Multimodal characterization of detection dog suitability: Combining behavioral, cognitive, and neurological measures for enhancing selection – Presented by Lucia Lazarowski

The growing demand for highly capable detection dogs and a lack of robust selection measures continue to pose a challenge to the industry.  The ability to predict success from an early age is critical to maximizing program efficiency. Since 2000, Auburn University has bred and raised Labrador retrievers to be employed as single-purpose detection dogs, constituting the longest continual running institutional program for producing detection dogs in the United States. Our first aim in this study was to validate a traditional standardized behavioral test for evaluating puppies’ (n= 60) future suitability as a detection dog.  Validation was achieved by demonstrating high inter-rater reliability across observers, convergence with other measures of dog behavior, and prediction of adult outcomes as early as 3 months of age. In subsequent efforts we have explored non-traditional measures to examine the bio-behavioral correlates of successful detection dog phenotypes. Utilizing a reinforcement sensitivity theory approach, we characterized approach-avoidance tendencies in n = 56 adult detection dogs. Metrics included behavioral, cognitive, emotional, and physiological measures. Additionally, neural activity will be recorded in a subset of dogs (n = 18) trained for awake and unrestrained fMRI. Patterns in responding across measures suggests that multi-modal assessments may be useful in enhancing the identification of desirable working dog phenotypes.