Effectiveness of marker training for detection dogs.

Abstract

Training detection dogs to alert to an odor requires precision in the timing and delivery of stimulus presentations in order to condition a strong association between odor and reward and to train a desired alert behavior that communicates the presence and location of the odor source. Marker training, in which a signal that predicts a reward is used to deliver immediate feedback for a correct response and bridge the delay between the desired behavior and reward, is a popular technique in the animal training industry. However, the application of marker training to detection dog training has not been examined, and empirical evidence of the purported benefits of marker training in general is lacking. The current study evaluated the effectiveness of marker training for odor detection learning and performance. Candidate detection dogs (n = 28) were trained to detect and alert to a target odor either with or without the use of a clicker as a marker (n = 14 per group). Effectiveness of marker training was assessed by comparing rate of learning the odor discrimination and the alert response, detection accuracy and topography of the alert behavior in an odor discrimination test, generalization of learned behavior from the odor recognition setting to a novel context (i.e., open-area operational searches), and resistance to extinction. Compared to dogs trained with the reward only, dogs trained with the marker as a signal for reward completed the training phase in fewer trials, performed the alert response more accurately in the odor recognition test, indicated the location of the odor source more precisely in the operational searches, and exhibited greater resistance to extinction when the primary reward for a correct response was withheld. These results provide evidence supporting the effectiveness of markers in animal training, and demonstrate benefits specific to the challenges commonly faced in detection dog training.