Memory and Expectancy in Delayed Discrimination Procedures
Diane L. Chatlosh
California State University, Chico Edward A. Wasserman
University of Iowa
In a delayed conditional discrimination task, an organism must base its responses on previously presented stimuli. For example, consider a typical delayed matching-to-sample (MTS) task for pigeons. The occurrence of a red or a green center keylight signals whether a peck to a red or a green side keylight will later produce food. The red or green center keylight is the sample stimulus for the trial. The offset of the sample stimulus is followed by a brief delay during which no events are scheduled. Once this retention interval has elapsed, red and green test stimuli are simultaneously presented on the two side keys. The red and green side keylights are the test stimuli; their locations (left and right) are reversed on a random half of the trials. The bird's task is simply to peck the red test stimulus after a red sample and to peck the green test stimulus after a green sample, regardless of their locations on the side keys on a particular trial. Performance on a delayed MTS task is thought to be a reflection of pigeons' short-term memory; elucidating the nature of this short-term memory is the goal of many studies of animal cognition.
What kind of memory mechanism is it that makes it possible for subjects to respond correctly during the choice test of a delayed discrimination task? According to one view of short-term memory, a sample stimulus forms a memory trace, which is thought to be an isomorphic representation of the actual stimulus (e.g., Roberts & Grant, 1976; for a somewhat different interpretation, see Roitblat, 1980). The trace is thought to grow while the stimulus is presented and