Judgment Bias Responses in Domestic Dog Puppies at 60 Days of Age

Judgment bias tasks are widely used to assess affective state in adult dogs but have rarely been applied before adoption age. To test whether a spatial version is viable, 82 puppies from 17 litters were tested at 60 days of age; 80 provided valid observations for analysis. After discrimination training, positive (P), negative (N) and ambiguous (A) locations were presented in up to three trials; latency to the bowl was recorded (30 s cut-off). Latencies from the first two trials (476 observations, 85 right-censored) were analyzed with mixed-effects Cox proportional hazards models including puppy and litter as random intercepts. Median recorded times were 2.54 s (P), 13.25 s (N) and 3.86 s (A), with timeouts in 0%, 38.4% and 15.2%, respectively. Cue location affected approach rate (χ² = 207.03, df = 2, p < 0.001); hazard ratios were N versus P 0.11 (95% CI 0.09-0.15), A versus P 0.32 (0.25-0.41) and N versus A 0.36 (0.27-0.47), all significant (Holm-adjusted). Trial number (p = 0.076) and sex (p = 0.483) were not significant. Random-effect variances were 0.180 (puppy) and 0.261 (litter). Puppies discriminated between outcome-associated locations and approached the ambiguous location at an intermediate rate, indicating that the task is feasible at 60 days of age.