Health Benefits and Positive Acute Effects of Psilocybin Consumption: A Quantitative Textual Analysis of User Self-Reported Data

ABSTRACT

There has been growth in the use of psychedelics by the global population in recent years. In addition to recreational and ritualistic use, recent research into psychedelics has brought advances for treating mental disorders. Understanding the specific circumstances in which psilocybin leads to positive outcomes may have important implications for the future of its clinical use and for harm reduction initiatives. This study aimed to investigate the positive effects from the consumption of psilocybin through public online self-reports. We sought to investigate health benefits promoted by the consumption of the substance, positive acute effects, and contextual details of these experiences. We analyzed 846 reports with the assistance of the IRaMuTeQ textual analysis software, adopting the procedures of Descending Hierarchical Classification, Correspondence Factor Analysis, and Specificities Analysis. The texts were grouped in 5 clusters, describing the content of mental experiences, cognitive processes, somatic experiences, perceptual alterations, and context of administration. The findings of this study reinforce central axes of the psychedelic experience, such as the presence of somatic and visual alterations, connectedness and feeling one with the world and effects of setting, as well as the beneficial character of mystical experiences this substance promotes, and the importance of the ego-dissolution phenomenon.

KEYWORDS:

• Psilocybin
• psychedelics
• ego dissolution
• erowid
• magic mushrooms

Introduction

Outside clinical settings there has been a growing use of psychedelics in the general population. From 2015 to 2021, “magic mushrooms” (mushrooms that contain psilocybin) had increases in lifetime users from 25% to 33% and from 9% to 16% for use in the last 12 months (GDS. Global Drug Survey 2022). Recent decriminalization of psilocybin mushrooms in several cities/states in the U.S. (e.g., Oregon, Seattle, Detroit), following previous efforts in this direction worldwide (e.g., Canada, Netherlands, Portugal), may lead to an increase in their documented use. This leads to a research priority to address the potential impact, both in terms of benefits and harms, of psilocybin.

In addition to recreational and ritualistic use, recent research into psychedelics has brought advances for treating mental disorders like depression and anxiety, for which there has been very little pharmacological innovation. Despite great public health need, the beginning of the 21^st century was characterized by a paucity of pharmacological innovation in the treatment of mental illness (Hyman 2014). Between 20 and 30% of patients suffering from depression do not respond to treatment, leading to lower quality of life and higher medical costs (Olchanski et al. 2013). Psilocybin has been shown to be effective against treatment-resistant depression in randomized placebo controlled studies (Luoma et al. 2020), and in a head-to-head comparison it has shown better but not statistically superior results to escitalopram (Carhart-Harris et al. 2021). Recent meta-analytic approaches have reinforced these results (Yu et al. 2022). These findings include psilocybin as a new therapeutic intervention for mood disorders, together with other breakthrough therapies, such as transcranial magnetic stimulation (Gaynes et al. 2014) and ketamine (Serafini et al. 2014). Psilocybin has also produced substantial and sustained decreases in anxiety, depression, and suicidal ideation in life-threatening cancer patients (Griffiths et al. 2016; Ross et al. 2021) and can promote positive outcomes in treatment of obsessive-compulsive disorder (Lowe et al. 2021), addiction (Daniel and Haberman 2017), sociocognitive impairments in psychiatric illnesses (Preller et al. 2016), and chronic pain (Castellanos et al. 2020).

However, these findings have been discussed critically, with some considerations about the harm potential of magic mushroom use (Amsterdam, Opperhuizen, and Van den Brink 2011). Healthy individuals may also experience mental confusion, agitation, extreme anxiety, fear and psychotic episodes, including bizarre and frightening images, severe paranoia, and loss of sense of reality (Johnson et al. 2018). In addition, an epidemiological study found significant associations between the consumption of hallucinogens throughout life and some negative outcomes including suicide attempts (Shalit, Rehm, and Lev-Ran 2019).

Despite psilocybin being considered a toxicologically safe substance (Johnson et al. 2018; Rucker et al. 2016), individual and contextual elements are crucial to predict whether an experience will be perceived as positive and produce therapeutic effects. Therefore, despite the growing number of clinical studies with psilocybin, it is also important to investigate its effects in uncontrolled real-life event settings. In a previous study, we found evidence of negative outcomes such as bad trips, anxiety, search for emergency medical care and health problems (Bienemann et al. 2020) using self-reports extracted from an online platform.

Understanding the specific circumstances in which psilocybin leads to positive outcomes may have important implications for the future clinical use of this substance, also providing relevant information for harm reduction initiatives. Even so, there is still a scarcity of quantitative analyses of self-reported user data, which is frequently neglected in clinical studies. Considering this, the aim of this work was to investigate positive effects resulting from the consumption of psilocybin through public self-reports extracted from an online platform. Specifically, we sought to investigate health benefits promoted by the consumption of the substance, positive acute effects, and contextual details of these experiences.

Methods

Extraction of data and construction of textual corpus

Text data are from reports manually extracted from the EROWID website (www.erowid.org), a database dedicated to reporting psychoactive substances and documenting actual reports from users. Reports are reviewed before publication and authors are asked to meet certain criteria, such as: Description of the context in which the experience was had and their previous mental states, details of the preparations made for ingestion, details of dosage and timing, observations of any other medicines, herbs or supplements ingested, and a description of the physical and mental effects experienced. The reports are published anonymously and are available on the website in different categories (Erowid 1995). For this study, we selected the reports in the subcategories “health benefits,” “medical uses,” “glowing experiences,” and “mystical experiences” included in the category “mushrooms” (magic mushrooms; psilocybin-containing mushrooms). Reports describing the use of mushrooms with substances other than psilocybin as the main active ingredient (e.g., Amanita muscaria) were not included.

The texts (n = 846) were transcribed manually, and any grammatical or typing errors were corrected. In addition, some symbols were deleted or replaced (e.g., hyphens, inverted commas, dashes) to allow analysis by the software. The average length of the reports was 36.48 words. In addition, the reports were also categorized according to the following variables: the subcategories mentioned above (“health benefits,” “medical uses,” “glowing experiences,” and “mystical experiences”), presence of substances other than mushrooms, dosage, route of administration, and gender of the user. For dosage, a binary variable (doses below and above 5 g) was created for the analyses, taking into account what has been described as a high dose with qualitatively different experiences (Oss and Oeric 1993). Missing values were classified as null and classifications differing from those above were classified as other. To determine the reliability of the analysis, inter-rater reliability was calculated for all categories not previously provided by the website (i.e., presence of substances other than mushrooms, dosage, route of administration, form of consumption, and sex of the user).

Data analysis

The participants’ answers were first evaluated qualitatively to create familiarity with the content. In this phase, the reports were read in detail one after the other by two members of the research team. Subsequently, the texts were analyzed quantitatively using IRaMuTeQ 0.7 alpha 2 (Ratinaud and Marchand 2012) and R 3.1.2 (R Core Team 2017). The analysis was conducted in the text corpus created from the reports and their categorizations using text segments (TS). TS are subdivisions of the text that define the context in which words appear. TS are automatically divided according to corpus extension; in this study, we used the standard division provided by IRaMuTeQ (40 words per text segment). We used the procedures of Descending Hierarchical Analysis (DHA, Reinert method), specificities analysis, and Correspondence Factor Analysis (CFA). DHA aims to obtain content text clusters with specific meanings based on the similarity, association, and frequency of their vocabularies. CFA leads to a graphical visualization of the proximity, contrasts, and tendencies of text segments (TS) or corpora clusters. It locates these elements in a Cartesian graph with factors generated from their classifications and allows a graphical visualization of the co-occurrence of words and the possible communities in which they cluster (Loubère and Ratinaud 2014). The analysis of specificities shows the index of co-occurrence between words, i.e., the relationship of words to each other and the communities formed by groups consisting of the most strongly associated words.

The criteria for including words and categories in their respective clusters by DHA are a frequency greater than the mean of occurrences in the corpus and a chi-square value with the cluster greater than 3.84. The words of interest (active forms) selected for analysis were adjectives, nouns, pronouns, verbs, adverbs, and forms not recognized by the IRaMuTeQ dictionary. In addition, when words with other associated forms occur (e.g., test, testing, tested), the most frequent form was selected for graphing. The chi-square test values indicate how strongly words and categories are associated with their clusters (Loubère and Ratinaud 2014). We also reported Cramer’s V, a measure of effect size of association (Kim 2017). To avoid inflation of type I errors, α was set to .01.

All materials were anonymized, preventing identification of the subjects. In view of the fact that the data were public, in accordance with national ethics regulations (Ministry of Health, National Health Council 2016), application for ethics committee approval was dispensed (Ministry of Health, National Health Council 2016).

Results

Descending hierarchical analysis

The analysis by DHA retained 84.5% of the total corpus, a percentage indicated as acceptable for the corpus to be considered for this type of analysis. The corpus was divided into 37,033 TS, relating 29,093 words that occurred 1,351,032 times (mean of occurrence for TS = 36.5). Of these, the active forms formed 21,181 words, with 2849 words with frequency equal or greater than 19.

As can be seen in dendrogram form (Figure 1), DHA resulted in five clusters of words. Initially, the clusters were grouped into two distinct branches: one composed only of Cluster 3 (15.08% of the total forms classified) and another split into two other branches, combining, respectively, Clusters 5 (22.55%) and 1 (24.1%), and 4 (17.78%) and 2 (20.49%). For the association between words and clusters (degrees of freedom = 4), considering the 25 words with the highest association in each cluster, Cramer’s V indicated medium and, particularly in Clusters 2, 3, and 4, large effect sizes.

Figure 1. Dendogram with the 25 words with highest χ2 in each cluster.

Large (red), medium (blue), and small (green) effect sizes, according to Kim (2017).

Figure 1 also shows the association of previous categories of the reports and variables of interest with each cluster. It is observed that Cluster 1 was significantly associated with the subcategory mystical experiences [χ² (3) = 235.1; p < .001; V = .53] and health benefits [χ² (3) = 5.0; p = .026; V = .08], one single dose [χ² (4) = 65.9; p < .001; V = .28], and more frequently reported by male users [χ² (2) = 15.4; p < .001; V = .13]. Cluster 2 was significantly associated with the subcategory glowing experiences [χ² (3) = 74.6; p < .001; V = .30], with weak doses [χ² (1) = 15.5; p < .001; V = .14] and with a second dose larger than the first [χ² (4) = 7.3; p = .007; V = .09]. Cluster 3 was strongly associated with more than two doses [χ² (4) = 62.5; p < .001; V = .27], moderately associated with subcategory medical use [χ² (3) = 27.2; p < .001; V = .18] and slightly associated with consumption of other substances besides mushrooms [χ² (1) = 58.8; p < .001; V = .26], female users [χ² (2) = 30.8; p < .001; V = .19], glowing experiences [χ² (3) = 15.4; p < .001; V = .14], two equal doses [χ² (4) = 13.1; p < .001; V = .12], strong doses [χ² (1) 10.4; p = .001; V = .11] and health benefits [χ² (3) = 8.7; p = .003; V = .10]. Cluster 4 was moderately associated with the subcategory glowing experiences [χ² (3) = 30.0; p < .001; V = .19], with two equal doses [χ² (4) = 19.5; p < .001; V = .15], and slightly associated with use of just mushrooms [χ² (1) = 17.1; p < .001; V = .14] and female users [χ² (2) = 6.8; p = .009; V = .09]. Cluster 5 did not show significant associations with any category.

Correspondence factor analysis

The CFA carried out in order to visualize the relation between the clusters indicated that the clusters are divided mainly into three areas, with cluster 1 (red) and 5 (pink) being strongly related to each other. Cluster 2 (gray) and 4 (blue) are slightly related (Figure 2). We can see that Cluster 3 is isolated from others and located on the other side of the chart.

Figure 2. Relationship between clusters and words in each cluster.

Red – Cluster 1; gray – Cluster 2; green – Cluster 3; blue – Cluster 4; pink – Cluster 5

Specificities analysis

The specificities analysis, indicating the index of co-occurrence between the words, can be seen in supplementary material (S1, S2, S3, S4, and S5).

Discussion

The results indicated the occurrence of 5 clusters of words, that is, environments of meaning present in the participants’ speech (Salvador et al. 2018). On those clusters, we can see a relationship between clusters 1 and 5, with the words very close together on the AFC Cartesian graph (Figure 2), and clusters 2 and 4 slightly related, with the words close together on the same axis of the graph (Figure 2). Still in relation to the proximity of the clusters, it is noted that cluster 3 differs from the others both in the CFA graph (Figure 2) and in the dendrogram, as it belongs to a different arm from the others that are components of the same origin (Figure 1). When we analyze the clusters semantically, the isolation of cluster 3 makes sense, since it seems to concern the setting and the context in which the experience took place, while the others refer to contents arising from the experience itself. That is, while cluster 3 gathers words external to the experience, the other clusters gather words more related to the first-person experience.

Analyzing the words individually and their associations with each other, it is assumed that cluster 1, which gathered 24.1% of the TS, refers to elements that concern the content of mental experiences typical of the experience with psychedelics, such as the ego-dissolution phenomenon (Letheby and Gerrans 2017) and oceanic feeling (Preller and Vollenweider 2016), by the presence of keywords such as “life,” “mind,” “universe,” “consciousness,” “nature,” “being,” “world,” among others. A similar cluster to this was also found in previous analysis of reports of negative experiences with psilocybin (see Bienemann et al. 2020). However, unlike in the analysis of negative experiences, where words with negative connotations were also seen (e.g., “fear,” “death,” “insane,” “crazy,” “cry” “die”), in the current study, only positive and neutral valence words showed significant associations (with the exception of the word “death,” which appears strongly linked to the word “life”). Thus, while the mental contents present in the reports of negative experiences approach a structure of negative affect that are linked to paranoia (Kramer et al. 2014), in the reports of positive experiences the words that appear make more direct reference to the ego-dissolution phenomenon (Nour et al. 2016). These data are in agreement with studies that show a positive correlation between ego dissolution and higher levels of life satisfaction, positive affect, among other positive mental health outcomes (e.g., Uthaug et al. 2018; van Oorsouw et al. 2021), which is also supported by the association of this cluster with the Health Benefits subcategory, although the association had a low effect size and was marginally significant (p = .026).

In addition, in this study, mystical words also appeared in this class, such as “spiritual” and “jesus,” besides the strong association of Cluster 1 with the subcategory “mystical experiences.” It is also noted that the central and strongest word of this cluster was the word “life,” and that there is a connection between the words “jesus” - “mind” - “life,” as well as “spiritual” – “life” which, in turn, also connects to the word “existence” (a graphical visualization of the co-occurrence between the words can be seen in the supplementary material). That is, these associations give an auspicious connotation to mystical words, which tend to be evident in a more positive than negative or anxiogenic sense. This fact is in accordance with data that point to mystical experiences as one of the possible elements that underpin the therapeutic mechanisms of psychedelics in mental health (Wheeler and Dyer 2020). The strong association of this cluster with single doses may indicate the use of higher and more effective doses, which may be linked to the fact that psilocybin induces dose-dependent alterations in perception and positively experienced ego dissolution (Hirschfeld and Schmidt 2021). While in a previous analysis of negative experiences (Bienemann et al. 2020), the cluster similar to this one was associated with bad trips and female users, in this analysis there was an association with health benefits and male users. It would be important for other studies to seek to further investigate gender differences in the way of experiencing ego dissolution and other effects from stronger doses of psychedelics. Possibly, the effects of higher doses should be expressed more strongly in women, due to enzymatic, hormonal and even social factors. For example, there is evidence of changes in the menstrual cycle in women caused by the use of classic psychedelics (Gukasyan and Narayan 2022). Future studies should investigate further the relationship between gender and ego dissolution.

Cluster 5 retained 22.5% of TS and was not associated with any subcategory. This cluster is contained in its central words “what,” “me,” “know,” “think,” among others (Figure 1) and it was a cluster that did not appear in the previous analysis (Bienemann et al. 2020), which only 4 clusters were obtained using the Reinert Method. Some of the words in this cluster (e.g., “think,” “know,” “remember,” “explain”) suggest a cognitive component to the experience. It is possible that this cluster is, in a way, part of the same sense environment composed by cluster 1 because, in addition to being part of the same arm of the dendrogram (Figure 1), when we analyze the Corresponding Factor Analysis (Figure 2), we see that his words seem quite close and related to each other. Therefore, possibly, it is a cluster that also relates to the mental phenomena typical of the psychedelic experience. However, unlike cluster 1, this cluster brought some words that possibly demonstrate some kind of ansiogenic experience, such as “cry,” “confuse,” “die,” “worry,” “fuck.” These data demonstrate that even in psychedelic experiences with positive outcomes, there may be a difficult part of the experience that promotes some anxiety, but still contributes to the positive effects in general. This is in confluence with the scientific psychedelic literature that reinforces that there is a therapeutic character in challenging experiences (Carbonaro et al. 2016), but at the same time it is necessary that there is a post-experience contextual integration work (Carhart-Harris et al. 2018) that can help integrate and make sense of the challenging story into the personal narratives of users (Gashi, Sandberg, and Pedersen 2021). In that way, it would be important for further studies to seek to identify which subjective aspects during and after these experiences helped to transform them into positive outcomes.

The second cluster of words had 20.49% of TS and, as well as in negative-experiences analysis, showed terms related to action, such as “walk,” “back,” and “sit,” which occupy central position in the cluster, with some nouns complementing. It is also noted that the presence of positive terms, such as “laugh,” and the absence of terms that appeared in the previous analysis, such as “ambulance,” “police,” “hospital,” “cop” (Bienemann et al. 2020), which would possibly indicate some problematic situation resulting from the experience. Therefore, while in negative experiences some of these actions indicated a possible occurrence of emergencies (Bienemann et al. 2020), when we unravel the relationship between the words, this nucleus does not appear in positive experiences. It is possible this is explained by the association of this cluster with lower doses, leading to an experience that preserved functional capacity. Additionally, this cluster was strongly associated with glowing experiences, pointing to positive outcomes linked to somatic feelings. The occurrence of this cluster in both positive and negative experiences, albeit with different word associations, suggests that a bodily and motor component is central to psychedelic experiences.

This somatic component was closely linked to cluster 4, which held 17.78% of TS, with words such as “color,” “eye,” “pattern,” “look,” “sky,” among others, which refer to visual and perceptual changes characteristic of the psychedelic experience. This may suggest a strong perceptual nucleus to subjective experiences of psilocybin users, encompassing both somatic and visual elements. This was the cluster that less differed from the previous analysis (Bienemann et al. 2020), with most of the words found here also having appeared in this same cluster previously, highlighting the characteristic hallucinogen effects of psilocybin. In the current study, this cluster was associated with the “glowing experiences” subcategory, which may point to the fact that people who had this type of experience, and consequently registered them in this subcategory of the erowid site, have had more visual and perceptual changes in their experiences.

The third cluster held 15.08% of the ST and seems to be related to context and details of the experience in general, with words such as “mushroom,” “eat,” “gram,” “smoke,” among others. A very similar cluster was also found in the previous analysis of negative experiences (Bienemann et al. 2020). It is noted that, in the previous study, words had appeared that possibly indicated the use of other drugs along with mushrooms, such as “lsd,” “acid,” “weed,” “marijuana,” “pot,” while in this study only the word “weed” appeared, related to “smoke” which, in turn, relates to “friend.” In the previous study, the word “smoke” was related to “trip” which, in turn, was related to “experience.” Although we know that the concomitant use of mushrooms and different substances can be harmful (van Amsterdam, Opperhuizen, and Van den Brink 2011), many users manage to make moderate use of marijuana during the psychedelic experience with positive outcomes and this may be the case in these reports.

Conclusion

This study aimed to analyze positive reports resulting from psilocybin consumption, from the perspective and words of the users themselves. Although psilocybin use has been made for centuries, reports of subjective experiences of users are still little scientifically analyzed. We believe that these analyses reveal important data on the mechanisms underlying the different possible outcomes that the same substance can promote in its users. The findings of this study reinforce central axes of the psychedelic experience (such as the presence of somatic and visual alterations, connectedness and feeling one with the world and effects of setting), as well as the beneficial character of the so-called mystical experiences that the substance promotes, and the importance of the ego-dissolution phenomenon positively experienced. Regarding the comparison with a previous study that focused on negative experiences, results suggest that the thought content during positive experiences is more related to aspects of ego-dissolution and transcendence, whereas in the previous study these contents signaled anxiety and paranoia.

In addition, findings reinforce the importance of studies aiming to investigate gender differences in psilocybin experiences, as well as investigating possible coping mechanisms to transform challenging experiences into beneficial experiences. It is possible that, in addition to integration works post-psychedelic sessions, there are coping mechanisms that arise during the acute experience with the substance, which may help in harm reduction initiatives and crisis management of people under the effect of psilocybin. It may be interesting to carry out studies in which this type of data analysis is applied to reports made during the experience with psychedelics, in order to avoid memory biases, allowing more accurate analysis of the mental content linked to these experiences.

The study is limited by the high number of missing values that prevented the analysis of contextual variables, incomplete representation of psilocybin users due to sampling biases, and the reliance on online self-reports. There are potential confounding effects of psilocybin’s illegal status, which could be linked to negative outcomes resulting from black market influences, such as the use of different substances or a lack of information about mushroom freshness, rather than the psilocybin content itself. In this regard, it would be important to determine the location of the collected reports since areas where psilocybin use has been decriminalized may show more positive outcomes due to the absence of police and social repression of the practice.

Acknowledgments

The authors acknowledge funding from Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES - Brazil), Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPQ - Brazil) and Fundação de Amparo à Pesquisa do Estado do Rio de Janeiro (FAPERJ - Brazil).

Disclosure statement

No potential conflict of interest was reported by the authors.

SUPPLEMENTARY MATERIAL

Supplemental data for this article can be accessed online at https://doi.org/10.1080/02791072.2023.2226414

Additional information

Funding

The author(s) reported that there is no funding associated with the work featured in this article.