The welfare of animals is the top priority of scientists working with them. Mice are the most commonly used species in scientific research, with almost 3 million used last year alone. To give some perspective, over ten times fewer rats were included in tests over the same time period. Unsurprisingly then, a considerable amount of research has focussed on how to improve conditions for mice used in research. At present, Home Office guidelines require that mice are checked daily by trained animal care teams. Those teams assess whether the mice look healthy and happy and will intervene if there are any signs of a problem. Unfortunately, mice are a prey species and, because of this, they try to hide visible signs of injury/illness. This means that simply looking at the mice is not the best way to assess their welfare. However, mice are known to vocalise about issues such as illness and pain before they show visible signs. Most mouse vocalisations go unheard by humans as they are ultrasonic. I will use specialised equipment to record vocalisations and record what the mouse is doing while vocalising in the cage the mouse is housed in. I will then match vocalisations with particular behaviours in order to find out what the mice are 'saying'. To ensure that the research is beneficial for welfare, I will focus on identifying vocalisations associated with potential welfare concerns, such as pain or aggression. Thousands of mice are lost each year because of fights between animals that are housed together in the same cage, and thousands more hide signs of illness, preventing treatment until a later stage. This research aims to decrease the distress felt by animals and enable earlier interventions to maximise welfare. As part of the research I will develop a large repository of information about which vocalisations are associated with a given behaviour. This will be used by researchers to examine what mouse vocalisations and/or behaviours mean for welfare and will enable them to intervene as necessary. I hope that this research will eventually lead to the development of a welfare tool that automatically translates mouse vocalisations into information about behaviours and presents this information to animal care teams and researchers in a way that means they can intervene quickly to maximise welfare.