PT  - JOURNAL ARTICLE
AU  - Balázs Hangya
AU  - Joshua I. Sanders
AU  - Adam Kepecs
TI  - A mathematical framework for statistical decision confidence
AID  - 10.1101/017400
DP  - 2015 Jan 01
TA  - bioRxiv
PG  - 017400
4099  - http://biorxiv.org/content/early/2015/04/01/017400.short
4100  - http://biorxiv.org/content/early/2015/04/01/017400.full
AB  - Decision confidence is a forecast about the probability that a decision will be correct. For human decision makers, confidence is a deeply subjective sense that can be difficult to study due to its inherently introspective nature. However, confidence can be framed as an objective mathematical quantity – the Bayesian posterior probability, providing a formal definition of statistical decision confidence. Here we use this definition as a starting point to develop a normative statistical framework for decision confidence. We analytically prove interrelations between statistical decision confidence and other observable decision measures. Among these is a counterintuitive property of confidence – that the lowest average confidence occurs when classifiers err in the presence of the strongest evidence. These results lay the foundations for a mathematically rigorous treatment of decision confidence that can lead to a common framework for understanding confidence across different research domains, from human behavior to neural representations.