Back

Neural Convergence of Graded Belief and Binary Belief Uncertainty

Gerchen, M. F.; Ullrich, S.; Lessau, M.; Kirsch, P.

2025-07-18 neuroscience
10.1101/2025.07.14.664652 bioRxiv
Show abstract

Belief was defined by William James as "the psychological process or function of cognizing reality", and the recent philosophical literature has emphasized that there are two types of belief: categorical and graded. The relationship between these two belief types is complex, and it is often assumed that the degrees of graded belief reflect confidence. However, this claim has not yet been addressed empirically. In this functional magnetic resonance imaging study we let N=29 young healthy participants rate their graded belief in propositions from the conspiracy theory spectrum and estimated certainty scores from the ratings that we associated with brain activation during the presentation of the propositions. We found associations of uncertainty (i.e. negative associations with certainty scores) with brain activation in the dorsomedial prefrontal cortex, right dorsolateral prefrontal cortex, and posterior parietal cortex. The clusters in the dorsomedial and dorsolateral prefrontal cortex replicated prior associations with explicit ratings of uncertainty in binary belief decisions from a study on domain-general belief and the dorsomedial prefrontal cortex effect, localized in the anterior pre-supplementary motor area, corresponds to the concepts of Alexander Bain about the close relationship between belief and action and a central role of doubt in belief. Our results provide empirical evidence for a link between graded belief and uncertainty in human cognition and emphasize the role of doubt as a relevant process to take into account in concepts of graded belief. Highlights- Converging neural associations of graded belief levels and belief uncertainty - Conceptual replication of belief uncertainty associations with a modified procedure - Correspondence to classical concepts of belief by Alexander Bain and William James - Our results emphasize doubt as a relevant process in graded belief

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.