BACK TO QUEBEC CITY 2002 TITLES
A previously-learned conditional
rule inhibits acquisition of a new rule by five Holstein heifers learning
an operant
social discrimination
task
J.M. Watts, J.M. Stookey,
J.L. Clavelle and D.B. Haley
University of Saskatchewan,
Saskatoon, Canada
We tried to teach heifers to discriminate between conspecifics based on social familiarity. Intending to simplify this, we trained the response first to a familiar and later to an unfamiliar animal. The testing arena contained two pens capable of holding a stimulus animal. A switch-operated dispenser was mounted in front of each pen. Switch-pushing was taught by shaping.
In phase 1 heifers learned to respond to a penmate in one holding pen. Heifers trained to associate penmate with reward had to push the switch on the occupied pen. Heifers trained to non-penmate had to push on the empty pen’s switch. If wrong the subject exited the pen and re-entered with conditions unaltered. If correct she consumed the food reward and was removed while conditions were reset for the next trial. The reward location was assigned randomly to either pen 5 times/heifer/test session. In phase 2 they were shown a non-penmate in one of the pens. To be correct, heifers that previously were required to push in front of the stimulus animal, now needed to push in front of the empty pen, and vice-versa.
Heifers differed in the numbers of errors they made (Kruskall-Wallis Statistic=12.72, P=0.0127). During phase 1 the group error rate declined over 18 successive sessions (Spearman, r=-0.9241, P<0.002) from 24.75 errors/session (first 4 sessions) to 1.5/session (last 4 sessions). During phase 2 the group error rate declined (Spearman, r=-0.8218, P<0.002), but not as far or as rapidly as before. The initial group error rate (first 4 sessions) was 29. After 23 sessions the rate (last 4 sessions) was 7.75. Despite five extra sessions, the mean error rate per session for the group was higher (13.4±5?2.08) compared with 9.28±?2.4) in phase two (Mann-Whitney, U=276, P=0.0351).
Although motivated to seek the reward, the animals probably learned first that the reward’s location was related to the presence or absence of an animal, not its social familiarity. In phase 2, it thus seemed that the rule governing the location of the reward was reversed. For these animals, learning the original rule was easier, perhaps because no pre-existing rule had to be unlearned.