Age-related differences in latent belief updating and its neural engagement
Yu-Shiang Su, Joshua Oon Soo Goh, National Taiwan University, Taiwan
Session:
Posters 3B Poster
Presentation Time:
Sat, 26 Aug, 13:00 - 15:00 United Kingdom Time
Abstract:
Human brains respond to incoming external information in a manner that can be formulated as Bayesian inference such that the posterior inference is a product of exploiting internal prior beliefs and integrating likelihood from external observations. We postulated that age alters the involved neural circuits that could affect either operands, and used a novel three-latent-state belief updating task with functional imaging scanning to probe age-related differences in such Bayesian processing in younger and older adults. Convergent behavioral responses and computational models showed that older adults have comparable prior beliefs for competing states such that they switched between alternatives whereas younger adults quickly settled on the correct latent states. No evidence supported older adults having disproportionate adherence to prior beliefs. Neural data analysis revealed that, compared to younger adults, older adults had less congruent representations of iterative beliefs in posterior parietal cortex that was accompanied with reduced medial orbital frontal cortex and hippocampus neural modulation by certainty of internal beliefs. Our behavioral and neural results suggest that older adults harbor less selective internal prior models about the environment compared to younger adults. These findings have important implications for understanding how information is encoded in aged intelligent systems.