IMPROVING CONDITIONING WITH UNSIGNALED, BUT NOT SIGNALED, UNCONDITIONED STIMULI

TWO experiments evaluated whether tone-food conditioning in Long-Evans rats could be improved by adding unsignaled, but not signaled, food p resentations. One group of subjects first received 10 food pellets at regular 10.5-s temporal intervals in the middle of a session. After th is temporal regularity was learned, later conditioning to a tone was p oor when the tone signaled the second food pellet. Two other groups al so received 10 food Pellets at regular 10.5-s temporal intervals and t he tone signaling the second food pellet. However, for one of these tw o other groups, subjects received 30 extra food pellets delivered at m ore widely spaced 1-min intervals. Tone conditioning was significantly stronger in this second group. A third group was identical to the sec ond, except that each of the 30 extra food pellets was signaled by a l ight. This third group learned that food pellets were delivered at 10. 5-s intervals but food pellets signaled by lights were not. Tone condi tioning was also poor in this third group. Session length was identica l for all groups and results were similar whether tone conditioning wa s measured by direct behavioral observation (Experiment 1) or conditio ned reinforcement testing (Experiment 2). Results violated a contingen cy view of learning and were unanticipated by current conditioning the ories. An explanation, however, was provided by acknowledging the bloc king ability of an unconditioned stimulus (US) when subjects learn a U S-US association (Goddard & Jenkins, 1988). (C) 1994 Academic Press, I nc.