Stimulus Control of Odorant Concentration: Pilot Study of Generalization and Discrimination of Odor Concentration in Canines

Abstract

Despite dogs’ widespread use as detection systems, little is known about how dogs generalize to variations of an odorant’s concentration. Further, it is unclear whether dogs can be trained to discriminate between similar concentration variations of an odorant. Four dogs were trained to an odorant (0.01 air dilution of isoamyl acetate) in an air-dilution olfactometer, and we assessed spontaneous generalization to a range of concentrations lower than the training stimulus (Generalization Test 1). Dogs generalized to odors within a 10-fold range of the training odorant. Next, we conducted discrimination training to suppress responses to concentrations lower than a concentration dogs showed initial responding towards in Generalization Test 1 (0.0025 air dilution). Dogs successfully discriminated between 0.0025 and 0.01, exceeding 90% accuracy. However, when a second generalization test was conducted (Generalization Test 2), responding at the 0.0025 concentration immediately recovered and was no different than in Generalization Test 1. Dogs were then tested in another generalization test (Compound Discrimination and Generalization) in which generalization probes were embedded within discrimination trials, and dogs showed suppression of responding to the 0.0025 concentration and lower concentrations in this preparation. These data suggest dogs show limited spontaneous generalization across odor concentration and that dogs can be trained to discriminate between similar concentrations of the same odorant. Stimulus control, however, may depend on the negative stimulus, suggesting olfactory concentration generalization may depend on relative stimulus control. These results highlight the importance of considering odor concentration as a dimension for generalization in canine olfactory research.

Expert considerations and consensus for using dogs to detect human SARS-CoV-2-infections

“…Dogs can even be trained to certain concentration differences of the target scent ( 228 ). For example, this is used in diabetes alert dogs, which detect increases or decreases of blood glucose values of patients beyond predetermined levels ( 229 ).…”

Case Study: An Evaluation of Detection Dog Generalization to a Large Quantity of an Unknown Explosive in the Field

“…Nonetheless, the substantially poorer alert rate in Experiment 1 (28%), suggests that the inability of the deployed dogs to detect the 13 kg explosive could have been due also to its large quantity, which would also agree with previous literature. For instance, using an air dilution olfactometer, DeChant et al found that dogs spontaneously generalized to a change in concentration of only approximately 10 fold [ 13 ]. In our case, a 13 kg sample represented an increase of 433 fold relative to the quantity dogs were trained to detect.…”

“…Olfactory generalization refers to the phenomenon where a dog produces the same response to perceptually similar odors that could share the same outcome [ 11 , 12 ]. In the field, olfactory generalization occurs when a dog alerts or shows the same trained behavioral response to variations of a target odor [ 13 , 14 ]. Olfactory discrimination is the opposite phenomenon where dogs do not alert to variations of the trained odor [ 11 , 14 ].…”

“…The concentration of the odorant used in training is also an important dimension to consider, because a recent study found that dogs will generalize spontaneously to a change in concentration not greater than 10 fold [ 13 ]. Further, explicitly training dogs for lower concentrations leads to improvements in odor detection thresholds [ 20 ].…”

Can scent-detection dogs detect the stress associated with trauma cue exposure in people with trauma histories? A proof-of-concept study

“…The dog received a reward after an error correction, but the trial was recorded as incorrect. As the difficulty level was very high in training, error corrections were to reduce the dog's frustration/confusion and maintain motivation ( 88 ). Error corrections were only used as much as necessary and as little as possible to prevent the dog from developing an alternative strategy of signalling different samples until hearing the click.…”