Behav Pharmacol. 1991 Nov;2(4):379-389.
Behavioural pharmacology
D.E. McMillan, W.C. Hardwick
PMID: 11224081
An attempt was made to train pigeons to discriminate phencyclidine (PCP) from saline using a three-key color-tracking procedure under which birds were trained under a second order schedule [FR10 (FR5)] "without errors." Training without errors was done by not lighting the side key on which responses were not reinforced during the early phases of discrimination training and then gradually increasing the light intensity on that key. Initially, this procedure prevented drug discrimination errors, but as the light intensity increased on the incorrect side key, errors began to appear and then surpassed the error rate of birds trained "with errors." When stimulus control stabilized, there were no differences in stimulus control by the phencyclidine stimulus as a function of the training procedure, as measured by error rates. Birds trained to discriminate PCP from saline with and without errors also showed no differences in the shape of the PCP stimulus generalization curve. A second attempt at training without errors under a second-order schedule was made in a second group of pigeons, using cocaine as the training drug and eliminating the use of the color-tracking procedure. The results were very similar to those with PCP. Few errors occurred in birds trained "without errors" until the incorrect key-light intensity was increased, at which time large numbers of errors began to occur, but ultimately differences in stimulus control as a function of the training procedure disappeared. There was no difference between birds trained with and without errors in the shape of the cocaine stimulus generalization curve. The pigeons trained to discriminate cocaine from saline under a second order FR 10 (FR 5) schedule showed complete generalization to d-amphetamine and partial generalization to phencyclidine and chlorpromazine. Because baseline stimulus control was deteriorating, the schedule was changed to a second order FR 4 (FR 25) and stimulus control by cocaine was reestablished. Under this schedule, partial generalization was observed from cocaine to pentobarbital and chlorpromazine before baseline stimulus control weakened again. Both the broad generalization of the cocaine stimulus and the difficulty in maintaining stimulus control with cocaine have been observed by others; such data suggest that under some conditions the cocaine stimulus may lack specificity.
