Measuring prediction errors in dreams: the Dream Prediction Error Scale (DPES)
Sophie Boudrias, Michael Schredl
International Journal of Dream Research May 26, 2026 DOI: 10.11588/ijodr.2026.2.116071 (opens in new tab)
Study at a glance
AI-extracted from the abstract| Characteristics | Scale development study Peer reviewed |
|---|---|
| Sample size | 50 |
| Population | Dream reports |
| Key points | The authors report that interrater agreement was sufficiently good for identifying the presence or absence of a prediction error in dream reports and for judging its emotional valence. Reliability for identifying specific prediction errors was lower, and the authors state that further research is needed to improve it. |
Abstract
Prediction error is a discrepancy between a prior expectation and what is experienced. Predictive processing posits that prediction errors lead to updates of previously acquired models, beliefs, and expectations. They can be valenced (i.e., better or worse than expected) or non-valenced. The presence of prediction errors in dreams remains debated, while dream reports do contain surprise related content. This article aims to develop a scale for measuring prediction errors and their valence in dreams. Interrater agreement was measured on a sample of 50 dream reports, analyzed by three independent raters after training. The results indicate sufficiently good interrater reliability regarding the presence or absence of a prediction error in a dream report, as well as the emotional valence of prediction errors. Further research is needed to improve the interrater reliability in identifying specific prediction errors. The scale can be used for research on dreams and consciousness and could have clinical implications.