Cognitive Bias Task Research Articles

Cognitive bias in rats evoked by ultrasonic vocalizations suggests emotional contagion

Emotional contagion occurs when an individual acquires the emotional state of another via social cues, and is an important component of empathy. Empathic responses seen in rodents are often explained by emotional contagion. Rats emit 50kHz ultrasonic vocalizations (USVs) in positive contexts, and emit 22kHz USVs in negative contexts. We tested whether rats show positive or negative emotional contagion after hearing conspecific USVs via a cognitive bias task. We hypothesized that animals in positive emotional states would perceive an ambiguous cue as being good (optimistic bias) whereas animals in negative states would perceive the same cue as being bad (pessimistic bias). Rats were trained to respond differently to two sounds with distinct pitches, each of which signaled either a positive or a negative outcome. An ambiguous cue with a frequency falling between the two stimuli tested whether rats interpreted it as positive or negative. Results showed that rats responded to ambiguous cues as positive when they heard the 50kHz USV (positive vocalizations) and negative when they heard the 22kHz USV (negative vocalizations). This suggests that conspecific USVs can evoke emotional contagion, both for positive and negative emotions, to change the affective states in receivers.

A Screen-Peck Task for Investigating Cognitive Bias in Laying Hens.

Affect-induced cognitive judgement biases occur in both humans and animals. Animals in a more negative affective state tend to interpret ambiguous cues more negatively than animals in a more positive state and vice versa. Investigating animals’ responses to ambiguous cues can therefore be used as a proxy measure of affective state. We investigated laying hens’ responses to ambiguous stimuli using a novel cognitive bias task. In the ‘screen-peck’ task, hens were trained to peck a high/low saturation orange circle presented on a computer screen (positive cue–P) to obtain a mealworm reward, and to not peck when the oppositely saturated orange circle was presented (negative cue–N) to avoid a one second air puff. Ambiguous cues were orange circles of intermediate saturation between the P and N cue (near-positive–NP; middle–M; near-negative–NN), and were unrewarded. Cue pecking showed a clear generalisation curve from P through NP, M, NN to N suggesting that hens were able to associate colour saturation with reward or punishment, and could discriminate between stimuli that were more or less similar to learnt cues. Across six test sessions, there was no evidence for extinction of pecking responses to ambiguous cues. We manipulated affective state by changing temperature during testing to either ~20°C or ~29°C in a repeated measures cross-over design. Hens have been shown to prefer temperatures in the higher range and hence we assumed that exposure to the higher temperature would induce a relatively positive affective state. Hens tested under warmer conditions were significantly more likely to peck the M probe than those tested at cooler temperatures suggesting that increased temperature in the ranges tested here may have some positive effect on hens, inducing a positive cognitive bias.

Emotion Evaluation and Response Slowing in a Non-Human Primate: New Directions for Cognitive Bias Measures of Animal Emotion?

The cognitive bias model of animal welfare assessment is informed by studies with humans demonstrating that the interaction between emotion and cognition can be detected using laboratory tasks. A limitation of cognitive bias tasks is the amount of training required by animals prior to testing. A potential solution is to use biologically relevant stimuli that trigger innate emotional responses. Here; we develop a new method to assess emotion in rhesus macaques; informed by paradigms used with humans: emotional Stroop; visual cueing and; in particular; response slowing. In humans; performance on a simple cognitive task can become impaired when emotional distractor content is displayed. Importantly; responses become slower in anxious individuals in the presence of mild threat; a pattern not seen in non-anxious individuals; who are able to effectively process and disengage from the distractor. Here; we present a proof-of-concept study; demonstrating that rhesus macaques show slowing of responses in a simple touch-screen task when emotional content is introduced; but only when they had recently experienced a presumably stressful veterinary inspection. Our results indicate the presence of a subtle “cognitive freeze” response; the measurement of which may provide a means of identifying negative shifts in emotion in animals.

No Significant Evidence of Cognitive Biases for Emotional Stimuli in Children At-Risk of Developing Anxiety Disorders.

This paper explores whether the increased vulnerability of children of anxious parents to develop anxiety disorders may be partially explained by these children having increased cognitive biases towards threat compared with children of non-anxious parents. Parents completed questionnaires about their child’s anxiety symptoms. Children aged 5–9 (n = 85) participated in two cognitive bias tasks: 1) an emotion recognition task, and 2) an ambiguous situations questionnaire. For the emotion recognition task, there were no significant differences between at-risk children and children of non-anxious parents in their cognitive bias scores for reaction times or for accuracy in identifying angry or happy facial expressions. In addition, there were no significant differences between at-risk children and children of non-anxious parents in the number of threat interpretations made for the ambiguous situations questionnaire. It is possible that these cognitive biases only become present subsequent to the development of an anxiety disorder, or only in older at-risk children.

Emociones, toma de decisiones y consumo de alcohol en jóvenes universitarios

According to the somatic marker hypothesis, decision-making is adaptive when emotional processes are integrated into cognitive activity. The primacy of the affective system over cognitive control in young people is related to decision making based on instant gratification, regardless of eventual future harmful consequences. Binge alcohol consumption —very common in young people— is associated with an increased maladaptive decision making. The aim of this study was to evaluate decision making in young people of both sexes who are moderate and binge alcohol consumers, according to the type of emotional induction to which they have been exposed. Using a cross-over design, decision making was assessed in 79 participants aged 18 to 29, before and after the positive or negative emotional induction. The Iowa Gambling Task and the Cognitive Bias Task were administered in order to evaluate decision making. Emotions were induced using film clips. Heart rate was also recorded during the procedure. As a result, women negatively induced and men positively induced achieved a worse performance on the Iowa Gambling Task. Heart rate decreased compared to baseline during the experimental sessions. Binge drinkers had a worse performance compared to a moderate performance in the Cognitive Bias Task. It is concluded that emotional experience and the type of alcohol consumption influence decision making in university students.

Effects of repetitive transcranial magnetic stimulation on non-veridical decision making.

We test the emerging hypothesis that prefrontal cortical mechanisms involved in non-veridical decision making do not overlap with those of veridical decision making. Healthy female subjects performed an experimental task assessing free choice, agent-centered decision making (The Cognitive Bias Task) and a veridical control task related to visuospatial working memory (the Moving Spot Task). Transcranial magnetic stimulation (TMS) was applied to the left and right dorsolateral prefrontal cortex (DLPFC) using 1 Hz and 10 Hz (intermittent) rTMS and sham protocols. Both 1 Hz and 10 Hz stimulation of the DLPFC triggered a shift towards a more context-independent, internal representations driven non-veridical selection bias. A significantly reduced preference for choosing objects based on similarity was detected, following both 1 Hz and 10 Hz treatment of the right as well as 1 Hz rTMS of the left DLPFC. 1 Hz rTMS treatment of the right DLPFC also triggered a significant improvement in visuospatial working memory performance on the veridical task. The effects induced by prefrontal TMS mimicked those of posterior lesions, suggesting that prefrontal stimulation influenced neuronal activity in remote cortical regions interconnected with the stimulation site via longitudinal fasciculi.

Clinical and genetic correlates of decision making in anorexia nervosa

ABSTRACTIntroduction: Decision-making (DM) abilities have been found to be impaired in anorexia nervosa (AN), but few data are available about the characteristics and correlates of this cognitive function. The aim of the present study was to provide data on DM functioning in AN using both veridical and adaptive paradigms. While in veridical DM tasks, the individual’s ability to predict a true/false response is measured, adaptive DM is the ability to consider both internal and external demands in order to make a good choice, in the absence of a single true “correct” answer. Method: The participants were 189 women, of whom 91 were eating-disordered patients with a lifetime diagnosis of anorexia nervosa, and 98 were healthy women. All the participants underwent clinical, neuropsychological, and genetic assessment. The cognitive evaluation included a set of neuropsychological tasks and two decision-making tests: The Iowa Gambling Task and the Cognitive Bias Task. Results: Anorexia nervosa patients showed significantly poorer performances on both decision-making tasks than healthy women. The Cognitive Bias Task revealed that anorexia nervosa patients employed significantly more context-independent decision-making strategies, which were independent from diagnostic subtype, handedness, education, and psychopathology. In the whole sample (patients and controls), Cognitive Bias Task performance was independently predicted by lifetime anorexia nervosa diagnosis, body mass index at assessment, and 5-HTTLPR genotype. Conclusions: Patients displayed poor decision-making functioning in both veridical and adaptive situations. The difficulties detected in anorexia nervosa individuals may affect not only the ability to consider the future outcomes of their actions (leading to “myopia for the future”), but also the capacity to update and review one’s own mindset according to new environmental stimuli.

The long-term impact of infant rearing background on the affective state of adult common marmosets (Callithrix jacchus)

Early life environment, including temporary family separation, can have a major influence on affective state. Using a battery of tests, the current study compared the performance of adult common marmosets (Callithrix jacchus), reared as infants under 3 different conditions: family-reared twins, family-reared animals from triplet litters where only 2 remain (2stays) and supplementary fed triplets. No significant differences were found in latency to approach and obtain food from a human or a novel object between rearing conditions, suggesting no effect on neophobia. There were no differences in cognitive bias task acquisition time, or proportion of responses to each ambiguous probe. Very minor differences were found in response to the probes, with only supplementary fed marmosets making fewer responses to the middle probe, compared to the probe nearest the rewarded stimuli. Similarly, in a test for anhedonia, no difference was found between rearing conditions in consumption of milkshake at different concentrations. There was just one very small difference in reward motivation, with only supplementary fed triplets demonstrating a lack of preference for milkshake over water at the lowest concentration. This consistent pattern of results suggests that the supplementary feeding of large litters of marmosets at this facility did not have a major effect on welfare, and is unlikely to influence performance in reward-related scientific tasks. Therefore, while family separation is not recommended, this particular practice should be used if it is necessary, such as to reduce infant mortality. Regular positive interactions with humans are also encouraged, to reduce fear and improve welfare of marmosets kept in captivity.

Rats became positive or negative states when listening to specific vocalizations

Emotional contagion is the process in which one became the same emotional states of the others via a behavioral signal, and many species use acoustic cues as this signal. For example, in human beings, emotional contagion occurs when they get emotional vocalizations of others, like laughter or crying. Likewise, rats emit 50 kHz or 22 kHz specific ultrasonic vocalizations (USVs) associated with positive or negative context. In order to measure whether these acoustic cues have the valence of attractiveness or aversiveness, we utilized the cognitive bias task in rats. Rats were trained to respond differently for two neutral stimuli. One resulted in a positive outcome while the other produced negative outcome. Then the intermediate stimulus between the two was used to test whether that was interpreted as being positive or negative. After the 50 kHz USVs stimulation, rat interpreted the neutral stimuli as being positive, while after the 22 kHz USV stimulation, the same stimuli were regarded as negative. Results suggest that USVs can bring about positive or negative emotional states of listeners, and our findings indicated that specific USVs can work as acoustic emotional signal in the conspecific. [Work supported by JSPS KAKENHI #23118003.]

Opposite Effects of Early-Life Competition and Developmental Telomere Attrition on Cognitive Biases in Juvenile European Starlings.

Moods are enduring affective states that we hypothesise should be affected by an individual’s developmental experience and its current somatic state. We tested whether early-life adversity, induced by manipulating brood size, subsequently altered juvenile European starlings’ (Sturnus vulgaris) decisions in a judgment bias task designed to provide a cognitive measure of mood. We predicted that starlings from larger broods, specifically those that had experienced more nest competitors larger than themselves would exhibit reduced expectation of reward, indicative of a ‘pessimistic’, depression-like mood. We used a go/no-go task, in which 30 starlings were trained to probe a grey card disc associated with a palatable mealworm hidden underneath and avoid a different shade of grey card disc associated with a noxious quinine-injected mealworm hidden underneath. Birds’ response latencies to the trained stimuli and also to novel, ambiguous stimuli intermediate between these were subsequently tested. Birds that had experienced greater competition in the nest were faster to probe trained stimuli, and it was therefore necessary to control statistically for this difference in subsequent analyses of the birds’ responses to the ambiguous stimuli. As predicted, birds with more, larger nest competitors showed relatively longer latencies to probe ambiguous stimuli, suggesting reduced expectation of reward and a ‘pessimistic’, depression-like mood. However, birds with greater developmental telomere attrition—a measure of cellular aging associated with increased morbidity and reduced life-expectancy that we argue could be used as a measure of somatic state—showed shorter latencies to probe ambiguous stimuli. This would usually be interpreted as evidence for a more positive or ‘optimistic’ affective state. Thus, increased competition in the nest and poor current somatic state appear to have opposite effects on cognitive biases. Our results lead us to question whether increased expectation of reward when presented with ambiguous stimuli always indicates a more positive affective state. We discuss the possibility that birds in poor current somatic state may adopt a ’hungry’ cognitive phenotype that could drive behaviour commonly interpreted as ‘optimism’ in food-rewarded cognitive bias tasks.

Dogs with separation-related problems show a “less pessimistic” cognitive bias during treatment with fluoxetine (Reconcile™) and a behaviour modification plan

BackgroundCanine separation-related problems (SRP) (also described as “separation anxiety” or “separation distress”) are among the most common behavioural complaints of dog owners. Treatment with psychoactive medication in parallel with a behaviour modification plan is well documented in the literature, but it is unknown if this is associated with an improvement in underlying affective state (emotion and mood) or simply an inhibition of the behaviour. Cognitive judgement bias tasks have been proposed as a method for assessing underlying affective state and so we used this approach to identify if any change in clinical signs during treatment was associated with a consistent change in cognitive bias (affective state).Five dogs showing signs of SRP (vocalising – e.g. barking, howling-, destruction of property, and toileting – urination or defecation- when alone) were treated with fluoxetine chewable tablets (Reconcile™) and set on a standard behaviour modification plan for two months. Questionnaires and interviews of the owners were used to monitor the clinical progress of the dogs. Subjects were also evaluated using a spatial cognitive bias test to infer changes in underlying affect prior to, and during, treatment. Concurrently, seven other dogs without signs of SRP were tested in the same way to act as controls. Furthermore, possible correlations between cognitive bias and clinical measures were also assessed for dogs with SRP.ResultsPrior to treatment, the dogs with SRP responded to ambiguous positions in the cognitive bias test negatively (i.e. with slower running speeds) compared to control dogs (p < 0.05). On weeks 2 and 6 of treatment, SRP dogs displayed similar responses in the cognitive bias test to control dogs, consistent with the possible normalization of affect during treatment, with this effect more pronounced at week 6 (p > 0.05). Questionnaire based clinical measures were significantly correlated among themselves and with performance in the cognitive bias test.ConclusionThese results demonstrate for the first time that the clinical treatment of a negative affective state and associated behaviours in a non-human species can produce a shift in cognitive bias. These findings demonstrate how the outcome of an intervention on a clinical problem can be evaluated to determine not only that the subject’s behaviour has improved, but also its psychological state (welfare)Electronic supplementary materialThe online version of this article (doi:10.1186/s12917-015-0373-1) contains supplementary material, which is available to authorized users.

Individual differences in cognitive biases: Evidence against one-factor theory of rationality

In this paper we seek to gain an improved understanding of the structure of cognitive biases and their relationship with measures of intelligence and relevant non-cognitive constructs. We report on the outcomes of a study based on a heterogeneous set of seven cognitive biases — anchoring effect, belief bias, overconfidence bias, hindsight bias, base rate neglect, outcome bias and sunk cost effect. New scales for the assessment of these biases were administered to 243 undergraduate students along with measures of fluid (Gf) and crystallized (Gc) intelligence, a Cognitive Reflection Test (CRT), Openness/Intellect (O/I) scale and Need for Cognition (NFC) scale. The expected experimental results were confirmed — i.e., each normatively irrelevant variable significantly influenced participants' responses. Also, with the exception of hindsight bias, all cognitive biases showed satisfactory reliability estimates (αs>.70). However, correlations among the cognitive bias measures were low (rs<.20). Although exploratory factor analysis produced two factors, their robustness was doubtful. Cognitive bias measures were also relatively independent (rs<.25) from the Gf, Gc, CRT, O/I and NFC and they define separate latent factors. This pattern of results suggests that a major part of the reliable variance of cognitive bias tasks is unique, and implies that a one-factor model of rational behavior is not plausible.

Stress is a bad advisor. Stress primes poor decision making in deluded psychotic patients.

Stress is implicated in the onset of psychosis but the complex links between stress and psychotic breakdown are yet poorly understood. For the present study, we examined whether two prominent cognitive biases in psychosis, jumping to conclusions and distorted attribution, in conjunction with neuropsychological deficits play a role in this process. Thirty participants with schizophrenia and acute delusional symptoms were compared with 29 healthy controls across three conditions involving a noise stressor, a social stressor or no stressor. Under each condition participants had to perform parallel versions of cognitive bias tasks and neuropsychological tests including a probabilistic reasoning task (jumping to conclusions), the revised Internal, Personal and Situational Attributions Questionnaire (IPSAQ-R; attributional style), and the Corsi block-tapping task (nonverbal memory). Stress, particularly noise, aggravated performance differences of patients relative to controls on memory. Participants with psychosis demonstrated an escalated jumping to conclusion bias under stress. At a medium effect size, patients made more monocausal attributions, which increased under social stress. The present study is partially in line with prior studies. It suggests that stress negatively affects cognition in psychosis more than in controls, which is presumably insufficiently realized by patients and thus not held in check by greater response hesitance. Raising patients' awareness about these response tendencies and encouraging them to be more cautious in their judgments under conditions of increased stress may prove beneficial for improving positive symptoms.

Cognitive manifestations of drinking–smoking associations: Preliminary findings with a cross-primed Stroop task

BackgroundDespite tremendous growth in research examining the role of cognitive bias in addictive behaviors, scant consideration has been paid to the close association between smoking and drinking behavior. This study sought to determine whether an association between smoking and drinking could be observed at an implicit level using a novel cognitive bias task, as well as characterize the relationship between performance on this task and clinically relevant variables (i.e., heaviness of use/dependence). MethodsIndividuals (N=51) with a range of smoking and drinking patterns completed a modified Stroop task in which participants identified the color of drinking, smoking and neutral words that were each preceded by drinking, smoking or neutral picture primes. Participants also provided information regarding the heaviness of their smoking and drinking behavior and completed self-report measures of alcohol and nicotine dependence. ResultsResponse times to smoking and drinking words were significantly slowed following the presentation of either smoking or drinking picture primes. This effect did not differ across subgroups. However, the strength of the coupling between smoking and drinking prime effects was greater among heavier drinkers, who also exhibited a concordant looser coupling of the effects of smoking and drinking primes on smoking words. ConclusionsAssociations between smoking and drinking can be observed at an implicit level and may be strongest for heavier drinkers.

Developing an optimism index using results from a cognitive bias task

Developing an optimism index using results from a cognitive bias task

Housing conditions affect rat responses to two types of ambiguity in a reward–reward discrimination cognitive bias task

Decision-making under ambiguity in cognitive bias tasks is a promising new indicator of affective valence in animals. Rat studies support the hypothesis that animals in a negative affective state evaluate ambiguous cues negatively. Prior automated operant go/go judgement bias tasks have involved training rats that an auditory cue of one frequency predicts a Reward and a cue of a different frequency predicts a Punisher (RP task), and then measuring whether ambiguous cues of intermediate frequency are judged as predicting reward (‘optimism’) or punishment (‘pessimism’). We investigated whether an automated Reward–Reward (RR) task yielded similar results to, and was faster to train than, RP tasks. We also introduced a new ambiguity test (simultaneous presentation of the two training cues) alongside the standard single ambiguous cue test. Half of the rats experienced an unpredictable housing treatment (UHT) designed to induce a negative state. Control rats were relatively ‘pessimistic’, whilst UHT rats were quicker, but no less accurate, in their responses in the RR test, and showed less anxiety-like behaviour in independent tests. A possible reason for these findings is that rats adapted to and were stimulated by UHT, whilst control rats in a predictable environment were more sensitive to novelty and change. Responses in the new ambiguity test correlated positively with those in single ambiguous cue tests, and may provide a measure of attention bias. The RR task was quicker to train than previous automated RP tasks. Together, they could be used to disentangle how reward and punishment processes underpin affect-induced cognitive biases.

Does stocking density modify affective state in pigs as assessed by cognitive bias, behavioural and physiological parameters?

Recent studies suggest that emotional state can affect cognitive abilities of humans and non-human animals, determining biases in information processing. Negative mental states, such as anxiety or depression, induce pessimistic judgments of ambiguous stimuli. These assumptions may be used to derive indicators of emotional state in captive animals, providing a novel approach to the assessment of animal welfare. This study used a spatial judgement task, in which farmed pigs were trained to expect food inside a bowl in one location and not in another, to determine whether pigs housed in ways that might be expected to result in relatively positive or negative emotional states respond differently to ambiguous stimuli of intermediate spatial locations. Forty growing pigs were housed in groups of 10 at different density for 8 weeks prior to the start of the test. After training, the pigs successfully discriminated between the rewarded and the unrewarded locations as assessed by increased latency to arrive at the unrewarded location, with no rearing treatment difference. Then, pigs were tested on 3 days in which three ambiguous locations, intermediate between the known rewarded and the unrewarded sites, were introduced and latency recorded. In order to compare the novel cognitive bias task with other welfare indicators, during the 8 weeks of the study four behavioural observations, two measurements of skin lesions, two salivary samples for cortisol and α-amylase, and six individual weights to assess growth were collected.Considering the mean of the three test days, there was no difference between the treatments in the pigs’ judgement of the three ambiguous locations. However, the latency trend during the testing days led to difference between treatments on the third day (P=0.026). Pigs housed a higher density seemed to learn faster that the ambiguous stimulus near the unrewarded location was also not reinforced by a reward and showed a higher latency to approach on day 3 (44 vs 15.6s). These animals were also observed to have a higher frequency of sitting posture (P=0.01), and more total skin lesions (P=0.035) due to aggressiveness, in particular at the ear location (P=0.009), but did not differ in other physiological parameters.Although the results showed no immediate effect of stocking density on cognitive bias, differences in latencies to reach the bowl over the three testing days suggest a different learning process between treatments.

Pain and Pessimism: Dairy Calves Exhibit Negative Judgement Bias following Hot-Iron Disbudding

Pain is defined as an unpleasant sensory and emotional experience associated with actual or potential tissue damage, but emotional states are difficult to directly assess in animals. Researchers have assessed pain using behavioural and physiological measures, but these approaches are limited to understanding the arousal rather than valence of the emotional experience. Cognitive bias tasks show that depressed humans judge ambiguous events negatively and this technique has been applied to assess emotional states in animals. However, limited research has examined how pain states affect cognitive processes in animals. Here we present the first evidence of cognitive bias in response to pain in any non-human species. In two experiments, dairy calves (n = 17) were trained to respond differentially to red and white video screens and then tested with unreinforced ambiguous colours in two or three test sessions before and two sessions after the routine practice of hot-iron disbudding. After disbudding calves were more likely to judge ambiguous colours as negative. This ‘pessimistic’ bias indicates that post-operative pain following hot-iron disbudding results in a negative change in emotional state.

Metacognitive training for patients with schizophrenia: Preliminary evidence for a targeted, single-module programme

Metacognitive training is an eight-module, group-based treatment programme for people with schizophrenia that targets the cognitive biases (i.e. problematic thinking styles) thought to contribute to the genesis and maintenance of delusions. The present article is an investigation into the efficacy of a shorter, more targeted, single-module metacognitive training programme, administered individually, which focuses specifically on improving cognitive biases that are thought to be driven by a 'hypersalience of evidence-hypothesis matches' mechanism (e.g. jumping to conclusions, belief inflexibility, reasoning heuristics, illusions of control). It was hypothesised that a more targeted metacognitive training module could still improve performance on these bias tasks and reduce delusional ideation, while improving insight and quality of life. A sample of 28 patients diagnosed with schizophrenia and mild delusions either participated in the hour-long, single-session, targeted metacognitive training programme (n = 14), or continued treatment as usual (n = 14). All patients were assessed using clinical measures gauging overall positive symptomology, delusional ideation, quality of life and insight, and completed two cognitive bias tasks designed to elucidate the representativeness and illusion of control biases. After a 2-week, post-treatment interval, targeted metacognitive training patients exhibited significant decreases in delusional severity and conviction, significantly improved clinical insight, and significant improvements on the cognitive bias tasks, relative to the treatment-as-usual controls. Performance improvements on the cognitive bias tasks significantly correlated with the observed reductions in overall positive symptomology. Patients also evaluated the training positively. Although interpretations of these results are limited due to the lack of an optimally designed, randomised controlled trial and a small sample size, the results are promising and warrant further investigation into targeted versions of the metacognitive training programme.

Validation of a novel cognitive bias task based on difference in quantity of reinforcement for assessing environmental enrichment

Cognitive bias tasks purport to assess affective states via responses to ambiguous stimuli. We hypothesized that a novel cognitive bias task based on positive reinforcement using quantity differences would detect changes in affect in captive grizzly bears (Ursus arctos horribilis). We trained bears (n = 8) to respond differently (nose or paw touch) to two stimuli (light or dark gray cue cards), with responses counterbalanced across bears. The two cues signaled a small or large food reward, respectively. Responses to ambiguous probe stimuli (i.e., shades of gray) intermediate to the trained stimuli were classified as either 'optimistic,' appropriate for the larger reward, or 'pessimistic,' appropriate for the smaller reward. In Experiment 1, we explored the contrast in reward size necessary to detect a change in response across probe stimuli (large reward, 3 or 6 apple slices: small reward, 1 slice). We observed a change in response across probe stimuli, with no difference in response between reward-value groups, indicating that a contrast of 3:1 apple slices was sufficient to affect responses. In Experiment 2, we investigated cognitive bias after 2.1 h of exposure to enrichment items varying in attractiveness. Results were unaffected by enrichment type or time spent interacting with enrichments, indicating that the task failed to demonstrate criterion validity for comparing mood following exposure to different enrichment items. However, greater time spent pacing prior to testing was associated with 'optimistic' judgments. The data provide some support for use of cognitive bias tasks based on quantity differences in animal welfare assessments involving captive wildlife.