How Emotional State Shapes Animal Decision-Making — and What It Reveals About Welfare
One of the most significant advances in animal welfare science over the past two decades is the development of cognitive bias testing as a tool for assessing emotional states in animals. Borrowed from human psychology — where pessimistic cognitive biases are associated with depression and anxiety — this methodology has been applied across livestock species to answer a question that long seemed unanswerable: is this animal in a positive or negative emotional state overall?
Cognitive bias research in humans established that depression and anxiety are associated with systematic distortions in information processing — depressed individuals tend to interpret ambiguous situations negatively, remember negative events more readily, and expect negative outcomes. This "pessimistic cognitive bias" is now a diagnostic marker for affective disorders in clinical psychology.
Researchers led by Mike Mendl, Elizabeth Paul, and colleagues at the University of Bristol pioneered the application of cognitive bias testing to non-human animals in the 2000s. Their foundational 2004 paper in Current Biology demonstrated that rats in negative affective states showed more pessimistic judgments in ambiguous spatial tasks — establishing that the cognitive bias framework is valid across mammalian species.
The standard procedure involves training an animal to associate two distinct stimuli (a "positive cue" and a "negative cue") with different outcomes:
Once the animal has learned these associations, it is presented with "ambiguous cues" — stimuli intermediate between the two trained cues. The animal's response speed and decision to approach or avoid the ambiguous cue reveals its cognitive bias: animals that respond to ambiguous cues as if they were the positive cue are classified as "optimistic"; those that treat ambiguous cues as negative are "pessimistic."
Beyond whether animals approach or avoid, the latency (speed) of their response provides additional information. Animals that quickly approach an ambiguous cue are showing high expectation of positive outcomes; slow or reluctant approaches suggest intermediate states. This continuous measure allows more nuanced assessment than binary approach/avoid coding.
Pigs have been extensively studied with cognitive bias paradigms. Key findings:
Sheep cognitive bias research has demonstrated that social isolation — a significant stressor given sheep's strong social nature — induces pessimistic bias, while social reunification restores optimism. This work has implications for practices like individual housing of sheep during veterinary procedures or post-transport isolation.
Studies in cattle have shown that:
Cognitive bias research in poultry is more challenging due to differences in cognitive architecture, but studies have shown that:
More recent work has extended cognitive bias testing to fish — particularly zebrafish and rainbow trout. This research has important implications for aquaculture welfare, as it provides evidence that fish experience emotional states that influence cognition, supporting the case for their sentience.
Cognitive bias tests have been used to compare welfare across housing systems in a way that transcends observer bias. When enriched pig pens produce more optimistic pigs than barren pens in a controlled trial, the finding is methodologically robust and difficult to dismiss. This has provided welfare scientists with powerful evidence for policies like the EU's enrichment requirements for pig housing.
Cognitive bias tests have proven particularly valuable in demonstrating the welfare impact of pain. When studies show that castration or dehorning without analgesia produces persistent pessimistic bias that is reversed by adequate pain relief, the case for mandatory pain management becomes scientifically concrete rather than philosophically argued.
The effect of stockperson behavior on cognitive bias — with gentle, positive handling producing more optimistic animals — has direct practical implications. Training programs for farm workers that emphasize positive human-animal interactions can be evaluated not just by fear scores but by measurable cognitive state improvements.
Welfare science has traditionally focused on identifying and reducing negative states. Cognitive bias research has enabled a complementary focus: measuring positive emotional states and optimism. The "optimistic shift" produced by play behavior, environmental enrichment, and social bonding can now be quantified, supporting the positive welfare framework (not just absence of suffering, but presence of positive experiences).
| Intervention | Species Studied | Cognitive Bias Effect |
|---|---|---|
| Environmental enrichment | Pigs, chickens | Shift toward optimism |
| Positive handling/HAR | Pigs, cattle | Shift toward optimism |
| Social housing | Sheep, cattle | Shift toward optimism |
| Chronic pain (lameness) | Cattle | Shift toward pessimism |
| Unpredictable housing | Pigs | Shift toward pessimism |
| Isolation | Sheep, pigs | Shift toward pessimism |
| Painful procedures without analgesia | Cattle, pigs | Shift toward pessimism |
| Play induction | Rats, pigs | Shift toward optimism |
Cognitive bias testing is not without limitations:
Current cognitive bias protocols are largely laboratory-based. Active research aims to develop automated, sensor-based systems that can assess cognitive indicators of emotional state in commercial farm settings. Eye-tracking, movement analysis, and machine learning approaches are being explored as scalable alternatives to traditional judgment bias tasks.
Cognitive bias research represents a landmark advance in our ability to understand and measure animal emotional life. Its application to livestock species has provided compelling scientific evidence for welfare improvements in housing, handling, pain management, and human-animal relationships. As methodologies mature and become more practical for on-farm use, cognitive bias assessment has the potential to transform how we evaluate and certify animal welfare — moving from proxy indicators to direct measures of what animals actually experience.