Directed forgetting in pigeons: The role of retention interval keypecking on delayed matching accuracy

Pigeons were trained on a delayed matching task involving presentation, during the delay on each trial, of one of three cues: on some trials a cue that signaled that a memory test would always occur (a remember, R, cue), on other trials a cue that signaled that it would never occur (a forget, F, cue), and on still other trials a cue that signaled that it would occur unpredictably on half of the trials (an ambiguous, A, cue). For birds in the omission group, absence of a memory test meant that the trial was over, whereas for birds in the substitution group, it meant that the trial ended with a simple, simultaneous discrimination. Pigeons in the substitution group pecked all cues equivalently and, relative to accuracy on R-cue trials, showed no decrement in matching accuracy on either A-cue- or F-cue-probe trials. Pigeons in the omission group pecked the R cue more than the A cue but matching accuracy on R- and A-cue trials was comparable. Results from the omission group suggest a dissociation between delay-cue pecking and matching performance.