The use of an intermittent schedule of reinforcement to evaluate detection dogs' generalization from smokeless-powder

Abstract

Odor generalization is essential for detection dogs. Regardless of its importance, limited research is available on detection dog odor generalization. The objectives of this study were (1) evaluate the use of an intermittent schedule of reinforcement to assess generalization in dogs and (2) evaluate olfactory generalization from a single exemplar of smokeless powder (SP). Dogs (N = 5) were trained to detect SP in an automated olfactometer under an intermittent schedule of reinforcement where only 60% of correct responses were reinforced. After training, eight non-reinforced probe trials were inserted within a session. A total of 15 testing odors were evaluated across 15 consecutive sessions (one odor/session). Six of the testing odors were control and the remaining testing odors were objects indirectly exposed to SP, objects that contained or were directly exposed to SP, single-base SP and diphenylamine (the main volatile present in the headspace of SP). Dogs' response rate to all testing odors except for the cotton gauzes and t-shirt cloths exposed to the headspace of SP, the simulated IED, and Getxent tubes exposed to direct contact with SP were statistically lower than their response rate of actual SP. The response rate to SP was not different across all 15 testing sessions suggesting that the intermittent schedule of reinforcement, maintained dog motivation and performance. Data show that the outlined method is a good approach to study generalization in detection dogs. These results also highlight dog generalization to SP varieties and associated odors.

Keywords: Detection dog; Explosive generalization; Olfactory generalization; Reinforcement schedule.