The Challenging Experience Questionnaire: Characterization of challenging experiences with psilocybin mushrooms
This paper develops and validates the 'Challenging Experience Questionnaire' that finds seven different underlying factors (grief, fear, death, insanity, isolation, physical distress, and paranoia).
Authors
- Barrett, F. S.
- Bradstreet, M. P.
- Griffiths, R. R.
Published
Abstract
Acute adverse psychological reactions to classic hallucinogens (“bad trips” or “challenging experiences”), while usually benign with proper screening, preparation, and support in controlled settings, remain a safety concern in uncontrolled settings (such as illicit use contexts). Anecdotal and case reports suggest potential adverse acute symptoms including affective (panic, depressed mood), cognitive (confusion, feelings of losing sanity), and somatic (nausea, heart palpitation) symptoms. Responses to items from several hallucinogen-sensitive questionnaires (Hallucinogen Rating Scale, the States of Consciousness Questionnaire, and the Five-Dimensional Altered States of Consciousness questionnaire) in an Internet survey of challenging experiences with the classic hallucinogen psilocybin were used to construct and validate a Challenging Experience Questionnaire. The stand-alone Challenging Experience Questionnaire was then validated in a separate sample. Seven Challenging Experience Questionnaire factors (grief, fear, death, insanity, isolation, physical distress, and paranoia) provide a phenomenological profile of challenging aspects of experiences with psilocybin. Factor scores were associated with difficulty, meaningfulness, spiritual significance, and change in well-being attributed to the challenging experiences. The factor structure did not differ based on gender or prior struggle with anxiety or depression. The Challenging Experience Questionnaire provides a basis for future investigation of predictors and outcomes of challenging experiences with classic hallucinogens.
Research Summary of 'The Challenging Experience Questionnaire: Characterization of challenging experiences with psilocybin mushrooms'
Introduction
Barrett and colleagues situate this work within the renewed empirical interest in classic psychedelic hallucinogens (for example psilocybin, LSD, mescaline, and DMT) and note that, alongside therapeutic and neuroscientific inquiry into positive and mystical-type effects, acute psychologically adverse reactions—commonly called "bad trips" or challenging experiences—remain a documented phenomenon. Prior clinical and research reports describe a heterogeneous set of acute adverse phenomena that can include affective symptoms (panic, depressed mood), cognitive disturbances (confusion, derealisation, fears of losing sanity), perceptual and delusional phenomena (paranoia, frightening hallucinations), and somatic responses (nausea, sympathetic arousal). Existing widely used scales (for example the Hallucinogen Rating Scale, Dittrich's Altered States of Consciousness questionnaires) capture some of these elements but lack a concise set of subscales specifically designed to parse distinct dimensions of challenging experiences. The present papers aimed to construct and validate a dedicated instrument, the Challenging Experience Questionnaire (CEQ), to measure discrete dimensions of challenging experiences occasioned by psilocybin. Study 1 used secondary analysis of an Internet survey to derive and replicate a factor structure from pooled items taken from the HRS, 5DASC and the States of Consciousness Questionnaire (SOCQ), test factorial invariance across sex and prior psychiatric struggle, and relate CEQ factors to global impact ratings (difficulty, meaningfulness, spiritual significance, and change in well-being). Study 2 then tested the CEQ as a stand-alone 26-item instrument in an independent online sample to replicate the factor structure and validate associations with impact ratings.
Methods
Overall approach: The investigators developed the CEQ using a two-study pipeline based on online retrospective survey data. Study 1 performed psychometric scale construction and cross-validation within a large pooled Internet sample by selecting items from three existing instruments; Study 2 tested the resulting 26-item CEQ as a stand-alone questionnaire in a separate online sample. Study 1 sample and eligibility: The original "Bad Trip Survey" collected 2085 responses via Internet recruitment (websites such as Erowid, email, word of mouth). Inclusion required English fluency, having ingested an active dose of psilocybin mushrooms that produced moderate to strong effects, reporting a difficult/challenging experience with psilocybin, being aged 18+ at survey completion and aged 18–70 at the time of the reported experience. After exclusions (research-context experiences, experiences of another person, co-ingested substances, and other validity concerns) 1993 participants remained for analysis. The analytical sample was stratified into two demographically matched sub-samples for exploratory (n=833) and confirmatory (n=934) analyses; 226 observations were removed during stratification yielding a final analysis sample of 1767. Item pool and measures: The initial item pool comprised 64 items judged to unambiguously assess potentially challenging aspects of experience, drawn from the Hallucinogen Rating Scale (HRS; 27 items retained), the Five-Dimensional Altered States of Consciousness questionnaire (5DASC; 13 items retained from its impaired cognition/control and anxiety subscales), and 24 "distractor" items from the SOCQ (excluding the MEQ43 mystical items). Additional items assessed overall impact (difficulty, meaningfulness, spiritual significance, change in well-being) and prior struggle with psychiatric disorders; demographic variables and past hallucinogen use were also collected. Response formats originally differed across instruments; for Study 1 items were analysed in their original forms. Scale construction and exploratory analyses: In the exploratory stratum the researchers inspected distributional properties and removed items with restricted range or strong non-normality. A mixed correlation matrix appropriate for mixed item types was computed (polychoric, tetrachoric, Pearson correlations as needed). Exploratory factor analysis (maximum likelihood extraction, oblimin rotation) used parallel analysis and scree plots to select factor number. Items with loadings <0.4 were discarded and iterative item-total and reliability (Cronbach's alpha or McDonald's omega) checks guided item removal, with finalisation of a reduced item set and factor solution. Model replication and factorial invariance: The exploratory solution was tested by confirmatory factor analysis (CFA) in the second stratum. Model fit was evaluated using CFI, SRMR and RMSEA, with conventional thresholds (CFI>0.90; SRMR and RMSEA<0.10 cited by the authors). Factorial invariance—testing whether the factor structure and parameters are comparable across groups—was assessed via multiple-group CFA across sex and across prior struggle with psychiatric disorder. The authors sequentially tested configural, weak (equal loadings), strong (equal intercepts) and strict (equal residuals) invariance, monitoring changes in CFI, RMSEA and SRMR to detect non-invariance. Structural modelling and convergent validity: A structural equation model (SEM) for the entire sample regressed reported overall impact ratings (difficulty, meaningfulness, spiritual significance, change in well-being) onto latent CEQ factor scores to examine external validity and the relationships of CEQ dimensions to subjective outcomes. Study 2 (stand-alone CEQ): The 26-item CEQ (24 items from Study 1 plus two paranoia items identified in Study 1) was adapted to a single response format (the six-point SOCQ-style scale: 0 none to 5 extreme) and minor prose edits harmonised item wording. The stand-alone CEQ was administered in a separate online sample together with the same overall impact measures and demographics. Analyses mirrored Study 1: KMO and Bartlett's tests assessed suitability for factor analysis, CFA and factorial invariance tested replication against Study 1, and SEM regressed impact ratings on CEQ factor scores. The extracted text does not provide a full protocol for recruitment or compensation in Study 2 beyond noting demographic differences versus Study 1.
Results
Study 1 exploratory findings: From the initial 64-item pool, distributional inspection and iterative item trimming reduced the set to 24 items after removing items with low loadings or poor item-total correlations. Parallel analysis and scree plot supported a six-factor solution that the investigators labelled physiological distress, grief, fear, insanity, isolation and death. This six-factor solution explained a cumulative 67% of variance. Measures of sampling adequacy supported factorability (Kaiser-Meyer-Olkin KMO=0.92; Bartlett's test χ2=44,899, df=25, p<0.0001). Addition of paranoia: Although a distinct paranoia factor did not emerge in the initial exploratory modelling, two SOCQ items explicitly referencing paranoia (feeling that people were plotting against you; experience of antagonism toward people around you) existed in the item pool. When CFA was performed in the confirmatory stratum the model fit improved by including those two items as a seventh factor for paranoia, yielding a 26-item, seven-factor model. Confirmatory analyses and model fit: CFA in the confirmatory stratum first tested the 24-item six-factor model (RMSEA=0.070, 90% CI 0.067–0.074; SRMR=0.054; CFI=0.913), considered acceptable by the authors. Adding the two paranoia items to form a 26-item seven-factor model produced similar fit (RMSEA=0.066, 90% CI 0.063–0.070; SRMR=0.052; CFI=0.912). Factor loadings, intercorrelations and reliabilities for the six original factors remained similar after adding the paranoia factor. Factorial invariance: Multiple-group CFA indicated negligible changes in fit indices across levels of invariance for both sex and prior struggle with psychiatric disorder, which the authors interpret as evidence of strict factorial invariance. This suggests the CEQ measures the same constructs across men and women and across participants with and without self-reported prior psychiatric struggles; the extraction does not provide detailed numeric changes but reports that thresholds for non-invariance were not exceeded. Associations with overall impact (Study 1 SEM): The SEM regressing global impact ratings on CEQ latent factors demonstrated good fit (CFI=0.915; RMSEA=0.060, 90% CI 0.058–0.062; SRMR=0.045). Key associations reported were: the death factor was associated with a decrease in well-being and positively associated with other impact ratings (difficulty, meaningfulness, spiritual significance); fear scores were positively associated with rated difficulty and with change in well-being, and negatively associated with spiritual significance and meaningfulness; isolation scores were positively associated with increased well-being but negatively with meaningfulness and spiritual significance; insanity and grief scores were positively associated with difficulty, meaningfulness and spiritual significance; physical distress was positively associated with spiritual significance and associated with decreased well-being. Paranoia scores did not associate with impact ratings in Study 1. Study 2 stand-alone CEQ: The independent sample for Study 2 differed demographically from Study 1: participants were older, more educated, more racially/ethnically diverse, had a higher proportion of females, and reported more prior hallucinogen use. Sampling adequacy remained acceptable (KMO=0.89; Bartlett's χ2=538.01, df=26, p<0.0001). Strict factorial invariance between Study 1 and Study 2 was supported, indicating replication of the 26-item factor structure. Compared with Study 1, Study 2 showed higher mean factor scores for fear and physical distress and lower scores for insanity and paranoia. Study 2 SEM and associations: The SEM in Study 2 also fit well (CFI=0.912; RMSEA=0.064, 90% CI 0.061–0.067; SRMR=0.048). Results broadly replicated some Study 1 associations: the death factor was negatively associated with change in well-being and positively associated with other impact ratings; fear was positively associated with difficulty and change in well-being and negatively associated with spiritual significance; insanity was positively associated with rated difficulty and meaningfulness. Some associations differed between studies: for example, paranoia was positively associated with change in well-being in Study 2 (but not Study 1), and isolation did not predict impact ratings in Study 2. The authors note inconsistencies across studies and suggest they may reflect sample differences or uncontrolled variables such as dose or set and setting. Other findings: Across Study 1, participants endorsing prior struggle with psychiatric disorder had higher CEQ factor scores for all factors except paranoia, whereas factor scores generally did not differ by sex except for slightly higher fear scores in women. The final CEQ thus comprises seven factors—grief, fear, death, insanity, isolation, physical distress, and paranoia—capturing affective, cognitive/affective and somatic dimensions of challenging psilocybin experiences.
Discussion
Barrett and colleagues interpret their findings as indicating that the CEQ is a psychometrically coherent instrument that captures multiple, qualitatively distinct dimensions of challenging experiences occasioned by psilocybin. The seven-factor structure displays face validity and aligns with prior clinical descriptions of acute adverse reactions, encompassing affective domains (fear, grief), physiological distress, cognitive/affective dimensions (isolation, paranoia, feelings of insanity) and a subjective ‘‘death’’ experience. Replication of the factor structure across independent samples and evidence for strict factorial invariance across sex and prior psychiatric struggle are emphasised as support for both internal and external validity. The relationship between CEQ factors and ratings of overall impact was complex. Scores on fear, grief, insanity and death were consistently positively associated with participants' ratings of how difficult the experience was, suggesting these may constitute the core dimensions of a challenging experience. Yet some dimensions—particularly death and fear—showed divergent associations with meaningfulness, spiritual significance and change in well-being: the death factor correlated positively with rated meaningfulness and spiritual significance but negatively with change in well-being, while fear sometimes predicted increased well-being despite reducing ratings of meaningfulness and spiritual significance. The authors suggest that certain facets of extreme experience (for example profound experiences of death or loss of sanity) may, in some cases, co-occur with spiritually meaningful appraisals while nonetheless being associated with poorer immediate well-being outcomes. Limitations and uncertainties are acknowledged. The data are retrospective and self-selected, recruited largely via websites frequented by individuals with favourable attitudes to psychedelics, so respondent bias is likely. Precise psilocybin dose cannot be known from retrospective reports of mushroom ingestion, and set-and-setting variables were not controlled; these unmeasured factors may account for between-sample differences and inconsistencies in predictor–outcome associations. The crude dichotomy used for prior psychiatric struggle does not differentiate diagnostic categories, and self-report of psychiatric history is imperfect. The paranoia subscale is noted to be a relatively crude measure of clinical paranoia because one of its two items taps antagonism rather than unfounded fear of harm. The authors therefore recommend cautious interpretation of some scales and call for prospective, controlled studies to clarify predictors and consequences of the CEQ dimensions. Implications and future directions: The CEQ provides a structured tool for research on acute challenging experiences, enabling finer-grained study of predictors (for example individual traits, dose, set and setting) and downstream outcomes, including possible differential lasting effects depending on the dominant challenge dimension. The instrument may assist both basic neuroscientific studies—where distinguishing facets of challenge could aid interpretation of imaging or behavioural findings—and clinical trials, where better characterisation of adverse acute phenomena could inform screening, preparation and support procedures to optimise safety and therapeutic benefit.
Conclusion
The authors conclude that the CEQ, validated across two independent Internet samples, is a useful instrument for characterising psychologically difficult aspects of experiences occasioned by psilocybin and likely applicable to other classic hallucinogens. They argue that improved measurement of challenging experiences can advance understanding of both the psychological and neural mechanisms of these drugs and help optimise their therapeutic application, while noting the need for further validation in controlled prospective studies.
View full paper sections
INTRODUCTION
Classic psychedelic hallucinogens, such as psilocybin, lysergic acid diethylamide (LSD), mescaline, and dimethyltryptamine (DMT; contained in the sacramental beverage ayahuasca) share a primary pharmacological site of action at the serotonin 2a (5HT2a) receptor, and display striking similarities in their subjective effects. The first era of modern human research on psychedelics, from roughly the 1950s to the 1970s, capitalized on the powerful effects of psychedelics on perception, emotions, and consciousness by investigating the therapeutic and psychotomimetic properties of these substances. The recent resurgence of empirical research on psychedelics has revived their investigation as therapeutic agentsand tools to understand psychosis. Recent studies have also examined mystical-typeor spiritualexperiences, changes in self-referential processing or "ego dissolution", altered social processing, visual imagery, changes in emotional experience, personality change, positive mental health outcomes, and enduring positive changes in attitudes, mood and behavioroccasioned by these compounds. However, both recent experimental reportsand past clinical reportsindicate that challenging psychological experiences during the acute effects of psychedelics are not uncommon.
PHENOMENOLOGY OF CHALLENGING PSYCHEDELIC EXPERIENCES
individuals and more than 25,000 separate drug administrations) noted instances of fear, delusions, dissociation, depersonalization, and sympathetic nervous system responses during acute effects of these drugs. Reports of psychosis lasting more than 48 h were exceedingly rare (<0.001% of healthy individuals, and <0.002% of individuals undergoing therapy), however this summary must be viewed tentatively, given numerous limitations including, for example, that none of the reporting investigators had conducted follow-up with their patients. A subsequent review, integrating both clinical literature (including emergency department reports) and research literature, presented a summary of acute adverse effects of classic hallucinogens that included frightening illusions and hallucinations, overwhelming anxiety or panic, confusion, aggression and possible violence, depression with suicidal ideation, gestures, or attempts, and fear to the point of paranoid delusions. When encountered, these symptoms typically responded to verbal reassurance and typically subsided within 48 h, often spontaneously. If severe, these symptoms typically responded to medication. However, the review noted that emergency department reports (e.g.are limited by uncertainty of substance ingested and unknown mental state of the patient prior to putative ingestion of a psychoactive substance. Recent research reports that address challenging aspects of experiences with psychedelics (hereby referred to as "challenging experiences" for brevity) are more specific with regards to substance and dose administered. Approximately 30% of participants in each of three highly controlled experimental studies (total sample: 69 participants and 204 sessions -involving a high dose of psilocybin (0.429 mg/kg psilocybin) experienced marked periods of anxiety or fear, while between 17-39% experienced paranoia. Only three of 42 (7.1%) participants receiving up to a moderate dose of psilocybin (0.315 mg/kg psilocybin) in a separate series of studies reported marked periods of anxiety, fear, or dysphoria, while only one (2.4%) of those participants reported prolonged symptoms. Clinical and experimental literature on psychedelics suggests a possible profile of challenging experiences that includes the following categories of experience: fear or panic, paranoia, sadness or depressed mood, anger, cognitive effects (e.g. confusion, loss of ego, loss of sanity, delusions, dissociation, depersonalization), perceptual effects (e.g. illusions), and physiological symptoms (e.g. increased heart rate, nausea/emesis, sympathetic system response). These categories are supported both by reviews of the safety of hallucinogensand published clinical guidelines for the assessment and management of adverse reactions to hallucinogens. While these categories of experience appear in aggregate in the literature, they may not all appear in any single instance of a challenging experience.
CURRENT ASSESSMENTS OF CHALLENGING EXPERIENCES
The two most widely used questionnaire instruments assessing the subjective effects of classic hallucinogens are the Hallucinogen Rating Scale (HRS)and the various forms of Dittrich's Altered States of Consciousness questionnaires, including the Altered States of Consciousness Rating Scale (OAV) and the Five-Dimensional Altered States of Consciousness Rating Scale (5DASC). While some of the sub-scales of these instruments, and many items within these instruments, assess some aspect of cognitive, emotional, or physiological experiences that may be challenging, no questionnaire includes a comprehensive list of clearly defined sub-scales that separately measure the categories of challenging experience that are suggested by previous literature. The HRS was developed explicitly as a means to quantify the subjective effects of classic hallucinogens. Of the six sub-scales of the HRS (i.e. affect, cognition, intensity, perception, somaesthesia, and volition), one might hypothesize that the affect, cognition, and somaesthesia subscales might be most sensitive to challenging experiences. However, although the affect scale includes items that assess negative affect (e.g. anxious, frightened, panic), it also includes items that assess positive affect (e.g. euphoria, love, awe). Items in the cognition and somaesthesia scales include items that assess possibly challenging cognitions and sensations (e.g. "sense of chaos" or "nausea"), but they also include items that assess general but not necessarily "challenging" cognitions and sensations (e.g. "new thoughts" or "change in body temperature"). Thus, scores on the six scales of the HRS do not distinguish between very strong challenging experiences and very strong non-challenging experiences. The OAV and 5DASC (and the preceding Abnormal Mental States (APZ) questionnaire) have been widely used in behavioral and neuroimaging research of pharmacologicallyas well as non-pharmacologically induced altered states of consciousness. The OAV sub-scale "dread of ego dissolution" (DED) covers a wide range of negative experiences, and is generally considered an overall "bad trip" scale. This meta-scale of possible negative effects covers many (e.g. panic, loss of ego/control, feelings of insanity) but not all (e.g. sadness/grief/ depression) possible categories of challenging experiences. The DED scale also averages responses from a number of proposed categories of experience (panic, loss of ego, insanity) rather than giving an individual score for each.revealed a rescoring of the 5DASC that includes a separate scale for impaired control and cognition, and for anxiety. While these represent psychometrically justifiable subscales, these two sub-scales do not address shortcomings of the DED scale (e.g. they do not address the wide range of potential dimensions of challenging experience that are suggested by previous literature). Many other instruments were either used or developed to assess various aspects of subjective experience of hallucinogens, including the Linton-Langs questionnaire, the Abramson questionnaire, the Phenomenology of Consciousness Inventory, the Minnesota Multiphasic Personality Inventory, the States of Consciousness Questionnaire, and the Addiction Research Center Inventory. These scales all suffer from at least one of the shortcomings of the 5DASC and HRS, namely: lack of specificity (i.e. lack of separable or independent scales that measure fine-grained facets of a challenging experience), lack of content coverage (missing a proposed dimension of challenging experience), or lack of sensitivity (i.e. not sensitive to the degree of challenge in an experience) to challenging experiences.
WHY AN ASSESSMENT OF DIMENSIONS OF CHALLENGING EXPERIENCE IS NEEDED
With the recognition of the importance of set (participant psychological state) and setting (interpersonal and physical environment), acute adverse effects of psychedelics may generally be minimized and successfully managed in research contexts. Guidelines have been described to minimize adverse reactions in controlled settings, and incidence of lasting symptoms (including psychosis) related to administration of classic hallucinogens in controlled settings is quite low. However, challenging experiences can still occur in the presence of substantially controlled and supportive conditions, and occurrence of acute and persisting adverse effects of psychedelics may be greater in uncontrolled settings, or with unscreened individualsLittle is known regarding the interdependence, causes, and consequences of any one of the individual categories of challenging experience suggested by the clinical and prior research literature. It is of interest to basic science and clinical application of classic hallucinogens to understand, for example, whether a challenging experience primarily involving panic and physiological distress might lead to persisting effects that differ from those that persist after a challenging experience primarily involving depression or confusion. If that is the case, it would be of value to know whether there are reliable indicators that could be used to predict whether an individual is more likely to experience panic or grief. Various facets of challenging experience may generally covary, but they may not necessarily be interdependent or collinear. Different facets of challenging experiences may lead to different outcomes. Psychedelics offer a tool for studying the biological basis of emotional experiences using modern neuroscientific tools such as brain imaging. Anger, fear, and sadness have putatively different neural substrates as well as behavioral implications. To the extent that challenging experiences are encountered in brain imaging studies, it may be necessary to parse interindividual differences in challenging experience during acute effects of psychedelics in order to be able to fully appreciate the neural basis of psychedelic experiences. The field is lacking an empirically derived measurement instrument that can assess individual differences in challenging experiences, provide a "profile" of challenging experiences, and be applied to study predictors and persisting effects of challenging experiences. Such a fine-grained knowledge regarding the profile of challenging experiences and the consequences of these experiences will be of value both to optimize our understanding of the neurobiology of psychedelics and to optimize their use in treatment settings.
AIM OF THE CURRENT STUDIES
In the following two studies, a Challenging Experience Questionnaire (CEQ) was constructed. Study 1 consisted of a psychometric analysis of responses to pooled items from a series of questionnaires (the 5DASC, HRS, and an additional measure, the States of Consciousness Questionnaire (SOCQ)) that have been used in an online survey of challenging experiences with psilocybinto establish, replicate, and validate the factor structure and factor scores of the CEQ in a pair of stratified sub-samples. In Study 2, administration of the CEQ as a stand-alone instrument in an online survey demonstrated both replication of the CEQ factor structure and validation of the CEQ factor scores.
STUDY 1
This study is a secondary analysis of data from an online survey of challenging experiences with psilocybin. Participants in this survey completed a series of questionnaires in reference to a self-identified challenging experience. The survey sample was stratified into two demographically matched groups. Though the HRS, 5DASC, and SOCQ do not have sub-scales that measure all facets of challenging experience suggested by the clinical literature, they all have individual items that assess these facets. Thus, the combined items from the HRS, 5DASC, and SOCQ that address potential aspects of challenging experience form a desirable initial item pool for the development of the CEQ. Exploratory factor analysis was applied to items from the HRS, 5DASC, and SOCQ in the first stratum that could reasonably assess an aspect of challenging experience. The resulting factor structure was validated using confirmatory factor analysis in the second stratum. Structural equation modeling was applied to the entire sample to demonstrate factorial invariance of the instrument across levels of gender. Clinical literature indicates that lasting psychosis related to challenging experiences may be present in those who have a personal or family history of psychological difficulty. Therefore, factorial invariance of the instrument was also assessed across levels of struggle with a psychiatric disorder that preceded the reported challenging experience. Ratings of the meaningfulness, spiritual significance, and difficulty of the experience, as well as ratings of the impact of the experience on a person's overall well-being, have been used in the past to assess the overall impact of psilocybin experiences on an individual. Therefore, ratings of the overall impact of the challenging experience in this sample were regressed on CEQ factor scores to assess the relationship between these outcome measures and aspects of challenging experience. A description of measures analyzed in the current report is provided below. A more detailed description of the complete methods, measures, and initial findings of the online survey can be found in the original report.
PARTICIPANTS
The Bad Trip Surveywas completed by 2085 participants. Participants were recruited via Internet advertisements, email invitation, and word of mouth. An Internet link to the survey was posted on websites, such as Erowid (an online information library on psychoactive substances, www. erowid.com), that are frequented by individuals interested in hallucinogens. Participants were not provided compensation for their responses. Inclusion criteria consisted of reading, writing, and speaking fluency in English, having ingested an active dose of psilocybin mushrooms that produced moderate to strong psychoactive effects, having had a difficult or challenging experience (i.e. "bad trip") after ingesting psilocybin mushrooms, age of 18 years or older at the time of completing the survey, and age of between 18-70 years old at the time of the reported challenging experience. Participants who did not meet inclusion criteria after providing demographic information were directed to an abbreviated version of the survey, and their data were not included in the analysis. This approach was taken to obscure to participants the fact that they were excluded, and therefore discourage them from re-taking the survey with false responses in an attempt to be included. Participants were also asked to refrain from completing the survey more than once. Data from completed participants were excluded if they reported that their challenging experience was (a) encountered in the context of a research study, (b) experienced by another person, or (c) attributed to another substance in addition to psilocybin. Participants were also excluded if free-response comments provided at the end of the survey raised concerns about the validity of their reports. Data for 92 completing participants were excluded, yielding a sample of 1993 participants.
MEASURES
Participants completed an online survey consisting of the belowreferenced questionnaires (the HRS, 5-DASC, and SOCQ) among a series of other questions regarding demographics, substance use behaviors, and details of the participant's reported challenging experience. A more detailed description of the full complement of measures administered in the online survey can be found in the original report. We identified 64 items from the HRS, 5-DASC, and SOCQ (listed in Supplementary Material, Appendix 2) that unambiguously assessed a challenging aspect of experience with classic hallucinogens, and treated these 64 items, as worded and responded to in their original form in the HRS, 5-DASC, and SOCQ, as an initial item pool for the construction of the initial form of the CEQ. Additional items (including questions regarding the overall impact of the challenging experience, and previous struggle with a psychiatric disorder) were used in regression and factorial invariance analyses, respectively, after the generation of the CEQ. Finally, demographic variables were assessed and used to describe the sample and also to generate two demographically matched strata for separate analyses within this study.
HRS.
The HRS is a 99-item instrument that is rated on a fivepoint scale (1 -"not at all", 2 -"slightly", 3 -"moderately", 4 -" very much", 5 -"extremely"). It assesses the greatest degree to which a respondent encountered different subjective effects during the course of an experience with a specified drug. It consists of six subscales assessing general dimensions of subjective experience (intensity, somaesthesia, affect, perception, cognition, and volition). Twenty-seven items (spanning all six scales of the HRS) that were judged by the authors to assess a potentially challenging aspect of experience with classic hallucinogens were retained for the initial item pool for the CEQ. 5DASC. The 5DASCis the latest version of the Altered States of Consciousness (APZ) questionnaire, which was designed to assess subjective aspects of a wide range of altered states of consciousness, including but not limited to those occasioned by classic hallucinogens. The 5DASC consists of 94 items that are rated on a 20-point rating scale with two anchors: 1 -"No, not more than usually" and 20 -"Yes, much more than usually". A recent psychometric analysis identified 11 plausible sub-scales of the 5DASC, which include the impaired cognition and control (ICC), and anxiety (ANX) scales. The 13 items of the 5DASC that constitute the ICC and ANX sub-scales were retained for the initial item pool for the CEQ. While the 5DASC was originally developed in German and has not been empirically validated in English, this does not undermine the use of items from this questionnaire in the current study, as these items contribute to an initial item pool for validation of a new questionnaire (the CEQ).
SOCQ.
The SOCQ is a 100-item questionnaire that is rated on a six-point scale (0 -"none; not at all", 1 -"so slight cannot decide", 2 -"slight", 3 -"moderate", 4 -"strong (equivalent in degree to any previous strong experience or expectation of this description)", 5 -"extreme (more than ever before in my life and stronger than 4)"). The SOCQ contains 43 items from the Mystical Experiences Questionnaire (MEQ43), which was developed to assess several domains of mystical experience. The other 57 items of the SOCQ were based on a wide range of possible subjective effects of classic hallucinogens that were suggested by a sample of clinicians, but these items have not previously been analyzed, and they are typically treated as distractor items. Twenty-four of these distractor items were identified and retained for the initial item pool for the CEQ. These 24 distractor items assess potentially challenging aspects of experiences with classic hallucinogens (such as emotional, social, and physical discomfort, pain, and suffering, disorientation, ego loss, loss of perception of time, isolation, and confusion). No items from the MEQ43 were included in the present analysis. While these items have not been a part of an independently validated questionnaire, this does not undermine the use of these items to contribute to an initial item pool for development and validation of the CEQ. Questions regarding the overall impact of the challenging experience. These questions comprise a set of items that have been administered in laboratory studies of the acute and persisting effects of psilocybin. Participants were asked to report on the meaningfulness and difficulty of their reported psilocybin experience using the following response options: "no more than routine, everyday experiences", "similar to (difficult or challenging/meaningful) experiences that occur on average once a week", "similar to (difficult or challenging/meaningful) experiences that occur on average once a month", "similar to (difficult or challenging/ meaningful) experiences that occur on average once a year", "similar to (difficult or challenging/meaningful) experiences that occur on average once every 5 years", "among the top 10 most (difficult or challenging/meaningful) experiences of my life", "among the top five most (difficult or challenging/meaningful) experiences of my life", and "the single most (difficult or challenging/meaningful) experience of my life". Participants indicated the degree to which the experience was spiritually significant to them, using the following rating scale: "not at all", "slightly", "moderately", "very much", "among the top five most spiritually significant experiences of my life", and "the single most spiritually significant experience of my life". Participants also responded to a question inquiring about the effect of the challenging experience on their well-being or life satisfaction ("Do you believe that the experience and your contemplation of that experience have led to a change in your current sense of personal well-being or life satisfaction?") using the following rating scale: "Increased very much (+3)", "Increased moderately (+2)", "Increased slightly (+1)", "No change (0)", "Decreased slightly (-1)", "Decreased moderately (-2)", "Decreased very much (-3)". These inquiries conform to previous methods of measuring participants' assessment of the overall impact of a psilocybin session. Meaningfulness, spirituality, difficulty, and well-being were regressed on latent variable (i.e. factor) scores derived from the CEQ. Previous struggle with a psychiatric disorder. Participants were asked to indicate whether they had struggled with a psychiatric disorder at some point before their reported challenging experience. Participants were then able to separately endorse having struggled with the following disorders ("Please select all psychiatric disorders you have struggled with:"): anxiety, depression, substance use disorder, or "other psychiatric disorder (e.g. schizophrenia)". This variable was coded "1" to indicate any endorsement of previous struggle with a psychiatric disorder, or "0" to indicate no endorsement of previous struggle with a psychiatric disorder. This coded variable was used to indicate group membership in later factorial invariance analyses. Demographic questions. These included age, gender, education, race, and total incidence of past hallucinogen use. Age was originally coded as interval data and was converted to an ordinal category variable, with the following five categories:<25, 25-34, 35-44, 45-54, 55-64, and ≥65 years. Participants used the following categories to rate the total number of times of past hallucinogen use: 1, 2-5, 6-10, 11-20, 21-50, 51-100, 101-300, >300. Gender, education, and race were coded as categorical data.
ANALYSIS
Matched exploratory and confirmatory analysis strata. The sample of 1993 participants was stratified into two sub-samples that were matched on age, sex, education, race, and total incidence of past hallucinogen use. Stratification was carried out with the strata function of the sampling packagein R (R Core Team, 2012), using simple random sampling without replacement. Before stratification, frequencies were calculated for each cell in the factorial model assumed by the stratification variables to verify that there were no cells with fewer than two observations. A total of 226 observations were removed in this procedure, yielding a final total analysis sample of 1767 (exploratory stratum n=833, confirmatory stratum n=934). After stratification, demographic data were compared between strata using a two-sample t-test for age, and chi-squared independence tests for all other demographic data. Demographic data for the strata are presented in the left-hand portion of Table. No variables differed significantly between strata.
INITIAL SCALE CONSTRUCTION.
The first stratum (n=833) was used to identify the latent structure of the items of the CEQ, using item analysis and exploratory factor analysis. Distributional properties of items were assessed to identify and remove items with restricted range or heavily non-normal distribution. The mixed.cor function from the psych package in Rwas then used to estimate the correlations between the remaining response variables in the exploratory stratum. The mixed.cor function calculates Pearson product moment correlations between continuous variables, polychoric correlations between polytomous variables, tetrachoric correlations between dichotomous variables, and polyserial or biserial correlations between mixed variables, thus generating a "mixed" correlation matrix. Exploratory factor analysis was conducted on this mixed correlation matrix using the fa function in the psych toolbox in R, with maximum likelihood factor extraction and oblimin rotation. Latent variables that contribute to the concept of a challenging experience may reasonably be expected to co-vary. Therefore, an oblique factor rotation (oblimin) was chosen over an orthogonal rotation. Visual inspection of the scree plot and parallel analysis with the fa.parallel function in the psych toolbox in R was used to determine the number of factors to extract. An initial exploratory factor analysis was fit to the data to identify common factors. Items with no loading above 0.4 on any factor were discarded. The remaining items were entered into iterative scale analyses. An initial scale was created for each factor from the items that loaded most strongly onto that factor. For each scale, at each iteration, a parallel analysis was conducted on the surviving items for that scale to assess dimensionality of the scale. When more than two underlying dimensions were identified for a given scale, McDonald's omega was used to estimate the reliability of the scale. Otherwise, Cronbach's alpha was used to calculate scale reliability. Itemtotal correlations were calculated at each iteration, as well as change in scale reliability with removal of each item. Items were removed from a scale if they demonstrated low item-total correlation (below 0.4) and if their removal increased scale reliability and average item-total correlation for the scale. The final remaining items were entered into an exploratory factor analysis to determine the initial model for the CEQ.
MODEL REPLICATION.
The resulting exploratory factor model was replicated using confirmatory factor analysis in the second stratum (n=934). Confirmatory factor analysis in a separate sample is a conservative test of an exploratory factor model, and it establishes reliability and internal validity of the structure of the CEQ. A combination of fit indices was used to assess confirmatory factor model fit, including the comparative fit index (CFI)the standardized root mean square residual (SRMR), and the root mean square error of approximation (RMSEA). Values of SRMR and RMSEA<0.1, and CFI>0.90. Factorial invariance. Factorial invariance analysis is a formal test of whether a set of variables has a similar (or "invariant") factor structure in different groups (i.e. that they show "factorial invariance" across groups). Factorial invariance analysis is conducted by performing a series of multiple-group confirmatory factor analyses, adding more constraints in each sequential model for parameters to be equal across groups and then testing to see whether model fit improves or degrades across sequential models. Configural factorial invariance sets the number of factors between groups to be equal, with the same items loading onto the same factors in each group. Weak factorial invariance adds the additional constraint of the factor loading parameter estimates of each item to be equal across groups. Strong factorial invariance additionally constrains the intercepts of the observed variables to be equal across groups. Strict factorial invariance additionally constrains measurement residuals to be equal across groups. Demonstrating strong or strict factorial invariance provides evidence that the CEQ is measuring constructs in a similar fashion in each group being compared (men vs women, or having vs not having previous struggle with a psychiatric disorder). Establishing strong or strict factorial invariance also allows direct comparison of latent variable means between groups. This is accomplished by setting the latent variable means for the comparison group to zero, and estimating the latent variable means for the other group. This identifies the model and standardizes latent variable mean estimates for the other group. Latent variable means in the other group that are significantly different from zero indicate that latent variables are significantly different from those in the comparison group. Factorial invariance was assessed in a series of multiplegroup confirmatory factor models using the cfa function in the lavaan toolbox in R, with maximum likelihood estimation. Model fit in factorial invariance models was assessed using a combination of change in the CFI, SRMR, and RMSEA. These fit indices have been shown through simulation to be sensitive to both measurement invariance and lack of measurement invariance at the three levels (factor loadings, intercepts, and residuals, or weak, strong, and strict invariance) that are tested within the current sample. Decrease in CFI>0.01, increase in RMSEA>0.015, and increase in SRMR>0.01 between levels of factorial invariance (i.e. change between modeling steps) indicates noninvariance. CEQ and overall impact of the challenging experience. The confirmatory factor model for the entire sample was then extended to a structural regression model with a measurement component to test the relationship between latent variable scores of the CEQ and ratings of the overall impact of the challenging experience (difficulty of the experience, meaningfulness of the experience, spiritual significance of the experience, and effect of the experience on well-being). Regression of CEQ factor scores on ratings of difficulty of the experience provides evidence for the convergent validity of the CEQ. Structural equation modeling was conducted using the sem function in the lavaan toolbox in R, with maximum likelihood estimation.
INITIAL SCALE CONSTRUCTION
Descriptive statistics were calculated for each of the 64 potential CEQ items in the exploratory stratum, and distributional properties of each item (skew and kurtosis) were inspected for each item. Nine items had a skew or kurtosis greater than 1.5. After removing the nominally skewed and kurtotic items, the response distributions of remaining items were visually inspected to verify that the full range of responses had been utilized in each item. The following procedures were repeated after returning these nominally skewed and kurtotic items to the item pool, and there was no change in the outcome. An initial exploratory factor model was estimated to identify both a general factor structure for the response data and the bestloading items on each factor. A Scree plot and parallel analysis indicated that a six-factor solution is appropriate for the dataset. The seven-factor solution yielded one clear junk factor, and the five-factor solution did not yield a very clear qualitative interpretation. The six-factor solution yielded factors with a cohesive qualitative interpretation: physiological distress (factor 1), grief (factor 2), fear (factor 3), insanity (factor 4), isolation (factor 5), and death (factor 6). The six-factor solution was used as the basis of further item removal. Items with no loading above 0.4 on any factor were discarded. A total of 19 items were removed in this step. Items were removed from a scale if they demonstrated low item-total correlation (below 0.4) and if their removal increased scale reliability and average item-total correlation for the scale. An additional 12 items were removed in this step, yielding a final set of 24 items. An exploratory factor analysis was conducted on the 24 remaining items. The final exploratory factor structure, factor intercorrelations, and factor reliabilities are reported in Table. The factors in this solution explained a cumulative 67% of variance in the observed data. Measures of sampling adequacy, including the Kaiser-Meyer-Olkin test (KMO=0.92) and Bartlett's test of sphericity (χ 2 =44,899, df=25, p<0.0001), indicated suitability of the data for factor analysis. Addition of a paranoia factor. Clinical literature and anecdotal reports indicate that paranoia is a subjective experience frequently represented in challenging experiences with classic hallucinogens. While a clear paranoia factor did not emerge from the exploratory factor analyses, the initial item pool for construction of the CEQ did contain two items directly related to paranoia (SOCQ item 40 "Feeling that people were plotting against you" and SOCQ item 72 "Experience of antagonism toward people around you"). Interestingly, there are no items in either the HRS or the 5DASC that directly assess paranoia. When conducting confirmatory factor analysis of the second stratum in the current sample, the identified 24-item, six-factor CEQ model was amended by adding the two paranoia items, both loading onto a seventh factor.
MODEL REPLICATION
A mixed correlation matrix was calculated for the confirmatory stratum, using the items from the final exploratory factor model (Table) as well as two additional items loading onto a seventh factor for paranoia. A confirmatory factor model was fit to this correlation matrix, setting positive loadings of each item onto its intended factor, with all other item loadings equal to zero. The model was identified, and factor loadings standardized, by setting the variance of each latent factor to one. An initial model was fit including the 24-item six-factor structure identified in exploratory analyses of the first stratum. Fit indices for this model (RMSEA=0.070 (90% confidence interval (CI): 0.067-0.074), SRMR=0.054, CFI=0.913) indicate acceptable model fit. A subsequent model was then fit that also included two additional items (SOCQ items 40 and 72) that loaded onto a seventh factor for paranoia. Fit indices for this model (RMSEA=0.066 (90% CI: 0.063-0.070), SRMR=0.052, CFI=0.912) also indicate acceptable model fit. Fit indices, as well as factor loadings, factor correlations, and factor reliabilities for the first six factors did not differ substantially from the original model. This confirms both the initial 24-item six-factor structure of the CEQ, and supports inclusion of a seventh factor for paranoia. Factor structure for the 26-item, seven-factor confirmatory model is presented in Table. Factor correlations and factor reliabilities for this model are presented in the top portion of Table.
FACTORIAL INVARIANCE OF THE CEQ
Factorial invariance was tested separately for two levels of sex and for two levels of previous struggle with a psychiatric disorder (having vs not having had previous struggle). Model fit indices for consecutive factorial invariance models are presented in Table. Fit indices showed negligible change between all levels of factorial invariance for sex and previous struggle with a psychiatric disorder, which indicates strict factorial invariance between categories in both sets of groups. Factor correlations for each level of gender and previous struggle with a psychiatric disorder, and differences in factor scores between levels, are presented in the middle and lower portion of Table.
OVERALL IMPACT OF THE CHALLENGING EXPERIENCES
Model fit indices for the structural equation model regressing overall impact ratings on CEQ factor scores indicated good The upper matrix presents loadings of items onto exploratory factors for the final item set. Factor loadings greater than 0.4 are in bold. The h2 column shows the communality of each item. The lower matrix includes factor intercorrelations below the diagonal in non-italicized typeface, and factor reliabilities on the diagonal in italics. fit (CFI=0.915, RMSEA=0.060 (90% CI: 0.058-0.062), SRMR=0.045). Regression estimates are presented in Table. Scores on the CEQ death factor were associated with a decrease in wellbeing attributed to the experience, and were positively associated with all other ratings of overall impact. Scores on fear were positively associated with ratings of the difficulty of the experience and change in wellbeing, and were negatively associated with the spiritual significance and meaningfulness of the experience. Isolation scores were positively associated with an increase in wellbeing and negatively associated with meaningfulness and spiritual significance of the experience. Scores on both insanity and grief were positively associated with rated difficulty of the experience, meaningfulness and spiritual significance of the experience. Physical distress was positively associated with ratings of spiritual significance and associated with a decrease in wellbeing. Factor scores on paranoia were not associated with any overall impact ratings.
STUDY 2
The items comprising the CEQ in Study 1 were taken from three separate and extensive instruments (the HRS, with 104 items total, the SOCQ with 100 items total, and the 5DASC with 42 items), each with a different response format. In Study 2, the 26-item CEQ was validated as a stand-alone instrument (i.e. outside of the context of the other items in the HRS, SOCQ, and 5DASC). Item text from the HRS and 5-DASC items was altered to match the prose of the items from the SOCQ, and the six-item SOCQ response format was adopted for all items of the CEQ. Side-by-side comparison of original and altered items is presented in Supplementary Material, Appendix 2. Items of the stand-alone CEQ were administered along with demographic measures and measures of the overall impact of the reported challenging experience used in Study 1 and analyzed in order to replicate the factor structure and regression results obtained in Study 1. CEQ. Twenty-six items identified in Study 1 (including two paranoia items) were presented in fixed pseudorandom order, with the same response format for each item ("0 -none; not at all"; "1 -so slight cannot decide"; "2 -slight"; "3 -moderate"; "4 -strong"; "5 -extreme (more than ever before in my life)"). Since the original items from the 5-DASC, HRS, and SOCQ differed between scales in style and prose, modifications were made to the text of the 5-DASC and HRS items in order to better match the prose of these items in the stand-alone CEQ. Side-by-side comparison of original
RESULTS
Study 2 differed from Study 1 on all demographic variables (Table, right-hand portion). Participants in Study 2 were older, more highly educated, and more racially and ethnically diverse. Study 2 contained a greater percentage of females than Study 1. Participants in Study 2 also reported a greater number of experiences with classic hallucinogens in general and more specifically psilocybin mushrooms than participants in Study 1. Finally, participants in Study 2 rated their experiences as overall slightly less difficult, and slightly more individuals on average in Study 2 indicated that they would repeat their reported experience, including the challenging portion, if given the opportunity. Measures of sampling adequacy, including the KMO test (KMO=0.89) and Bartlett's test of sphericity (χ 2 =538.01, df=26, p<0.0001), indicated suitability of the data for factor analysis. Model fit indices for factorial invariance models are presented in the lower portion of Table. Change in fit indices for each level of factorial invariance between Study 1 and Study 2 did not exceed critical values for noninvariance. This supports strict factorial invariance of the 26-item CEQ between Study 1 and Study 2. Factor correlations and reliabilities for the stand-alone CEQ are presented in the top portion of Table. Factor score differences between Study 1 and Study 2 are presented in the bottom portion of Table. Factor scores were significantly higher in Study 2 on fear and physical distress factors, and significantly lower on insanity and paranoia factors than in Study 1. Model fit indices for the structural equation model regressing overall impact ratings on CEQ factor scores indicated good fit (CFI=0.912, RMSEA=0.064 (90% CI: 0.061-0.067), SRMR=0.048). Regression estimates are presented in Table. Scores on the CEQ death factor were negatively associated with ratings of change in wellbeing attributed to the experience, and positively associated with all other ratings of overall impact. Scores on fear were positively associated with ratings of the difficulty of the experience and change in wellbeing, and negatively associated with the spiritual significance of the experience. Scores on insanity were positively associated with rated difficulty of the experience and meaningfulness of the experience. Grief and physical distress were positively and negatively associated with ratings of the difficult of the experience, respectively, and paranoia factor scores were positively associated with change in wellbeing attributed to the experience, but neither factor score were associated with other impact ratings. Factor scores on isolation were not associated with any overall impact ratings.
GENERAL DISCUSSION
The CEQ was developed from responses to an Internet survey of challenging experiences, and the stand-alone CEQ was validated in responses to a separate Internet survey of challenging experiences. The seven factors of the CEQ display a simple factor structure, have high face validity, and represent a wide sampling of challenging aspects of psychedelic experience (referred to as "challenging experiences" for brevity) that aligns well with previous reports of presenting symptoms of acute adverse reactions to hallucinogenic substances. This includes affective (fear and grief), physiological (physical distress), and cognitive/affective (feelings of isolation, paranoia, feelings of insanity or loss of sanity, and the subjective experience of death) categories of subjective experience.
INTERNAL AND EXTERNAL VALIDITY OF THE CEQ
The factor structure of the CEQ was identified and confirmed in Study 1 within separate sub-samples of the data, and was shown to be invariant across levels of gender and struggle with psychiatric illnesses before the reported challenging experience. The factor Participants provided ratings of the degree of difficulty of the challenging experience that they reported. Scores on the fear, grief, insanity, and death factors of the CEQ were consistently positively associated with these difficulty ratings in both Study 1 and Study 2, providing evidence for external validity of the CEQ. The experiences represented by the items in these four factors may constitute the core of what may be considered a challenging experience, while the remaining three factors may represent less reliable aspects of challenging experience. Differences in factor scores between samples in four scales (fear, physical distress, insanity, and paranoia) were observed (Table), and should be explored further in future studies. Greater fear and physical distress scores, and lower paranoia and insanity scores, were observed in Study 2 when compared to Study 1. These differences may be attributable to differences between Study 1 and Study 2 samples in demographics, but it is equally if not more compelling to consider that additional factors that were not controlled in these samples, such as psilocybin dose and details of set and setting, may be associated with differences in CEQ scores. An accepted clinical definition of "paranoia" is "unfounded fears that others intend harm to the individual". While one item of the paranoia scale of the CEQ is consistent with this definition ("feeling that people were plotting against you"), the other ("experience of antagonism toward people around you"), while likely related, is not closely consistent with this definition. Thus, the CEQ paranoia scale may be viewed as a crude measure of the clinical construct of "paranoia", and the external validity of the paranoia scale may be somewhat restricted by this limitation.
CHALLENGING EXPERIENCES AND THE OVERALL IMPACT OF EXPERIENCES WITH PSILOCYBIN
Previous studies of the subjective effects of psilocybin have used ratings of the meaningfulness and spiritual significance of the experience to provide a very general characterization of the effects of psilocybin, in relation to other experiences that a volunteer has had. A consistent finding has been that moderate to high doses of psilocybin (20+ mg/70 kg) occasion experiences that are frequently rated in the top five most meaningful and spiritually significant experiences of a participant's life. Participants have also attributed positive change in well-being to their psilocybin experience, and this change in well-being has been shown to correlate positively with mystical experienceand egodissolution. Preliminary research further suggests positive psilocybin-occasioned behavior change in the context of addiction treatment, providing a general characterization of the longer-term impact of these experiences. The associations between CEQ factor scores and the overall impact ratings of meaningfulness, spiritual significance, and change in well-being were assessed in the current samples. Scores on the insanity and death factors of the CEQ were positively associated with ratings of the meaningfulness of the reported experience in both Study 1 and Study 2. Scores on the fear factor were negatively associated with spiritual significance, while scores on the death factor were positively associated with spiritual significance of reported experiences in both studies. To the extent that individuals might construe or relate the loss of self-referential processing that is often reported during mystical experiences as feeling as though they are losing a sense of sanity or experiencing their own death, an encounter with this facet of challenging experience may be expected to covary with both the meaningfulness and the spiritual significance of an experience. The subjective experience of one's own death and loss of control of the mind might somehow allow for the type of unity experience that leads to spiritual and meaningful experiences. However, scores on the death factor of the CEQ were negatively associated with change in well-being attributed to challenging experiences. While the experience of ego dissolutionand mystical experienceare positively associated with well-being, it is not completely clear that the items of the 'death' subscale of the CEQ ("Profound experience of your own death" and "Feel as if dead or dying") are collinear with either mystical experience or ego dissolution. The positive association between wellbeing and both mystical experience and egodissolution, contrasted against the negative association between wellbeing and the 'death' scale of the CEQ in both Study 1 and Study 2, suggests that there may be something unique about the subjective experience described as 'death' or 'dying' during a challenging experience that may detract from wellbeing, and this may have implications for therapeutic efficacy of psychedelics in clinical trials. Thus, future work may benefit from further elucidating the relationship between ego dissolution or mystical experience and the 'death' factor of the CEQ. Scores on the fear factor of the CEQ were associated with an increase in well-being attributed to challenging experiences and negatively predicted meaningfulness and spiritual significance of challenging experiences. Fear may generally detract from a spiritual experience, but the crucible of panic during a challenging experience might still lead to positive outcomes. Some discrepancies between Study 1 and Study 2 are noted in the significant predictors of meaningfulness, spiritual significance, and change in well-being attributed to experiences. Scores on the paranoia factor were negatively associated with the rated difficulty of the experience in Study 1 but not Study 2, while scores on the physical distress factor were negatively associated with the rated difficulty of the experience in Study 2 but not Study 1, and scores on the isolation factor were not associated with rated difficulty of the experience in either study. While fear and isolation were negatively associated and grief was positively associated with the rated meaningfulness of the experience in Study 1, they were not significantly associated with this rating in Study 2. Grief, physical distress, and insanity were positively associated, and isolation was negatively associated, with spiritual significance of an experience in Study 1, but not Study 2. Further, physical distress was negatively associated and isolation was positively associated with change in well-being in Study 1 but not Study 2, while paranoia was positively associated with change in wellbeing in Study 2 but not Study 1. It is possible that demographic differences between the samples in each study or the difference between studies in the relative rated difficulty of challenging experiences may have contributed to observed discrepancies in the associations between CEQ factor scores and ratings of the overall impact of the experiences (difficulty, meaningfulness, spiritual significance, and change in well-being attributed to the experience). It may also be that aspects of experience that were not controlled in these studies, such as particular features of set and setting, psilocybin dose ingested (which cannot be precisely known, especially from retrospective reports of ingested fungal matter), or traits of individuals such as personality or attachment type, interact with the relationship between factor scores and ratings of the overall impact of the experiences. Another limitation of this study is that individuals who have a positive attitude towards psychedelics were probably more likely to complete the survey than those with a negative attitude towards psychedelics, given that the main sources of recruitment were websites that are typically frequented by those with a positive attitude towards psychedelics. It may be that the relationships between CEQ scores and measures of the overall impact of the experiences could be different for those with a different regard towards psychedelics. While these questions deserve prospective scrutiny in controlled environments, it is encouraging that consistencies were found between studies (e.g. the relationship between fear, grief, insanity, and death factor scores and rated difficulty of the experience).
POTENTIAL RISK FACTORS FOR CHALLENGING EXPERIENCES
Factor scores in Study 1 did not differ by gender (except for slightly greater scores on the fear factor for women compared to men). In contrast to gender, factor scores were significantly greater in those who had previously struggled with a psychiatric disorder, compared to those who had not, for all factors except for paranoia. While this may seem to suggest that individuals with a history of psychiatric disorders may be more prone to a greater degree of challenge with psilocybin, it is not clear that such a generalization is warranted. The current survey does not inform or address the frequency or likelihood of a challenging experience, but rather the potential profile of subjective experience when a challenging experience is encountered. Also, the designation of having previously struggled with a psychiatric disorder is a coarse designation at best. This categorization does not differentiate well among many important and distinct disorders, including mood disorders or psychosis. Those who suffer from disorders on the psychotic spectrum, which can be accompanied by delusions, ideas of reference, and paranoia, may be at increased risk of experiencing paranoia while experiencing the effects of a classic hallucinogen. Also, reliance on self-identified struggle with a psychiatric disorder is imperfect for diagnostic purposes. While recommendations for safe conduct of hallucinogen research indicate a contraindication for individuals with personal or family history of psychosis, non-psychotic depression may not be contraindicated. Recent empirical work has shown the potential value of psilocybin as a treatment for depressionand anxiety secondary to a life-threatening cancer diagnosis, and recent large-sample surveys have demonstrated a lower risk for psychological distress and suicidality in those who have endorsed having consumed a classic hallucinogen in the past. Future studies should explore a more fine-grained understanding of the relationship of previous history with various psychiatric disorders to challenging experiences with psilocybin or other classic hallucinogens and clinical outcomes of psychedelic therapies.
TOWARDS A MODEL OF CHALLENGING EXPERIENCES
With the CEQ, one may begin to consider an empirical model for challenging experiences. A full model of challenging experiences should include the acute effects of challenging experiences, the conditions that may predict a challenging experience, and the persisting effects of that experience. The CEQ provides a structure to consider the acute effects of challenging experiences that may be used in prospective studies. Additional categories of challenging experience have been suggested in the literature, such as the experience of karmic or "astral" experiences, experience of "kundalini" energy, or the experience of such phenomena as age regression. Some of these experiences may be culturally bound or framed explanations of intense emotional or physiological responses to drug conditions. Moreover, many such experiences are non-specfiic to challenging experiences, as they may arise during both challenging and non-challenging experiences. In this case, they may not be uniquely informative of challenging experiences. There are many questions that can be asked regarding challenging experience, including the nature of challenging experiences, the predictors of emergence of challenging experiences, and the consequences of having had a challenging experience. The current findings deal primarily with challenging aspects of experiences with psilocybin, as reported by the study volunteers regarding their primary subjective experience. Future studies may benefit from this understanding, and will be able to use the CEQ to investigate predictors and consequences of challenging experiences.
CONCLUSION
The CEQ as validated in this article may serve as a valuable tool for characterizing psychologically difficult aspects of experiences occasioned by psilocybin and, very likely, by other classic hallucinogens. Better understanding of challenging experiences with classic hallucinogens may increase the precision of our understanding of both the psychological nature of and neural mechanisms underlying the effects of these drugs. Understanding challenging experiences with classic hallucinogens may also facilitate the optimization of therapeutic application of drugs within this class. Thus, the CEQ developed in this article makes a significant contribution to methods of assessment of acute psychologically adverse reactions to psilocybin and, with further validation, will likely make a significant contribution to methods of assessment of such experiences with other classic hallucinogens.
Study Details
- Study Typeindividual
- Populationhumans
- Characteristicssurvey
- Journal