with the likelihood and intensity of MTEs.
Conclusions:
The acute psychedelic experience, especially MTEs, was the most frequently reported predictor of therapeutic response. Future trials should explore a broader range of predictors, include longer-term follow-up and improve methodological consistency to strengthen the evidence base for reliable predictors of therapeutic response.
Introduction
After decades of destigmatization, advocacy, education and rigorous research, psychedelic drugs are currently being revisited for their therapeutic application in mental disorders. Despite relatively high response rates associated with psychedelic-assisted therapy (PAT), heterogeneity between patients confers differential responsiveness. There have been some attempts to decipher factors which contribute to favorable therapeutic outcomes. Romeo et al. (2021) investigated clinical and biological predictors of therapy response in mental disorders across 20 studies. The review found that the main predictive factor across diagnoses was the intensity of the acute psychedelic experience (Romeo et al., 2021). The acute psychedelic experience refers to the subjective effects elicited by the psychedelic drug and was assessed by questionnaires such as the Hallucinogen Rating Scale (HRS) (Riba et al., 2001), the Mystical Experience Questionnaire (MEQ) (Barrett et al., 2015) and the Five-Dimensional Altered States of Consciousness Questionnaire (5D-ASC) (Dittrich et al., 2010). For example, individuals who had intense mystical-type experiences (MTEs) typically reported feelings of unity, ecstasy, awe, tranquility and transcendence of time, space and self. Across several studies included in the review, this level of consciousness alteration predicted subsequent clinical improvement (Romeo et al., 2021). The authors also examined sociodemographic variables such as age, sex, relationship status and level of education, but found that these do not significantly predict responses to PAT (Romeo et al., 2021)
Further research has shown that individuals’ states and traits may predict the acute psychedelic experience (Aday et al., 2021). This relates to the concept of set, that is, an individual’s pre-existing mental state prior to the psychedelic experience. This encompasses thoughts, moods and expectations of the patient that may predict reactions to psychedelic drugs. Across fourteen studies, the review found that individuals who exhibited high levels of absorption, acceptance and surrender were more likely to report MTEs (Aday et al., 2021). Conversely, individuals low in openness and surrender or in apprehensive or confused psychological states were more likely to experience acute adverse reactions. This suggests that the individual’s “set” influences the intensity of the MTE, which can subsequently predict the patient’s degree of clinical improvement.
While these two reviews provide valuable summaries of predictors of therapeutic response and experiential factors, the evidence base has expanded considerably since their publication. In the past 4 years alone, more than 30 additional studies exploring predictive factors have been published, many of which involve larger and more rigorously designed clinical trials. The current review incorporates this growing body of evidence and aims to synthesize findings across disorders, considering predictors related to baseline characteristics, the acute experience and biological alterations. For instance, some studies have investigated neurobiological changes, such as inflammatory cytokine modulation, alterations in brain connectivity and amygdala activation. We treat this as both a putative neurobiological mechanism and correlate of improvement. In addition, this review explicitly examines predictors of the acute psychedelic experience itself, particularly those that influence the likelihood or intensity of MTEs.
We adopt a transdiagnostic perspective, synthesizing studies across psychiatric and subclinical populations. While our results are organized by clinical indication to aid interpretability, this broader lens reflects growing interest in shared mechanisms of psychedelics. Neurobiological models including 5-HT2A receptor activation (Preller et al., 2017), the psychoplastogen model (Vargas et al., 2021) and the REBUS model (Carhart-Harris and Friston, 2019) as well as psychological models such as ego dissolution and meaning-making propose processes that may underlie effects across disorders. These frameworks have been implicated in substance-use disorders (SUDs) (Johnson et al., 2016, 2017; Noorani et al., 2018), depression (Davis et al., 2021; Roseman et al., 2018b; Uthaug et al., 2021), post-traumatic stress disorder (PTSD) (Vargas et al., 2021) and existential distress (Belser et al., 2017; Ross et al., 2016), providing a conceptual basis for examining predictors across diagnostic boundaries.
PAT is resource-intensive, involving substantial time and effort from the patients, therapists, facilitators and study personnel. Thus, identifying factors related to improved PAT response is integral to optimizing patient outcomes. The goal of this systematic review is twofold: (1) to identify predictors of therapeutic response to PAT and (2) to better understand the factors that contribute to the emergence of specific acute psychedelic experiences, particularly those of a mystical nature, with consideration of other potential predictors and mechanisms.
Materials and methods
Study registration
The current systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (S1) guidelines (Page et al., 2022). The protocol for this study was registered on PROSPERO (record number CRD42023440486). We searched the Medline, Cochrane and Scopus databases for studies published from inception through to March 2024. The searches were executed between February and March 2024 using keywords related to psychedelic drugs and mental disorders.
Search strategy
To elucidate our research question, we searched the MEDLINE, Cochrane and Scopus databases using keywords related to psychedelic drugs and mental disorders. Specifically, classic psychedelics were included (“psilocybin,” “lysergic acid diethylamide (LSD),” “ayahuasca,” “dimethyltryptamine (DMT),” “5-MeO-DMT,” “ibogaine” or “mescaline”), along with mental disorders (“substance use disorder (SUD)/dependence/addiction,” “major depressive disorder (MDD),” “treatment-resistant depression (TRD),” “anxiety,” “obsessive-compulsive disorder (OCD),” “post-traumatic stress disorder (PTSD),” “personality disorder”) as well as terms related to predictors of response (“predict,” “response,” “reaction,” “improve”). The search was limited to studies published in English in peer-reviewed journals. Study selection was performed by GV and verified by FB.
Eligibility criteria
Eligibility criteria included (i) adult participants, (ii) classic psychedelic drugs, (iii) mental disorders and (iv) the analysis of predictive factors of response to PAT. Our review includes studies conducted in both clinical and nonclinical settings. While our primary focus is on PAT, we also considered studies involving participants with clinically relevant psychiatric symptoms assessed via validated measures (e.g., Beck’s Depression Inventory (BDI), Positive and Negative Affect Schedule (PANAS)), even if formal diagnoses were not confirmed. Inclusion of these studies provides complementary data on real-world predictors of symptom improvement and enhances the ecological validity of our findings. The inclusion of both contexts allows examination of whether predictive factors generalize beyond clinical trials, an important consideration for future translation of findings. Potential differences in predictive patterns between contexts are discussed. Studies investigating psychedelic microdosing were excluded, since such doses elicit responses below the perceptual threshold, unfolding their potential benefit by other means than the subjective (acute) experience (Bershad et al., 2019).
Outcome measures and data extraction
To evaluate factors that are associated with therapeutic responses to PAT, various predictive factors were examined. Overall, this included sociodemographic factors, details of diagnoses, study design and patient characteristics at baseline. Data related to set and setting factors and relevant biological test data were extracted. Lastly, data associated with the acute subjective effects of the dosing session were collected. In synthesizing the results, the studies were grouped by the primary indication. The exact effect sizes of these predictive factors can be found in the Supplemental Material (S2).
Risk of bias assessment
Risk of bias was assessed using a variety of validated quality assessment tools to cover multiple study designs (S3). Risk of bias for open-label trials and randomized controlled trials (RCTs) was assessed using the modified Newcastle–Ottawa Scale (Deeks et al., 2003) and the Cochrane risk of bias tool (Sterne et al., 2019), respectively. The checklist for retrospective database studies (Motheral et al., 2003) was used to assess the quality of retrospective studies. For qualitative and observational studies, the Critical Appraisal Skills Programme (CASP) qualitative studies checklist and the CASP cohort studies checklist (CASP, 2018) were employed, respectively. Quality assessment was performed by GV and verified by FB.
Synthesis table
Our approach to classifying evidence was informed by the logic of the Cochrane GRADE framework (Guyatt et al., 2008) and its qualitative extension (GRADE-CERQual) (Lewin et al., 2018), which assess the certainty of a body of evidence across domains such as risk of bias, consistency and adequacy of data. Because our review encompassed heterogeneous study designs and aimed to evaluate predictors at the individual level rather than outcomes more broadly, we adopted a system that applied predefined thresholds across quantitative and qualitative evidence. This enabled a pragmatic synthesis while remaining conceptually aligned with GRADE.
To synthesize predictors across studies (Table 7), we recorded the primary disorder(s) in which it was observed, the follow-up timeframe (short-term ⩽1 month, intermediate >1 month to ⩽6 months, long-term >6 months), evidence strength (established, suggestive, inconclusive) and the quality ratings. Predictors with ⩾2 RCTs or ⩾3 high-quality studies (any design) showing consistent effects were classified as established; those with ⩾2 studies but lower quality or inconsistent findings were classified as suggestive; all others were deemed inconclusive. Quality ratings for each predictor were derived from risk of bias assessments. For RCTs, “high” quality required low risk of bias in all domains; “moderate” allowed one or two domains at unclear/high risk; “low” reflected greater risk. For open-label trials, high, moderate and low corresponded to Newcastle–Ottawa scores of ⩾4/5, 3/5 and <3/5, respectively. For retrospective and observational studies, high quality required checklist scores ⩾20/26, moderate 16–19/26 and low <16/26. For qualitative studies, high, moderate and low were defined as CASP checklist scores ⩾16/20, 13–15/20 and <13/20, respectively.
Results
Article identification process
A total of 934 records were identified through database searches (MEDLINE, n = 517, Cochrane, n = 226, SCOPUS, n = 191). After removing duplicates, 634 records remained for screening. Following the initial screening of titles and abstracts, 514 records were excluded. The full texts of the remaining 120 articles were assessed for eligibility, resulting in the exclusion of 66 studies for the following reasons: review or meta-analysis (n = 24), no analysis of a predictive factor (n = 19), no addictive or psychiatric disorder included (n = 11), case study (n = 7) and microdosing studies (n = 5). A total of 54 studies met the inclusion criteria and were included in the qualitative synthesis (Figure 1).
Flowchart of the article identification process according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines.
Substance-use disorder
Across the eight studies investigating PAT for SUD, the quality of the acute psychedelic experience was consistently associated with improved clinical outcomes (Table 1). In particular, higher MEQ scores were associated with reductions in substance use and increased abstinence across alcohol, tobacco, opioids and stimulants (Bogenschutz et al., 2015; Garcia-Romeu et al., 2020; Johnson et al., 2016). These effects were observed in both open-label trials and retrospective surveys and in some cases remained significant up to 12 months post-intervention (Johnson et al., 2016), although sample sizes were generally small. Increased ratings of personal meaning and spiritual significance also correlated with favorable treatment outcomes (Davis et al., 2017; Garcia-Romeu et al., 2015, 2019, 2020; Johnson et al., 2016, 2017). Argento et al. (2019), in a qualitative analysis, found that increased connectedness to self, others, nature or spirit following ayahuasca use was associated with long-term reductions in substance use.
Substance-use disorder.
5D-ASC: Five-Dimensional Altered States of Consciousness; CBT: cognitive behavioral therapy; HRS: Hallucinogen Rating Scale; MEQ: Mystical Experience Questionnaire; MET: Motivational Enhancement Therapy; SOCQ: States of Consciousness Questionnaire.
Same study population.
n/a indicates that the time point was not specified, or that an effect size could not be extracted (e.g., qualitative studies or results reported solely as p-values or unstandardised beta coefficients).
These experiential factors appeared more predictive than baseline sociodemographic variables, although one retrospective study in tobacco addiction found that lower baseline substance consumption was modestly associated with better outcomes (Johnson et al., 2017). Conversely, two other retrospective online surveys reported the opposite pattern, with higher disorder severity associated with greater reductions in alcohol use (Garcia-Romeu et al., 2019) and cannabis, opioid and stimulant misuse (Garcia-Romeu et al., 2019). Older age was associated with more favorable outcomes in two surveys (Garcia-Romeu et al., 2020; Johnson et al., 2017). Across SUD studies, reported predictors included the depth and quality of the acute psychedelic experience (particularly mystical or insightful states) as well as baseline factors such as age and disorder severity.
Treatment-resistant depression
Across studies investigating TRD (Table 2), the association between the intensity and quality of the acute psychedelic experience was less consistent. Some clinical trials reported a connection between acute drug effects measured by 5D-ASC and HRS (Carhart-Harris et al., 2018; Palhano-Fontes et al., 2019; Roseman et al., 2018b) whereas an open-label trial (D’Souza et al., 2022) and interview study (Kaelen et al., 2018) found no association. Findings for specific experiential dimensions, such as the MEQ, were similarly mixed. While one RCT with ayahuasca reported that higher MEQ scores predicted symptom reduction (Palhano-Fontes et al., 2019), another RCT with psilocybin found that neither MEQ nor challenging experience questionnaire scores were associated with treatment outcomes (Sloshower et al., 2023).
Treatment-resistant depression.
ACT: acceptance and commitment therapy; ASC: Altered States of Consciousness; CEQ: Challenging Experience Questionnaire; DMT: dimethyltryptamine; GAD-7: General Anxiety Disorder-7; GRID-HAMD: GRID Hamilton Rating Scale for Depression; HAM-D: Hamilton Depression Rating Scale; HRS: Hallucinogen Rating Scale; MEQ: Mystical Experience Questionnaire; MADRS: Montgomery–Åsberg Depression Rating Scale; PANAS: Positive and Negative Affect Schedule; PSI: Preparation, Support, Integration; QIDS: Quick Inventory of Depressive Symptomatology; RCT: randomized controlled trial; VAS: Visual Analogue Scale; OBN: oceanic boundlessness.
or b or c or dSame study population.
n/a indicates that the time point was not specified, or that an effect size could not be extracted (e.g., qualitative studies or results reported solely as p-values or unstandardised beta coefficients).
Several psychological and contextual factors also emerged as predictors. Higher baseline neuroticism (at trend level) (Erritzoe et al., 2018), lower emotional expressivity during baseline interviews (Carrillo et al., 2018) and fewer prior unsuccessful treatment attempts (de Almeida et al., 2019) were each associated with greater clinical improvement. Biologically, reductions in C-reactive protein levels from baseline to 2 days post-dosing predicted symptom improvement (Galvão-Coelho et al., 2020), whereas changes in salivary cortisol did not (de Menezes Galvão et al., 2018).
Neuroimaging studies identified additional correlates of response. Roseman et al. (2018b) found that increased amygdala activation to fearful stimuli from baseline to 1 day post-dosing was associated with symptom improvement at 3 weeks. Mertens et al. (2020) reported that increased functional connectivity between the ventromedial prefrontal cortex and occipital-parietal regions from baseline to 1 day post-treatment predicted reductions in depressive symptoms at 1 week. Similarly, Daws et al. (2022) found that early-phase decreases in network modularity from baseline to 1 day post-dosing predicted sustained improvement at 6 months. Copa et al. (2024) demonstrated that higher baseline functional connectivity across visual, default mode, executive and salience networks predicted both early (1–4 weeks) and later (24 weeks) symptom reductions. Complementing these findings, Skosnik et al. (2023) observed that increased theta power from 1 day to 2 weeks post-treatment was associated with clinical response at 2 weeks. Finally, factors related to the dosing session also played a role. An RCT found that participants who were administered higher doses of psilocybin showed greater improvement (Goodwin et al., 2022). In addition, Kaelen et al. (2018) showed that specific features of the music experience during dosing, such as feelings of “liking,” “resonance” and “openness” were significantly associated with therapeutic outcome. Across TRD studies, reported predictors included acute psychedelic experience measures, biological markers, baseline psychological traits and session-specific factors.
MDD without treatment-resistance
As summarized in Table 3, several studies in major depressive disorder (MDD populations identified the quality of the acute psychedelic experience as a predictor of clinical response. The general intensity of subjective effects predicted outcomes in a retrospective online survey (Kopra et al., 2022) and a non-RCT 3 months post-dosing (Aaronson et al., 2023), but not in an RCT at 2 weeks (von Rotz et al., 2023). Findings for oceanic boundlessness (OBN) were mixed with an open-label trial showing prediction at 1–2 weeks post-dosing (Timmermann et al., 2024) and an observational study showing an association at 1 month but not 1 year later (van Oorsouw et al., 2022). Higher ratings of ego dissolution were associated with improvement at 1 month (van Oorsouw et al., 2022) and 1 day post-dosing but not at 4 weeks (Uthaug et al., 2019). Higher ratings of psychological insight and emotional breakthrough were associated with reduced depressive symptoms in an RCT (Davis et al., 2021) and prospective online survey (Nygart et al., 2022), respectively. Two RCTs found MEQ scores predicted improvement (Davis et al., 2021; Gukasyan et al., 2022) with the latter showing an association up to 12 months post-dosing. This association was also reported in a prospective psilocybin survey (Nayak et al., 2025) but not in another prospective online survey including all classic psychedelics (Nygart et al., 2022). Challenging experiences were not associated with improvement.
MMD (without treatment-resistance).
5D-ASC: Five-Dimensional Altered States of Consciousness; ACE: Accept, Connect, Embody; ACE**: Adverse Childhood Experience; ASC: Altered States of Consciousness; BDI: Beck’s Depression Inventory; BSI: Brief Symptom Inventory; CEQ: Challenging Experience Questionnaire; DMT: dimethyltryptamine; EBI: Emotional Breakthrough Inventory; EDI: Ego-Dissolution Inventory; GAD-7: General Anxiety Disorder-7; GRID-HAMD: GRID Hamilton Rating Scale for Depression; MADRS: Montgomery–Åsberg Depression Rating Scale; MEQ: Mystical Experience Questionnaire; PHQ-9: Patient Health Questionnaire-9; PSI: Preparation, Support, Integration; QIDS: Quick Inventory of Depressive Symptomatology; RCT: Randomized Controlled Trial; MDD: major depressive disorder; LSD: lysergic acid diethylamide.
Same study population.
n/a indicates that the time point was not specified, or that an effect size could not be extracted (e.g., qualitative studies or results reported solely as p-values or unstandardised beta coefficients).
Other reported predictors included higher baseline trait suggestibility at 6 weeks post-dosing (Szigeti et al., 2024), higher state of surrender at 2–3 months (Nygart et al., 2022), higher drug doses, having a medicinal motive and previous psychedelic experience (Nygart et al., 2022) as well as increased lifetime psychedelic use (Raison et al., 2022). An RCT reported that stronger therapeutic alliance predicted better depression outcomes (Murphy et al., 2022) and an open-label trial found that enhanced baseline connectivity from the anterior cingulate cortex was associated with improvement at 4 weeks (Doss et al., 2021). Overall, predictors spanned experiential, baseline and neurobiological domains.
Depression and anxiety in life-threatening illness
In patients with life-threatening cancer, higher MEQ scores during psilocybin sessions were associated with reductions in depression and anxiety symptoms in the short-term (Griffiths et al., 2016; Ross et al., 2016), as summarized in Table 4. Associations were observed across depression multiple measures with medium-sized correlations between MTEs and symptom improvement 5–6 weeks post-dosing. Findings at longer-term follow-up were mixed: Agin-Liebes et al. (2020) reported that while MEQ scores predicted symptom reduction initially, this association was no longer significant at 3.2–4.5-year follow-up. A later analysis of the same cohort found that reductions in loss of meaning remained significantly associated with therapeutic response at 4.5 years (Ross et al., 2021).
Depression and anxiety in patients with life-threatening cancer.
BDI: Beck’s Depression Inventory; BSI: Brief Symptom Inventory; HADS: Hospital Anxiety and Depression Scale; HAM-A: Hamilton Anxiety Rating Scale; HAM-D: Hamilton Depression Rating Scale; MEQ: Mystical Experiences Questionnaire; RCT: Randomized Controlled Trial; STAI: The State-Trait Anxiety Inventory.
Same study population.
n/a indicates that the time point was not specified, or that an effect size could not be extracted (e.g., qualitative studies or results reported solely as p-values or unstandardised beta coefficients).
Anxiety and related disorders (OCD, PTSD)
Although OCD and PTSD are no longer classified as anxiety disorders in current diagnostic frameworks, they are often grouped with related conditions due to overlapping phenomenology and treatment approaches. Across the limited studies available (Table 5), the predictive value of acute psychedelic experiences was mixed. In an RCT using LSD for generalized anxiety disorder, social phobia and panic disorder, higher MEQ and OBN scores were associated with greater reductions in anxiety symptoms at 4 month follow-up (Holze et al., 2022). Conversely, a nine-person open-label trial of psilocybin for OCD found no association between HRS scores and symptom change at 1 day (Moreno et al., 2006). In a study of veterans with PTSD undergoing a combined ibogaine and 5-MeO-DMT treatment, greater shifts in consciousness, younger age and higher baseline symptom severity were associated with short-term improvement (Xin et al., 2023). However, the observed association between changes in consciousness and symptom reduction was not sustained at 6 months.
Anxiety and related disorders.
HRS: Hallucinogen Rating Scale; MEQ: Mystical Experience Questionnaire; PEQ: Persisting Effects Questionnaire; PCL-5: Post-Traumatic Stress Disorder Checklist for DSM-5; PSI: Preparation, Support, Integration; RCT: Randomized Controlled Trial; VAS: Visual Analogue Scale; YBOCS: Yale-Brown Obsessive-Compulsive Scale; OCD: obsessive-compulsive disorder; LSD: lysergic acid diethylamide.
n/a indicates that the time point was not specified, or that an effect size could not be extracted (e.g., qualitative studies or results reported solely as p-values or unstandardised beta coefficients).
Depression and anxiety symptoms
Several studies examined predictors of therapeutic response in populations reporting combined depression and anxiety symptoms (Table 6). Associations between specific dimensions of the acute psychedelic experience and clinical outcomes were mixed. Higher MEQ scores were associated with symptom reductions up to 3 months post-dosing (Davis et al., 2019, 2020; Ruffell et al., 2021), but not at 6 months (Ruffell et al., 2021). Ego dissolution findings were inconsistent: one observational study found an association with symptom improvement (Uthaug et al., 2021), whereas a retrospective survey did not observe this effect (Agin-Liebes et al., 2021). In contrast, increased personal insight was associated with therapeutic response in two retrospective surveys (Agin-Liebes et al., 2021; Davis et al., 2020). Challenging experience scores were not associated with improvements.
Depression and anxiety symptoms.
5-MeO-DMT: 5-methoxy-N,N-dimethyltryptamine; BDI: Beck’s Depression Inventory; CEQ: Challenging Experience Questionnaire; CTQ: Childhood Trauma Questionnaire; DASS-21: Depression, Anxiety and Stress Scale-21; EDI: Ego-Dissolution Inventory; MEQ: Mystical Experience Questionnaire; PANAS: Positive and Negative Affect Schedule; PIQ: Psychological Insight Questionnaire; POMS: Profile of Mood States; PTSD: post-traumatic stress disorder; SUD: substance-use disorder.
n/a indicates that the time point was not specified, or that an effect size could not be extracted (e.g., qualitative studies or results reported solely as p-values or unstandardised beta coefficients).
Synthesis of predictors across included studies
Table 7 summarizes key predictors identified across the included studies, including associated disorders, follow-up timeframes, evidence strength, study designs and quality ratings. Predictors were most frequently observed in relation to the acute psychedelic experience, including MTEs, OBN, ratings of personal meaning, spiritual significance and psychological insight. Several baseline factors, contextual influences and post-dosing biological changes were also identified. Predictors are classified by timeframe: short-term (⩽ 1 month), intermediate (>1–6 months) or long-term (>6 months) and evidence strength (established, suggestive or inconclusive) based on the number, quality and consistency of supporting studies (see “Materials and methods” section). Quality ratings were derived from standardized risk-of-bias assessments. To further synthesize the main results from the systematic review, the frequency of various predictive factors identified in the included studies is shown in Figure 2.
Synthesis of positive predictive factors across studies with evidence strength and quality rating.
CEQ: Challenging Experience Questionnaire; MDD: major depressive disorder; MEQ: Mystical Experience Questionnaire; OBN: Oceanic Boundlessness; PIQ: Psychological Insight Questionnaire; PTSD: post-traumatic stress disorder; RCT: randomized controlled trial; SUD: substance-use disorder; TRD: Treatment-Resistant Depression.
Frequency of predictive factors identified in the included studies. *Unsuccessful refers to factors that did not find a significant association between predictive factors and therapeutic response.
Discussion
Substance-use disorder
In SUD populations, the most frequently reported predictors were phenomenological features of the acute psychedelic experience, particularly MTEs and ratings of personal meaning or insight (Bogenschutz et al., 2015; Garcia-Romeu et al., 2015, 2019, 2020; Johnson et al., 2016, 2017). These associations were generally more consistent than those observed for baseline demographic or clinical variables, although this may reflect a greater number of studies assessing experiential factors. Some studies also found that greater disorder severity predicted larger reductions in substance use (Garcia-Romeu et al., 2019, 2020), potentially relating to higher motivation for change in individuals with more severe presentations. Qualitative evidence indicating enhanced connectedness to self, others, nature or spirit (Argento et al., 2019) aligns with the possibility that relational and existential shifts contribute to sustained behavioral change in this population.
Treatment-resistant depression
In TRD populations, associations between the acute psychedelic experience and therapeutic response were less consistent than in SUD. While several trials reported links between global measures of acute drug effects or MTEs and symptom improvement (Carhart-Harris et al., 2018; Palhano-Fontes et al., 2019; Roseman et al., 2018b), others found no such associations (D’Souza et al., 2022; Sloshower et al., 2023). This variability may reflect differences in study design, sample size, dosing protocols or measuring tools to assess the acute experience, but could also indicate that in TRD, experiential intensity alone is insufficient to account for clinical change. A broader range of psychological, biological and contextual factors were also identified in this group. Higher baseline neuroticism (trend level) (Erritzoe et al., 2018) and lower emotional expressivity prior to dosing (Carrillo et al., 2018) may reflect psychological characteristics related to PAT response, although further research is needed to clarify these associations. Fewer prior unsuccessful treatment attempts were also linked to therapeutic response (de Almeida et al., 2019), possibly indicating a greater remaining capacity to benefit from treatment.
Neuroimaging and electrophysiological findings such as pre-post treatment changes in amygdala reactivity (Roseman et al., 2018a), network connectivity (Mertens et al., 2020; Copa et al., 2024) and post-treatment changes in theta power (Skosnik et al., 2023) point to potential neurobiological mechanisms underlying improvement, although the timing of these measurements means they may be better understood as correlates of therapeutic change rather than pre-treatment predictors. Contextual elements during dosing, including psilocybin dose (Goodwin et al., 2022) and subjective music experience (Kaelen et al., 2018), were also linked to outcomes, consistent with evidence that the sensory and relational environment can influence the therapeutic process.
Major depressive disorder without treatment-resistance
In MDD without treatment resistance, associations between the acute psychedelic experience and clinical response were generally more consistent than in TRD, though variability remained. Multiple trials and survey studies linked higher MEQ scores, as well as dimensions such as psychological insight and emotional breakthrough, to greater symptom reduction (Davis et al., 2021; Gukasyan et al., 2022; Nayak et al., 2025; Nygart et al., 2022). One trial found the MEQ-outcome association persisted up to 12 months (Gukasyan et al., 2022), suggesting potential long-term relevance of certain experiential qualities, although other studies did not replicate this. Findings for OBN and ego dissolution were mixed (Timmermann et al., 2024; Uthaug et al., 2019; van Oorsouw et al., 2022), likely reflecting differences in methodology, follow-up duration and sample characteristics. As in other diagnostic groups, challenging experiences were not associated with clinical improvement, indicating that positive or meaningful aspects of the experience may be more relevant.
Beyond the acute experience, potential predictors included baseline traits such as suggestibility (Szigeti et al., 2024) and surrender (Nygart et al., 2022), which may influence a participant’s capacity to engage with and integrate the dosing session. Contextual variables including dose, treatment motive, previous psychedelic experience and the strength of the therapeutic alliance (Murphy et al., 2022) align with broader evidence for the role of set and setting in shaping outcomes. Neuroimaging findings, such as enhanced baseline anterior cingulate connectivity as a correlate of improvement (Doss et al., 2021), suggest possible neural markers of responsiveness, though the small number of studies precludes strong conclusions. Overall, non-TRD MDD evidence suggests a more consistent pattern of association between subjective experience and outcome than in TRD, but the variability across measures and designs warrants cautious interpretation.
Depression and anxiety in life-threatening illness
In patients with life-threatening cancer, MTEs during psilocybin sessions have been consistently associated with short-term improvements in depression and anxiety (Griffiths et al., 2016; Ross et al., 2016). Medium-sized correlations across multiple affective measures suggest these experiences may play a role in the early therapeutic process. However, their influence over time is less certain. At over 3 years post-treatment, MEQ scores no longer predicted outcomes (Agin-Liebes et al., 2020), although reductions in loss of meaning remained associated with benefit at 4.5 years (Ross et al., 2021). This pattern suggests that existential or meaning-related changes may represent a more enduring pathway to sustained symptom reduction although further research is warranted to differentiate between experiential features that drive early change and those supporting long-term improvement.
Anxiety and related disorders
In anxiety and related conditions, available evidence points to a heterogeneous pattern of predictors. An RCT of LSD-assisted therapy for generalized anxiety disorder, social phobia and panic disorder found that higher MTE and OBN scores were associated with sustained anxiety reduction at 4 months (Holze et al., 2023), suggesting that positively valenced experiences may support longer-term benefit. In contrast, an open-label trial of psilocybin for OCD (n = 9) found no association between acute experience intensity and outcomes at 1 day (Moreno et al., 2006), a result likely influenced by short follow-up and limited power. In PTSD, a combined ibogaine and 5-MeO-DMT program linked greater shifts in consciousness, higher baseline severity and younger age to short-term improvement, though experiential effects were not sustained at 6 months (Xin et al., 2023). These mixed findings highlight potential disorder-specific influences and the need for larger, well-controlled studies with extended follow-up.
Depression and anxiety symptoms
In populations reporting combined depression and anxiety symptoms, predictors showed a mixed pattern. MTEs were linked to short-term improvements in several studies (Davis et al., 2019, 2020; Ruffell et al., 2021), though the association was not sustained at 6 months (Ruffell et al., 2021). Ego dissolution yielded similarly inconsistent results, with one observational study reporting a positive association (Uthaug et al., 2021) and a retrospective survey finding none (Agin-Liebes et al., 2021). Personal insight was related to positive outcomes (Agin-Liebes et al., 2021), consistent with suggestions that meaning-making processes may contribute to symptom change. Neuroimaging data indicate that acute increases in amygdala reactivity from baseline to 1 day post-psilocybin may also accompany early improvements (Barrett et al., 2020), although such changes are better understood as correlates than pre-treatment predictors. As in other diagnostic categories, challenging experiences were not associated with clinical benefit, supporting the view that positively valenced experiences may be more relevant to therapeutic response.
Consistency of MTEs across disorders and study design
Across diagnostic groups, the intensity of MTEs was the most frequently reported predictor. This was consistent across study design with an association between MEQ scores and symptom improvement present in six RCTs, three open-label studies, one observational study and five surveys. This association was consistently observed in both clinical (Griffiths et al., 2016; Gukasyan et al., 2022; Ross et al., 2016) and non-clinical contexts (Davis et al., 2019; Garcia-Romeu et al., 2019, 2020). The consistency of this association did, however, vary slightly by disorder. In SUD, five studies linked higher ratings on the MEQ to improved outcomes across RCTs, open-label trials and retrospective surveys, with some effects persisting up to 12 months (Johnson et al., 2016). In non-TRD MDD, three studies showed an association, also at the 12 month follow-up in one RCT (Gukasyan et al.. 2022). In studies investigating depression and anxiety in patients with life-threatening cancer, two studies showed an association. However, at the 3–4 year follow-up, MEQ no longer predicted therapeutic response (Agin-Liebes et al., 2020). TRD, showed the greatest inconsistency but direct comparison is complicated by differences in predictor focus: many TRD studies examined neurobiological changes associated with symptom improvement, with only two of nineteen assessing MEQ scores (one reporting an association, one not), whereas three of four cancer-related anxiety studies examined MEQ, all reporting short-term associations.
Temporal patterns of prediction
Across studies, most evidence relates to short-term prediction, with the majority of follow-up assessments conducted within 1 month of dosing. Only a small number of variables demonstrated associations with outcomes beyond 6 months, and these were typically reported in just one or two studies. These included age (Xin et al., 2023), MEQ scores (Gukasyan et al., 2022; Johnson et al., 2016), ratings of personal meaning and spiritual significance (Garcia-Romeu et al., 2015; Johnson et al., 2016) and reduced network modularity (Daws et al., 2022). The limited number and diversity of long-term predictors make it difficult to draw firm conclusions about their robustness or generalizability. Clinically, the predominance of short-term data underscores the need for caution when extrapolating predictors to sustained outcomes. Future trials should incorporate extended follow-up to identify factors that reliably forecast long-term benefit, as these may differ from those that characterize early response.
The mystical experience: Phenomenology and potentiators
The current systematic review identified the mystical subtype of psychedelic experiences as the most frequently reported predictor of therapeutic response across most disorders. An MTE is typically characterized by feelings of unity, bliss, awe, tranquility and a sense that the experience is ineffable (Barrett et al., 2015). Several studies included in this systematic review showed that MTEs, in particular, are associated with higher therapeutic benefits. This was found across a range of diagnoses including SUD (Bogenschutz et al., 2015; Garcia-Romeu et al., 2015, 2019; Johnson et al., 2016), MDD (Doss et al., 2021; Gukasyan et al., 2022; Kaelen et al., 2018; Murphy et al., 2022; Palhano-Fontes et al., 2019), depression and anxiety in patients with life-threatening cancer (Ross et al., 2016) and anxiety disorder (Holze et al., 2022). While the strength and consistency of this association varied—particularly in TRD and longer-term follow-ups, the overall pattern was robust enough to warrant further investigation into the therapeutic relevance of MTEs.
Given the association between MTEs and treatment outcomes, it is important to consider elements that influence the occurrence of such experiences. Factors related to set and settings are important factors influencing the nature of the psychedelic experience and subsequently, the success of PAT (Figure 3). As mentioned earlier, individuals high in traits of absorption, openness, acceptance and surrender are more likely to have MTEs (Aday et al., 2021; Haijen et al., 2018). Moreover, high levels of anxious attachment at baseline positively correlate with MTEs (Stauffer et al., 2020). This can potentially be explained by a transference in connection-seeking. Such individuals display a strong emphasis on maintaining proximity with others which may prime them to welcome such extreme interconnectedness that accompanies MTEs (Stauffer et al., 2020). The intention or motive of the participant has also been shown to impact the type of psychedelic experience. Haijen et al. (2018) found that having clear intentions and expectations for the psychedelic experience was positively correlated with MTEs. Finally, therapeutic rapport can influence the nature of the psychedelic experience. Murphy et al. (2022) found that a stronger therapeutic alliance, as assessed on a participant-rated scale, predicted more intense MTEs and emotional breakthroughs. Thus, these findings highlight several factors that differentially influence the occurrence and intensity of MTEs.
Moderators of MTEs. MTE: mystical-type experience.
The setting in which a psychedelic drug is ingested has also been shown to influence certain phenomenological features of the psychedelic experience. A 2018 study found that the patient’s experience of music during psilocybin-assisted therapy was associated with the occurrence of MTEs (Kaelen et al., 2018), specifically the degree to which the patient ‘liked’ the music positively predicted MTEs. In addition, the immediate environment in which the dosing session occurs is associated with the nature of the experience. Use of 5-MeO-DMT in supportive and structured contexts, that is, contexts with structured procedures for dosing and administration (e.g. preparatory and integration sessions) are more conducive to higher ratings of MTEs and lower ratings of challenging experiences, as opposed to non-structured contexts (e.g. at home or a festival) (Sepeda et al., 2019). However, that is not to conclude that all research settings are favorable. A previous study showed that a significantly higher number of challenging reactions were evoked by placing participants in a positron emission tomography scanner (Studerus et al., 2012). Thus, setting has an established influence on the quality of the psychedelic experience, including the occurrence of MTEs and must be considered in the context of optimizing therapeutic efficacy.
Neural correlates of mystical experiences
While substantial research efforts have been directed toward identifying moderators of MTEs, there is no widely agreed-upon consensus on how it is they occur. In fact, whether MTEs can be reduced to neurobiological processes remains an open question. The nature of MTEs is highly subjective; a deeply felt sense of unity, sacredness and ineffability is difficult to reduce to biological mechanisms. In addition, not all experiences with psychedelics are of the mystical subtype and MTEs can occur independently of drugs (e.g. through meditative states). Thus, psychedelics are neither necessary nor sufficient for producing such experiences. However, there is still value in understanding physiological processes that relate to MTEs. Various neural correlates of such experiences have been measured. At a neuropharmacological level, 5-HT2A receptor agonism is considered to play, at the very least, an initiatory role in the brain processes that relate to MTEs (Barrett and Griffiths, 2017). At the network level, the DMN is heavily involved in self-referential processing and is thought to be potentially related to MTEs. The nodes of the DMN are akin to a network of brain regions involved in maintaining a sense of self in time and space. The experience of unity is a hallmark of MTEs and involves reduced self-referential processing. Barrett and Griffiths (2017) proposed a neural model of MTEs comprising of the following criteria: (1) reduced activity in the medial and lateral nodes of the DMN and (2) reduced long-distance cortical communication. Furthermore, the medial and lateral nodes are thought to be associated with reduced self-referential processing and losing sense of time and space, respectively (Barrett and Griffiths, 2017). This suggests that modulation of DMN connectivity may partly account for the occurrence of MTEs. However, the literature presents a mixed picture; while some studies report a decrease in DMN connectivity during MTEs (Carhart-Harris et al., 2017), others show an increase in connectivity from baseline to post-session that is linked to better therapeutic outcomes (Copa et al., 2024; Mertens et al., 2020; Ruban and Kołodziej, 2018). This inconsistency highlights the complexity of pinpointing a clear neural basis for MTEs and indicates that DMN connectivity might play a nuanced or context-dependent role.
Limitations and implications
The systematic review is subject to several limitations. First, the wide scope of the review makes it difficult to draw concrete conclusions. There were several mental disorders included in the review, along with different types and doses of classic psychedelics and multiple study designs. Aside from open-label trials and RCTs, observational studies, qualitative analyses, semi-structured interviews and retrospective online surveys were included. Retrospective surveys in particular are subject to recall bias or misclassification bias, thereby limiting the quality of the data. Even within the open-label trials and RCTs included, the specific study designs were highly heterogeneous. Among trials that utilized the same psychedelic drug to treat the same mental disorder, the number of dosing sessions varied between one and four and some studies employed adjunctive psychotherapy while others did not. This heterogeneity makes it difficult to compare the obtained results. In addition, many of the included trials were based on the same study population, which limits the ability to generalize the findings. The dosage, however, remained quite stable across studies with high doses being employed for most trials.
The sample sizes of the studies included in the review were generally small. Thus, with insufficient power, only factors with a high effect size may be detected and more subtle factors may be lost. A critical limitation is the predominance of short-term studies, with the majority of findings derived from follow-up periods of ⩽1 month. This not only constrains the ability to assess sustained treatment effects but also raises concerns about the generalizability of results to long-term clinical practice. Another major shortcoming of the review was the inadequacy of the follow-up of patients. Studies examining PAT for SUD or depression and anxiety in patients with life-threatening cancer generally had thorough follow-up assessments beyond 6 months, however, studies investigating MDD (both TRD and non-TRD), which made up the majority of the review, generally did not assess primary and secondary outcomes beyond 4 weeks after the dosing session(s). This limits the ability to predict true long-term treatment outcomes. Furthermore, discrepancies in risk of bias ratings across systematic reviews illustrate the subjectivity involved in applying established assessment tools, which represents an additional limitation. Future studies should design trials to include adequate follow-up periods to assess whether therapeutic effects are sustained beyond a few weeks.
Finally, another important consideration to make is the distinction between the most studied versus the most robust predictive factors. There is a tendency for subsequent studies to examine associations that have already been established. The sheer frequency of investigation between MEQ scores and clinical improvement may well partially account for the overwhelming correlation. This also raises the possibility of circular reasoning in mystical experience research, whereby constructs like MTEs are defined and measured in ways that inherently align with desired therapeutic outcomes, potentially reinforcing the same association in future work. While this is important, it may preclude the detection of new factors or misrepresent the importance of a single factor. Thus, Figure 2 displaying the frequency of predictive factors should be interpreted with this in mind. It should also be acknowledged that the broader psychedelic research field is subject to a likely publication bias toward positive or favorable findings, which may inflate the apparent strength or consistency of certain predictors in the literature.
Despite these limitations, the findings of this systematic review offer valuable insights for enhancing future treatments. To promote a higher response in PAT, it is crucial to consider data-driven patient selection and thorough patient preparation. Screening patients for baseline characteristics such as the ability to surrender and having a specific, concrete treatment intention, may significantly enhance the effectiveness of PAT. By integrating these predictive factors into the screening process, clinicians can better determine who is most likely to benefit from psychedelic therapy. Additionally, focusing on these aspects in preparatory sessions can help patients achieve the kind of subjective experiences that are associated with positive therapeutic outcomes. Furthermore, optimizing set and setting, including appropriate music selection, dose calibration and development of a strong therapeutic alliance may facilitate beneficial subjective experiences such as MTEs and emotional breakthrough. By addressing these factors, future PAT can be more effectively tailored to individual patients, potentially leading to higher overall treatment success.
Conclusion
While PAT has demonstrated substantial efficacy across a range of mental disorders, patient response remains variable. This study systematically reviewed the literature to identify factors that may predict therapeutic response to PAT. Across 54 studies, we found that the most frequently identified variable predicting clinical outcomes was the intensity of the acute psychedelic experience, particularly MTEs. However, the consistency of this association varied slightly across disorders and follow-up durations. Factors related to set, setting and dose were found to influence the likelihood and intensity of MTEs, highlighting their relevance in shaping therapeutic outcomes. Future research should incorporate longer-term follow-up and expand the range of predictors assessed to develop a more robust and generalizable understanding of PAT response.
Supplemental Material
sj-docx-1-jop-10.1177_02698811251389581 – Supplemental material for Predictors of therapeutic response to psychedelic-assisted therapy: A systematic review
Supplemental material, sj-docx-1-jop-10.1177_02698811251389581 for Predictors of therapeutic response to psychedelic-assisted therapy: A systematic review by Grace Viljoen, Henrik Walter, Antonia Bendau, Michael Koslowski and Felix Betzler in Journal of Psychopharmacology
Supplemental Material
sj-docx-2-jop-10.1177_02698811251389581 – Supplemental material for Predictors of therapeutic response to psychedelic-assisted therapy: A systematic review
Supplemental material, sj-docx-2-jop-10.1177_02698811251389581 for Predictors of therapeutic response to psychedelic-assisted therapy: A systematic review by Grace Viljoen, Henrik Walter, Antonia Bendau, Michael Koslowski and Felix Betzler in Journal of Psychopharmacology
Supplemental Material
sj-docx-3-jop-10.1177_02698811251389581 – Supplemental material for Predictors of therapeutic response to psychedelic-assisted therapy: A systematic review
Supplemental material, sj-docx-3-jop-10.1177_02698811251389581 for Predictors of therapeutic response to psychedelic-assisted therapy: A systematic review by Grace Viljoen, Henrik Walter, Antonia Bendau, Michael Koslowski and Felix Betzler in Journal of Psychopharmacology
Footnotes
Author contributions
Conceptualization: all authors; methodology: all authors; investigation: GV and FB; Writing—original draft: GV; writing—review and editing: HW, AB, FB and MK; project administration: FB and MK.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Declaration of conflicting interests
The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: FB has received honoraria as a lecturer, served as an advisory board member and/or received research funding from Takeda and Medice Pharmaceuticals.
Data availability statement
Data will be provided upon request.
AI statement
This manuscript was edited for language using ChatGPT. The authors take full responsibility for the content.
Supplemental material
Supplemental material for this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
