Ceja, Vanessa; Ezzeldine, Yussuf; Peters, Megan AK; (2022) Models of confidence to facilitate engaging task designs. In: 2022 Conference on Cognitive Computational Neuroscience. (pp. pp. 24-27). Cognitive Computational Neuroscience

Preview

Text 0000024.pdf - Published Version Download (1MB) | Preview

Abstract

Decision confidence models classically depict decision-making circuitry as: 1) accumulating relative evidence for each choice alternative and 2) computing confidence estimates from the difference in evidence magnitude favoring each choice. Recently, however, new evidence suggests a dissociation between metacognitive (confidence) computations and those supporting low-level perceptual decisions, positing instead that confidence is predominantly influenced by evidence favoring the selected choice while simultaneously ignoring evidence for the non-selected choice. Low-level perceptual tasks completed by neurotypical subjects and/or within controlled experiments, coupled with computational modeling, have helped reveal the computations and brain areas involved, but we do not yet know to what degree these dissociations generalize to other types of perceptual or cognitive tasks or to clinical, developmental, or aging populations. Here, we begin to tackle this issue by proposing a task and computational modeling comparison framework aimed at understanding whether perceptual confidence computations are stable across varying levels of perceptual judgements, in service of creating more engaging tasks for use in wider and more diverse populations.

Download activity - last month

Export as CSVJSONXML

Download activity - last 12 months

Export as JSONXMLCSV

Downloads by country - last 12 months

Export as CSVJSONXML

Archive Staff Only

View Item