Can scent-detection dogs detect the stress associated with trauma cue exposure in people with trauma histories? A proof-of-concept study.
Level 4 - case-series / case-control
Proof-of-concept diagnostic scent-detection study with a small observational cohort of donors and two canine subjects
PubMed 38606088 · doi:10.3389/falgy.2024.1352840
What was done
Breath samples were collected from 26 humans with trauma histories (54% with PTSD) over 40 experimental sessions during a calm state (control) and a stressed state induced by personalized trauma cues (target). Two trained scent-detection dogs were evaluated using a two-alternative forced choice (2AFC) discrimination task to distinguish between states within individual donors, and a yes/no detection task to assess generalization across different donors and stress events. Signal Detection Theory was used to analyze performance alongside donor self-reported emotional responses.
What was found
The dogs achieved ~90% accuracy across all sample sets in the 2AFC discrimination experiment. In the cross-donor detection experiment, the dogs achieved 74% and 81% accuracy, respectively. One dog's performance correlated with donor self-reported fear, while the other dog's performance correlated with self-reported shame.
Why it matters
This proof-of-concept study provides initial evidence that dogs can detect olfactory signatures associated with acute trauma-cue-induced distress in human breath, offering biological support for PTSD service dog alert functions.
Limits
The canine sample was limited to only two dogs, restricting generalizability across animals. Specific volatile organic compounds were not chemically identified, the human sample was modest (n = 26), and laboratory breath-sniffing tasks may not directly translate to real-world service dog alerts.
Cited by
- supports Dogs are capable of detecting the autonomic nervous system tone and physiological states of humans.