The content and structure of dreams are coupled to affect

Luke Leckie,Anya K. Bershad,Jes Heppler,Mason McClay,Sofiia Rappe,Jacob G. Foster

2024-09-22

Abstract:Dreams offer a unique window into the cognitive and affective dynamics of the sleeping and the waking mind. Recent quantitative linguistic approaches have shown promise in obtaining corpus-level measures of dream sentiment and topic occurrence. However, it is currently unclear how the affective content of individual dreams relates to their semantic content and structure. Here, we combine word embedding, topic modeling, and network analysis to investigate this relationship. By applying Discourse Atom Topic Modeling (DATM) to the DreamBank corpus of >18K dream reports, we represent the latent themes arising within dreams as a sparse dictionary of topics and identify the affective associations of those topics. We show that variation in dream affect (valence and arousal) is associated with changes in topical content. By representing each dream report as a network of topics, we demonstrate that the affective content of dreams is also coupled to semantic structure. Specifically, positively valenced dreams exhibit more coherent, structured, and linear narratives, whilst negatively valenced dreams have more narrative loops and dominant topics. Additionally, topic networks of high arousal dreams are structurally dominated by few high arousal topics and incoherent topical connections, whereas low arousal dreams contain more loops. These findings suggest that affective processes are associated with both the content and structure of dreams. Our approach showcases the potential of integrating natural language processing and network analysis with psychology to elucidate the interplay of affect, cognition and narrative in dreams. This methodology has broad applications for the study of narrated experience and psychiatric symptomatology.

Neurons and Cognition

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: **How are the emotional contents of dreams related to their semantic contents and structures?**

Specifically, the authors focus on:

1. **Affective content**: The emotional states in dreams (such as positive or negative emotions, and the intensity of emotions).
2. **Semantic content**: The themes and topics involved in dreams.
3. **Structure**: The semantic structure and coherence in dream narratives.

Current research mainly focuses on analyzing a large number of dream reports through natural language processing (NLP) techniques to reveal the relationships between these aspects. However, existing methods can usually only obtain general - ized statistical information of the entire corpus, and are unable to capture the changes within a single dream, such as the changes in affective and semantic arcs.

To solve this problem, the authors combined **word embedding**, **topic modeling** and **network analysis**, and in particular, used **Discourse Atom Topic Modeling (DATM)** to study the relationship between the emotional content of dreams and their semantic content and structure. Specific steps include:

- **Word embedding**: Map words to a high - dimensional semantic space, so that semantically similar words are closer in the space.
- **Topic modeling**: Decompose dream texts into a series of topics through DATM, and analyze the emotional load (valence and arousal) of these topics.
- **Network analysis**: Represent each dream as a topic network, where nodes represent topics, edges represent transitions between topics, and quantify the semantic structure of dreams through network properties (such as modularity, path length, etc.).

Through these methods, the authors demonstrated that:

- There is a significant association between the emotional content (valence and arousal) of dreams and their semantic content.
- The emotional content of dreams also affects their semantic structure, for example:
- **Dreams with positive emotions** tend to have more coherent, structured and linear narratives.
- **Dreams with negative emotions** contain more narrative loops and dominant themes.
- **Dreams with high arousal** are dominated by a few high - arousal topics, and the semantic connections are incoherent.
- **Dreams with low arousal** contain more cyclic structures.

These findings not only reveal the dynamic relationships between dream emotions, content and structure, but also provide a new quantitative method that can be applied to the study of self - reported data in other fields.