Generalized Gaussian signal detection theory: A unified signal detection framework for confidence data analysis

Area of Science:

• Cognitive psychology
• Decision science
• Computational neuroscience

Background:

• Human decision-making involves confidence, a graded awareness of certainty.
• Measuring metacognitive efficiency (confidence vs. accuracy) is crucial but challenging.
• Existing frameworks have limitations in quantifying this relationship.

Purpose of the Study:

• To introduce a novel signal detection theory paradigm, the generalized Gaussian distribution (GGSDT).
• To provide a robust framework for quantifying metacognitive efficiency.
• To enable new research protocols for comparing metacognitive performance.

Main Methods:

• Developed a new signal detection theory paradigm using the generalized Gaussian distribution (GGSDT).
• Utilized shape and scale parameters to evaluate metacognitive efficiency and internal standard deviation ratio.
• Interpreted the shape parameter in relation to the metacognitive lapse rate, independent of decision accuracy.

Main Results:

• The GGSDT framework effectively quantifies metacognitive efficiency.
• The shape parameter provides a measure of metacognitive lapse rate, robust to decision accuracy variations.
• The framework allows for novel research designs, such as comparing different decision tasks.

Conclusions:

• The GGSDT offers a unique and powerful tool for assessing metacognitive efficiency.
• This framework has broad applicability across various fields of behavioral science.
• An accompanying R package (ggsdt) facilitates its implementation and analysis.