Developing Methods for Observing Awe Narration in Psilocybin-Assisted Therapy.
Elise C Tarbi, Ian Bhatia, Nabil Balach, Suzannah Buehler, Magdalena Demeo-Meres, Cailin Gramling, Tej Thambi, Julia Hart, Maija Reblin, Donna M Rizzo, Robert Gramling, Manish Agrawal, Emily Manetta
Healthcare (Basel, Switzerland) June 5, 2026 DOI: 10.3390/healthcare14111589 via PubMed
Summary
AI-generated from the abstractA direct observation coding system can reliably identify moments when people narrate experiences of awe during psychedelic-assisted therapy. In 32 video-recorded therapy sessions from a Phase 2 trial of psilocybin-assisted therapy for advanced cancer, two coders independently identified 246 moments of awe narration across 16,760 minutes. Coders were substantially more confident when vastness was present (odds ratio 4.3). The coding system was refined to split accommodation into two components: initial cognitive disruption and later engagement with that disruption. Awe narration is observable using explicit definitional criteria, providing a foundation for larger-scale studies.
Study at a glance
| Characteristics | Observational study using direct observation coding of clinical trial recordings Peer reviewed |
|---|---|
| Sample size | 8 |
| Population | Participants with advanced cancer in a Phase 2 clinical trial of psilocybin-assisted therapy |
| Topics | Psilocybin Psychedelic-assisted therapy |
| Keywords | Awe Interpersonal communication Psychedelics |
| Key finding | Awe narration during psilocybin-assisted therapy can be reliably identified by human coders using definitional criteria, with coders' confidence significantly higher when vastness is present. |
Abstract
Background: Understanding the benefits of psychedelic-assisted therapy (PAT) will require scientific attention to the causal interaction between the therapeutic context and process. Measuring what actually happens during PAT in large-scale studies will be an essential component of this work. Objective: We aim to develop and preliminarily evaluate the feasibility and reliability of a direct observation coding system for narrations of awe experiences during PAT, one hypothesized therapeutic mechanism. Methods: We analyzed 32 PAT clinical trial encounter recordings involving eight participants from a Phase 2 clinical trial study of psilocybin-assisted therapy in advanced cancer. Using a conceptually grounded structured codebook, two human coders independently identified start and stop times for moments exhibiting definitional characteristics of awe narration, including expressions of vastness, need for accommodation and ineffability. We used coder agreement and degree of confidence to refine the coding system. Results: During 16,760 total minutes of video, coders collectively recorded 246 moments of awe narration. Of those moments, 42% (104/246) were identified by one coder and 58% (142/246) by two coders. Coders felt substantially more confident in their judgments about a moment of awe when vastness was present compared to when vastness was absent (OR: 4.3; 95% CI: 2.4, 7.8). Iterative refinement of the coding system led to accommodation being operationalized as two distinct components: an initial cognitive disruption followed by variable engagement in the process of accommodation. Conclusions: Awe narration is directly observable using explicit definitional criteria. This work provides the empirical foundation for scalable coding systems of awe narration during PAT.