Signal-Detection Theory and the Neuroscience of Recognition Memory
I used my opportunity at the conference in honor of Henry Roediger to weighin on an issue that is dear to my heart but not to his, namely, the validity andutility of the signal-detection theory of recognition memory. This venerable theory has been applied to problems from virtually every domain of experimental psychology. With regard to recognition memory, the idealized version of signaldetection theory involves two equal-variance Gaussian distributions (one representing targets and the other representing lures) and one decision criterion. Any test item that generates a memory strength exceeding the criterion is declared to be Old, otherwise it is declared to be New (as illustrated in the upper panel of Figure 5.1). Although the idealized equal-variance detection model is often used to illustrate signal-detection theory, much evidence suggests that a quantitatively more accurate version of the theory is an unequal-variance model in which the standard deviation of the target distribution somewhat exceeds that of the lure distribution (Egan, 1958, 1975; Ratcliff, Shue, & Gronlund, 1992), as illustrated in the lower panel of Figure 5.1.