Using sniffing behavior to differentiate true negative from false negative responses in trained scent-detection dogs

Area of Science:

• Animal behavior
• Olfactory detection
• Canine science

Background:

• False negatives in chemical detection systems are a known issue.
• In detection dogs, false negatives can stem from handler errors or detection failures.
• Critical scenarios like search and rescue or explosives detection highlight the importance of accurate canine scent detection.

Purpose of the Study:

• To investigate differences in sniffing behavior of trained detection dogs.
• To differentiate between true positives, true negatives, false positives, and false negatives.
• To assess if sniffing patterns can improve the evaluation of canine scent detection performance.

Main Methods:

• Analysis of 200 video recordings of 10 working detection dogs.
• Frame-by-frame quantification of sniffing duration and sniffing episodes.
• Utilized a Go/No-Go single scent-detection task with an eight-choice apparatus.

Main Results:

• Sniffing duration for true negatives was significantly shorter than for false negatives, true positives, and false positives.
• Dogs exhibited only one sniffing episode towards true negatives.
• Two sniffing episodes were commonly observed for true positives, false positives, and false negatives.

Conclusions:

• Sniffing behavior patterns in detection dogs can differentiate between correct and incorrect scent detection outcomes.
• Quantifying sniffing duration and episodes offers a method to assess canine performance.
• Understanding these behavioral nuances can enhance the reliability of dogs as biological detection devices.