Comparing the characteristics of hallucinations and mental imagery: a large cross-sectional study in the general population.
Guillaume Pepin, Hélène Lœvenbruck, Alan Chauvin, Camille Jacquet, Jean-baptiste Eichenlaub, Catherine Bortolon
Consciousness and cognition January 1, 2026 Peer reviewed DOI: 10.1016/j.concog.2025.103974 via PubMed
Summary
Hallucinations and involuntary mental imagery share many features but differ in key ways. In a survey of 1,951 French-speaking adults, involuntary mental imagery was more frequent, vivid, emotionally positive, and felt self-generated. Hallucinations were more distressing and attributed to external sources. Moderate to strong correlations between the two on most dimensions support a shared experiential continuum, but differences in confidence and agency challenge existing cognitive models.
Study at a glance
| Design | cross-sectional study |
|---|---|
| Sample size | 1,951 |
| Population | French-speaking adults |
| Key finding | Involuntary mental imagery is more frequent, vivid, and self-generated than hallucinations, while hallucinations are more distressing and externally attributed, yet both share a phenomenological continuum. |
Abstract
Hallucinations (HAs), the perceptual experiences that occur without external stimuli, are classically associated with psychiatric disorders, but also occur in the general population. Involuntary mental imagery (IMI), the spontaneous occurrence of visual or auditory mental content, shares several features with HAs. The continuum hypothesis suggests that these phenomena exist on a spectrum ranging from benign to clinically significant experiences. However, the precise boundaries of these phenomena remain under-explored. This pre-registered cross-sectional study focusing on the experiential continuum, involved 1,951 French-speaking adults who completed an online questionnaire assessing the frequency and characteristics of their HAs and IMI experiences. These experiences could include hearing music or voices, or seeing shadows. Participants rated ten phenomenologically-informed dimensions, including vividness, emotional valence, controllability, distress, agency, and perceived location. Associations were tested using Spearman correlations and Wilcoxon tests with Bonferroni correction. IMIs were significantly more frequent than HAs and were rated as more vivid, emotionally positive, and self-generated. In contrast, HAs were associated with greater distress and external attribution. Moderate to strong correlations between HAs and IMIs on most dimensions support the hypothesis of a shared experiential continuum. However, the fact that confidence and agency cannot be fully separated from controllability raises questions about cognitive models of self-monitoring and metacognition. These findings provide empirical evidence for both convergence and divergence in the phenomenological profiles of HAs and IMIs. Refining these distinctions could enhance theoretical models of hallucinatory experiences and facilitate the early identification and prevention of distressing internal phenomena in individuals at risk.