The Altered States Database: Psychometric data from a systematic literature review

Earlier research has increased interest in systematically measuring altered states of consciousness (ASC), which are transient mental states characterised by marked changes in perception, self-experience, time, and sensation and which can be induced pharmacologically (for example by psychedelics) or non-pharmacologically (for example by meditation or sensory deprivation). Standardised, validated questionnaires have been developed to quantify the subjective phenomenology of ASC and to allow comparison across induction methods, study designs, and participant groups. Interest has recently intensified because of both basic science questions about consciousness and clinical trials exploring therapeutic uses of psychedelic drugs. Prugger and colleagues aimed to assemble and harmonise published psychometric questionnaire data on ASC into an open, reusable resource: the Altered States Database (ASDB). The project sought to identify all journal articles reporting questionnaire-based measurements of ASC, extract group-level summary statistics (means, standard deviations) for questionnaire factors and dimensions, and make these data available via the Open Science Framework and a dedicated ASDB website to support cross-study comparisons and dose–response analyses. This paper reports the systematic review methods, the extracted dataset, and the organisation and technical validation of the database.

The investigators conducted a systematic literature search of PubMed and ISI Web of Science to identify journal articles reporting ASC assessed with validated retrospective questionnaires. The PubMed query combined an extensive list of pharmacological agents (including classical psychedelics and other psychoactive drugs), a broad set of non-pharmacological induction techniques (for example meditation, breathwork, sensory deprivation), and the names and abbreviations of established psychometric instruments (Abnormal Mental States Questionnaire/APZ, 5D-ASC, 11-ASC, Hallucinogen Rating Scale/HRS, Mystical Experience Questionnaire/MEQ, and Phenomenology of Consciousness Inventory/PCI). Additional inclusion terms (e.g. phenomenology, psychometric, subjective experience) were added to capture relevant studies where questionnaire names might not appear in titles or abstracts, and exclusion terms were applied to remove animal, in-vitro, or otherwise irrelevant uses of abbreviations. The search was restricted to English-language publications dated 1975–2021 and used MeSH terms; the PubMed and ISI searches were performed on 2021-12-31. References retrieved from both sources were managed in Zotero. The initial yields were 6,414 items from PubMed and 614 from ISI Web of Science, totalling 7,028 records; 326 duplicates were merged, leaving 6,702 items for screening. Screening occurred in two stages: (1) title and abstract screening to exclude non-journal items, non-English, animal or in-vitro studies, reviews, studies using other questionnaires, or inaccessible articles, which removed 6,381 items and left 321 for full-text review; (2) full-text screening to assess whether accessible, extractable questionnaire data were reported, which excluded a further 38 full texts. The final extraction set included 165 journal articles. Data extraction recorded group-level means and standard deviations for each reported factor/dimension of the questionnaires. Where only the standard error of the mean was reported, it was converted to a standard deviation; where data were presented only graphically, values were digitised using WebPlotDigitizer v4.5. The authors defined an individual dataset as a unique combination of experimental conditions and questionnaire (for example, different induction methods or dosages in the same article could yield multiple datasets). The extracted data were organised into one Microsoft Excel file per questionnaire, structured to fit a mySQL database schema, and included fields such as number of subjects, control/experimental condition, questionnaire application time, and reference for the assessment. The OSF repository also contains a list of excluded studies with reasons for exclusion. To reduce bias, study selection and extraction were performed independently by multiple team members and cross-checked: screening was conducted independently by J.P. and J.D., and extraction was performed by J.P., E.D., and J.D. with subsequent cross-checking; discrepancies were resolved by consensus. No automated screening tools were used beyond duplicate detection in Zotero.

The systematic review yielded data from 165 journal articles comprising 674 datasets (individual experiments) and 4,689 group-level data points across all questionnaire factors and dimensions. These group-level entries derive from a total of 17,792 questionnaire applications to individual participants. An individual dataset corresponds to a unique combination of experimental conditions and questionnaire. Of the 165 included articles, 145 reported pharmacologically induced ASC and 20 reported non-pharmacologically induced ASC. Several articles reported results from more than one questionnaire. For pharmacological inductions, the most frequently reported instrument was the 5D-ASC (65 articles, 90 datasets, 1,792 applications), followed by the 11-ASC (43 articles, 144 datasets, 2,321 applications), the HRS (43 articles, 128 datasets, 1,804 applications), the MEQ30 (21 articles, 58 datasets, 5,183 applications), the APZ (15 articles, 45 datasets, 551 applications), and the PCI (1 article, 1 dataset, 22 applications). For non-pharmacological inductions, the PCI was most common (15 articles, 32 datasets, 1,450 applications), followed by the 11-ASC (4 articles, 7 datasets, 151 applications), the 5D-ASC (3 articles, 6 datasets, 133 applications), the APZ (2 articles, 4 datasets, 170 applications), and the MEQ30 (1 article, 9 datasets, 401 applications). Extracted numerical data included means and standard deviations for each reported factor/dimension; where necessary, standard errors were converted to standard deviations and graphical data were digitised. The complete dataset is hosted on the Open Science Framework in a folder labelled "ASDB_v2.0_12-2021" and is also accessible and visualisable via the ASDB website. The OSF files are organised per questionnaire and formatted to support import into a relational database, and the repository additionally provides a documented list of excluded studies with reasons for exclusion.