The efficacy of bridging stimuli during acquisition of an operant task and the use of food-based positive reinforcement training on unwanted oral investigative behaviors in horses, Equus caballus

Wednesday, July 23, 2014
Exhibit Hall AB (Kansas City Convention Center)
Megan R LaFollette , Truman State University, Kirksville, MO
Kaitlin A Cloonan , Truman State University, Kirksville, MO
Kelly W. Walter , Truman State University, Kirksville, MO
Abstract Text:

This study sought to determine the impact of food-based positive reinforcement training (PRT) on frequency and severity of unwanted oral investigative behaviors (UOIB) in horses. It also investigated the influence of various bridging stimuli (BS) on time and number of reinforcements to behavioral acquisition of an operant task. Eighteen horses were used in a randomized complete block design, where they were split into six blocks by age and assigned to one of three treatments at random. Treatments consisted of mechanically-produced BS, human-produced BS, or no BS. A standardized training protocol was used to train each individual to touch a target. Lag time between BS and food delivery was minimal. Before and after training, all horses were evaluated for number and type of UOIB (nose, lips, or teeth on skin, clothes, or treat bag) at 10 s intervals for 5 min.  These UOIB were later given a numerical value based on type and summed to evaluate overall severity. These data were analyzed using PROC MIXED of SAS. Treatment did not influence time to behavioral acquisition (TBA) (p = 0.7682) or number of reinforcements (p = 0.8881).  However, the youngest block had the shortest TBA (p = 0.0599) and received the most treats per minute (p = 0.0207). Before training, the youngest and oldest blocks tended to have fewer UOIB (p = 0.0692) and less overall severity (p = 0.0653) UOIB compared to the rest of the blocks. Analysis using a two-tailed t-test showed that, after training, horses had an average of 4.4 fewer UOIB (p = 0.0311) as well as a reduction in overall severity of UOIB (p = 0.0473). Our data indicates that use of food-based PRT decreased number and overall UOIB, which suggest that proper delivery of food rewards can allow for successful task acquisition without causing UOIB. The lack of treatment influence on TBA could be due to minimal lag time between BS and presentation of the food reward. Future research evaluating if the presence of a BS with a lag time in food delivery influences TBA is recommended.

Keywords: Operant conditioning, Food rewards, Clicker training,