Computational characterization of metacognitive ability in subjective decision-making

Area of Science:

• Cognitive psychology
• Neuroscience
• Decision science

Background:

• Metacognition involves reflecting on and controlling one's own thoughts and behaviors.
• Metacognitive ability is typically assessed by modeling confidence reports and choice behavior in objectively measurable tasks.
• Prior research investigated metacognitive ability conservation across tasks and domains like perception and memory.

Purpose of the Study:

• To determine if metacognitive assessment methods can extend to value-based decision-making, where objective accuracy is not evaluable.
• To compare metacognitive ability across diverse tasks, including perception and valuation.
• To utilize Bayesian hierarchical estimation for a computational model of confidence.

Main Methods:

• Employed Bayesian hierarchical estimation of a computational process model of confidence.
• Modeled metacognitive ability as an individual's uncertainty about their own decision uncertainty.
• Applied the model to decision variables representing both external features and internal subjective states.

Main Results:

• Metacognitive ability was reliably estimated in both objective and subjective decision-making tasks.
• Metacognitive ability demonstrated relative conservation across different tasks.
• Conservation was particularly notable within specific domains and similar confidence reporting paradigms.

Conclusions:

• The developed computational model effectively assesses metacognitive ability in diverse decision-making contexts.
• Metacognitive ability appears to be a relatively stable trait conserved across various tasks and domains.
• Findings support the generalizability of metacognitive assessment beyond tasks with objective performance measures.