Three well-established rodent models of somatic, visceral and neuropathic pain were used to test the hypothesis that a stress and anxiety evoked behaviour, namely ultrasound vocalisation, correlates with other well-characterised indices of pain behaviour, such as limb withdrawal and stereotypical behaviour. Persistent pain presents a significant clinical problem for which there remains relatively ineffective clinical management and animal models of pain are commonly employed to investigate the underlying pathophysiology and for pre-clinical evaluation of novel therapies. At present, the assessment of such animal models largely relies on the observation of simple reflex responses which may not entirely represent the full range of rodent pain behaviour. Therefore, additional integrated behavioural indices for the quantification of pain could improve the veracity of animal models. In stressful or harmful situations, it is thought that rodents produce ultrasound vocalisations to communicate within the social group. In this study, the number of ultrasound vocalisations (22 kHz) was measured during both evoked and ongoing pain. Ultrasound vocalisation was not associated with other pain behaviour in any of the inflammatory, visceral or neuropathic pain models examined and is therefore not a useful integrated correlate of pain behaviour.