Table 1 Example behavioral indicators of animal welfare, their uses, and limitations

Behavioural observations

Certain behaviours are more likely to occur depending on whether an animal is in a positive or negative state [5]

Stereotypies (e.g., bar chewing in pigs and tongue rolling in cattle) [8, 10, 11]

Aggressive behaviours (e.g., injurious feather pecking in poultry and tail biting in pigs) [12,13,14,15]

Rebound activity, which is also used to indicate that an animal is motivated to perform that behaviour [16]

Behavioural diversity, social behaviours (e.g., play) and maternal behaviours [17,18,19]

Focus generally on the presence of behaviours associated with negative mental states, such as abnormal or redirected behaviours

Unreliable and non-specific (i.e., multiple mental states can produce the same behaviour)

Influenced by level of arousal, which may be indicative of both positive and negative mental states

Subjective and requires training/understanding of species-specific behaviours

Influenced by past and learned experiences

Fear tests

Measurement of the ability of an animal to cope in a challenging situation

Novel object, restraint, or isolation tests

Used to assess fear in response and thus level of anxiety [20]

Focused solely on negative mental state

The fear response is complex and varies at an individual level due to genetics and previous experiences [21]

Can be contradictory in that animals may either respond actively (fight or flight) or passively (freezing) [21]

Cognitive bias tests (judgement, attention, and memory)

The emotional state of an animal will impact the way they assess risks (judgement), choose what to focus on (attention), or remember (memory) an event

Animals in a negative mental state are thought to show negative bias in judgement, pay more attention to threats, and show impaired memory

Higher frequencies of anticipatory behaviour in animals associated with more pessimistic judgements in cognitive bias tests [22]

Require animals to be trained to respond to cues associated with a rewarding or unrewarding event; influenced by previous experience; unreliable; and animals in poor welfare may show negative judgement bias, but the converse does not necessarily follow [23,24,25,26,27,28]

Preference and motivation tests

Measurement of what and how much animals ‘want’ an environment or resource

Preference tests require animals to ‘work’ or pay a ‘cost’ to perform a behaviour or use a resource

Motivation tests attempt to quantify the importance an animal places on a preference

Dependent on individual variation, environmental context and learned experience [29]

Unreliable and non-specific (i.e., behaviour does not necessarily correlate to the actual experience of an animal)

Qualitative behavioural assessments

Human assessment of an animal as a whole and how it interacts with its environment

Uses behavioural descriptors on a continuum ranging from low (e.g., calm, relaxed) and high (e.g., active, restless) levels of arousal

Specifically validated in several species, including farm animals [30,31,32,33,34,35,36,37,38,39,40]

Subjective, context sensitive and relies on the ability of the human assessor [29]

Influenced by past and learned experiences

High arousal can be correlated with both positive and negative mental states, as can low arousal (e.g., indicating learned helplessness, freezing response, or calm)

Facial expression

Facial expressions are reflective of emotional states

Facial expressions [41, 42]

Eye white and movements, ear and body postures, movement patterns [43, 44]

Mainly been used in relation to negative mental states, for example use of facial expressions to measure pain

Difficult to correlate specifically with the mental state of an animal

Require species-specific data for accurate analysis due to species and breed variations

Vocalizations

Vocalizations are expressed according to the experience lived by the animals

The use of vocalizations in animals as a form of welfare assessment have successfully been demonstrated in many farm animals including pigs, cattle, and horses [45,46,47]

Vocalization research in this field is measured under simulated situations where the animals are put in a setting that trigger vocalizations, resulting in ‘artificial’ vocalizations