Equine utilization of secondary reinforcement during response extinction and acquisition

Forty-eight horses were used to examine equine utilization of a secondary (learned) reinforcement signal. Phases I and 11 of the study investigated whether the secondary reinforcer could prolong extinction of a learned task. Phase III utilized the secondary reinforcer to train the horse to perform a new task. Horses were paired by age, sex and breed. All horses were taught to push a lever to obtain a feed reward. One horse of each pair served as the control (CON, feed reward only, no secondary reinforcement) and the other was given a feed reward paired with an auditory buzzer as a secondary reinforcer (SR). In phase 1, horses were given 30 trials of continuous reinforcement daily for 3 days. For SR horses, the feed reward was delivered approximately Is after the buzzer sounded. On days 4 and 5, extinction trials occurred in which horses did not receive any feed reward for pushing the lever, but the buzzer was still delivered to SR horses. Extinction of the learned response was defined as 5 min with no lever press. Responses and time to extinction were recorded for each horse. After 2 days of rest the horses were retrained to the lever and this procedure was repeated with the feed reward delivered simultaneously with 3-4 s of the SR (phase 11). In phase III horses were given 40 reinforcements on a variable ratio 5 reinforcement schedule for 3 days. On days 18 and 19, the lever was removed and the horses were shaped for 30 min daily to a new task (push a flap) using the SR. Numbers of reinforcements and responses were recorded for each horse. Data from CON and SR horses in all phases of the study were compared utilizing a paired t-test. During phases I and 11, no differences were found in responses (P > 0.9, P > 0.3, respectively) or in time to extinction (P > 0.3, P > 0.7, respectively). In phase 111, SR horses had more reinforcements during shaping on day 18 than CON horses (mean = 28.0 and 11.8, respectively, S.E.D. = 5.4, P < 0.01). The SR horses also had more responses than CON horses on day 18 (mean = 11.9 and 4.5, respectively, S.E.D. = 3.6, P < 0.05). Results suggest that secondary reinforcements may not prolong extinction of a learned response in horses; however, they may facilitate learning. (C) 2002 Elsevier Science B.V. All rights reserved.