Encoding and use of detail information in picture recognition

Abstract

Subjects participated in a yes/no picture recognition experiment in which exposure time varied from 50 to 1,000 msec at the time of initial study. Following each study trial, half of the subjects, the detail at study and test (ST) group, reported whether they had observed a detail in the picture that they thought might help in subsequent recognition. The other half of the subjects, the detail at test only (T) group, did not attempt to name details during study. All of the subjects reported at the time of each test picture whether they were basing their yes/no recognition response on a specific detail in the picture or on the picture's general familiarity. The data provided strong support for a model which assumed that (a) there is a constant probability of encoding a detail during each successive unit of time at study and (b) a detail is named at test either if it was encoded at study or with some bias probability. ST subjects showed superior recognition memory performance relative to T subjects. Within the context of the aforementioned model, this superiority stems from two sources: ST subjects encode details at a faster rate than do T subjects and an encoded detail provided a better discriminative feature for ST subjects.