Report on UFAW Meeting July 2017

Measuring animal welfare and applying scientific advances – Why is it still so difficult?


The 2017 UFAW international symposium was held at Royal Holloway from 27th to 29th June. The general theme of the conference was that despite progress in animal welfare research, it is still difficult to agree on validated measures of animal welfare, particularly those that can be applied practically outside of research settings. The meeting was opened by Dr Robert Hubrecht, Chief Executive and Scientific Director of UFAW. As well as welcoming delegates to the meeting, Dr Hubrecht also gave a brief obituary of Professor Jaak Panksepp. Professor Panksepp was due to give a keynote speech at the conference, but sadly passed away earlier this year. He was remembered by several of the speakers for his work in affective neuroscience and for his research into ultrasonic rat vocalisations as an indicator of positive welfare.

The first keynote speech of the conference was given by Professor Mike Mendl. Professor Mendl began by asking how to define animal welfare, stating that there is increased use of feeling descriptors in papers on animal welfare, indicating a growing consensus that welfare is what an animal consciously experiences. This means that we need to know about animal emotions, but there is a lack of consensus as to what we mean by emotions. Two schools of thought exist. The first is that emotions are mental categories, defined by discrete neuro-behavioural systems. The second is that emotions are based on a small number of underlying systems over which many categories of feeling can be expressed. The latter is the basis for the core affect model. Several other speakers touched on the discrete and dimensional theories of emotion, some favouring one or the other, and some favouring a combined approach. Professor Mendl moved on to discuss three challenges in measuring animal welfare. First, what affective states exist in our study species? The challenge here is that there is not a consensus in the human literature of which theory of emotion (discrete or dimensional) is correct, so it is difficult to extrapolate affective states to animals. Second, what measures reflect affective states? Our options here are to translate measures from humans, to use intuitive inference or to apply theoretical concepts. There are issues with each approach, all of which increase the further away an animal is taxonomically, compared to humans. Third, can animals consciously experience affective states? This debate is still ongoing, but was expanded upon by both Dr James Yeates and Professor Victoria Braithwaite. Dr Yeates, from the RSPCA, questioned whether animals can consciously experience affective states. He stated that in ethics there a line often must be drawn as to whether an animal is sentient or not. This contrasts with the scientific stance where there is a large grey area in which sentience is up for question. Professor Braithwaite suggested ways in which affective stance could be used to demonstrate felt emotion and stressed the need for biologically and ecologically relevant stimuli in these tests.

The second keynote speech was given by Dr Tom Smulders, who suggested three key areas where neuroscience had a role to play in measuring animal welfare. First, methods such as fMRI and EEG can be used to measure acute affective states in the brain. These have been applied in several rodent studies. They are too complex to be used for routine monitoring but are of use in experiments where behavioural data is lacking. Second, new measures are becoming available to detect chronic affective states in the brain, including hippocampal neuronal regeneration. Hippocampal neuronal regeneration can be applied to welfare issues ranging from disease to husbandry and may indicate long term positive welfare. This technique could be used for routine monitoring whether animals are killed (e.g. meat production), but may be more useful in an experimental setting. Third, Dr Smulders suggested a more controversial concept, where neuroscience could aid in reducing an animal’s capacity to suffer, such as targeting emotional circuits to be bred out.

The third keynote speech was given by Professor Georgia Mason. She asked: are abnormal repetitive behaviours (ARB) an indicator of poor welfare? In discussing this question Professor Mason outlined several key questions which should be addressed when validating any animal welfare measure. First, the measure should change in animals exposed to aversive stimuli. This is true for ARB, which increases in singly housed pigs and macaques. Second, the measure should change when animals are exposed to stimuli that are ancestrally bad for fitness. ARB is increased in early weaned mink, mice, chimps, horses and dogs, satisfying this condition. Third, does the measure have good specificity? ARB does have good specificity as it increases with an increasing number of projects that individual chimps are used in. Forth, does the measure have good sensitivity? This is where ARB falls down as a measure of animal welfare as ARBs are prone to false negatives and poor welfare does not always lead to the development of an ARB. Professor Mason concluded by stating that ARB is better for measuring chronic states of welfare and that the absence of an ARB is not sufficient to infer good welfare.

Other general themes of the conference included the concept of lifelong positive welfare and behavioural reflections of poor welfare. On the topic of lifelong positive welfare, there was suggestion that some negative states are allowable as it would not be possible, or even desirable for an animal to have permanent positive welfare. Colline Poirier, in her talk on the use of hippocampal grey matter to determine long term affective states in macaques, suggested that negative emotions were acceptable, but negative moods were not. Alistair Lawrence discussed the difference between feeling good and doing good for animals, with few validated measures of either. Several talks discussed how behavioural measures could be used to determine poor welfare, including boredom, abnormal repetitive behaviours, sleep deprivation, facial expressions, motivational tests, play, vocalisations in pigs and chickens and fault bar development in chickens. There was a general feeling that it was difficult to validate behavioural measures of welfare and that each one on their own may not be enough to determine poor welfare. Sarah Wolfensohn introduced a new tool that she has developed to integrate several measures of welfare into an animal welfare assessment grid which can help to calculate cumulative suffering and provide a visual representation to aid the discussion of welfare issues. Professor Wolfensohn welcomes input into developing the animal welfare assessment grid, further details of which can be found at


Report written by Nicola Davidson, University of Liverpool