Fig 1.
Four positions of a trained dog during the human scent line-up task.
1A: the dog sampled the target scent inside the jar. 1B: the dog was then required to walk the line from the starting point. 1C: the dog sniffed in all of the jars along the line. 1D: when an odour in the line matched the target sample, the dog lay down in front of the corresponding jar. An enlargement of the drawing representing the tripod we use in our procedure is represented at the bottom of the figure.
Fig 2.
Mean number of trials (± S.E.M.) to reach criterion at each step of initial training.
Number of dogs in the experimental groups: n = 5 in steps 1 to 3; n = 7 in step 4 and n = 9 in step 5. ***, ** and *: P < 0.001, 0.01 and 0.05 respectively, compared to step 5, with post-hoc Bonferroni tests.
Table 1.
Proportion of correct detections (sensitivity scores: number of Hits / [Hits + Misses]) and False Alarms (specificity scores: number of CR / [CR + FA]) throughout the 10 periods of continuous training.
Fig 3.
Effect of the type of scent combination used during continuous training on sensitivity scores.
Sensitivity scores represent the proportion of correct detections (number of Hits / [Hits + Misses]). Each curve represents the evolution of scores (± S.E.M.) according to period. Intergroup comparison: * indicates P < 0.05 between BS / BS; TS / TS versus BS / TS and TS / BS in periods 8 to 10. Intragroup comparison: ♦indicates P < 0.05 compared to scores obtained in other periods for BS / BS and TS / TS.