Components of Sensitivity

Detection theory provides strategies for partitioning the variance that limits performance between external and internal sources, and among subcategories of each. An important tool is the distribution-discrimination task, in which participants must determine which of two overlapping distributions led to an observation. The magnitude of the external noise arises from the overlap, and the best possible ideal performance is what would be found if this were the only limiting factor. If accuracy falls below this level, the discrepancy is attributed to internal noise. When stimulus classes vary on more than one dimension, external variance can be divided among the dimensions. Discrimination can be described with multidimensional detection theory, and the decision bound between the distributions depends on both dimensions. The weighting assigned to a dimension depends on the variability of the distributions along both dimensions (for the ideal observer) and on the slope of the psychometric conditional-on-a-single-stimulus (COSS) functions (for real observers).

Electronic engineering · Sensitivity (control systems) · Embodied and Extended Cognition · Engineering · Free Will and Agency · Philosophy and Theoretical Science