# An exploratory study of associations between judgement bias, demographic and behavioural characteristics, and detection task performance in medical detection dogs

## TL;DR

This study explores how medical detection dogs' decision-making styles and behaviors relate to their performance in detection tasks.

## Contribution

It identifies novel associations between dogs' 'optimistic' or 'pessimistic' decision-making and their detection task performance.

## Key findings

- Older dogs and those with higher confidence, food orientation, and playfulness showed better detection task performance.
- Optimistic decision-making in judgement bias tasks correlated with higher detection ability in trainee dogs.
- Pessimistic dogs tended to have higher scent specificity in trained biodetection dogs.

## Abstract

Medical detection dogs search for diseases from remote samples (biodetection) and assist patients with chronic conditions (medical alert assistance). There is scarce information on how dogs’ decision-making tendencies relate to task performance. This study explored the relationships between medical detection dog demographics, responses in a behavioural test battery, ‘optimistic’ or ‘pessimistic’ decisions in a judgement bias task, and their performance in detection tasks. A sample of 58 trainee and trained medical detection dogs were studied in a Go/NoGo spatial judgement bias test. For trainee dogs (n = 39), training outcome (pass/fail) and trainer ratings of behavioural traits; yielding a composite score of ability in detection tasks, were used as markers of task performance. For trained biodetection dogs (n = 27), scent sensitivity and specificity scores derived during training and testing trials were used. Older dogs (p < 0.001), those showing higher ‘Confidence’ (p = 0.009), ‘Food orientation’ (p = 0.014) and ‘Playfulness’ (p = 0.005) in the test battery, and those who made more ‘optimistic’ decisions in the judgement bias task (p = 0.002), had higher detection task ability scores. For trained dogs, latency to approach ambiguous stimuli was positively correlated with scent specificity levels (n = 25, p = 0.021), suggesting that more ‘pessimistic’ dogs tended to be more specific. Our findings suggest relationships between behaviour in judgement bias tests and other learning and discrimination tasks, which may reflect underlying individual or personality differences in affective and/or cognitive processes that influence dogs’ style of searching and performance ability in medical detection tasks. Future research is needed to explore these associations further and investigate the value of judgement bias tasks in predicting later search performance in medical and other types of search dogs.