Biomarkers and behavior: What salivary biomarkers reveal about canine temperament

When a client walks into a clinic with a dog, veterinarians and their teams are already making rapid assessments before a history is fully taken or an examination begins. A dog may stiffen in the waiting room, try to hide, and avoid eye contact. Another dog may bound up to the counter, tail wagging, trying to greet every single member of the team.

These early observations provide information about a dog’s temperament, influencing how the dog will be approached, handled, and whether additional behavioral modifications, training, and/or pharmacological support may be needed.

Being able to rely on experience and knowledge to make these judgments is essential for the safety of the client, the patient, and the staff. But, although there are certain known indicators of fear or stress in dogs, those judgements can be somewhat subjective, shaped by experience, context, and interpretation.

Because of this inherent subjectivity, the details guiding these judgments can be murky. What is observed may be situational behavior, affected by the specific circumstances, or reflect the dog’s physical condition..

A recent study published in PLoS ONE examines this gap in understanding. Rather than looking at assessments of behavior in isolation, it linked scores from temperament tests with measurable biological markers and showed significant associations between observed behavior and physiology. This work provides objective evidence supporting the differences in temperament that clinicians are seeing and identifying every day—and, perhaps even more importantly, provides physiological validation for a widely used behavioral tool.

Temperament assessment is more than observation

Formal temperament testing is typically performed by behaviorists and trainers rather than by primary care teams (PCT). In a primary care veterinary setting, however, informal and continuous assessments of a dog’s temperament take place as a matter of course.

Every consultation involves some level of behavioral evaluation. Whether conscious or not, veterinarians and their teams assess a dog’s confidence, sociability, stress tolerance, and reactivity, often within seconds of meeting the patient. When these assessments are not standardized, it can make them challenging to rely upon.

Two clinicians may interpret the same dog differently depending on their experience and the context of the visit. A dog that presents as stressed during one visit may simply be expressing transient fear due to circumstances, but it may also indicate an underlying behavioral issue requiring further support, such as referral to a behaviorist. Recognizing the difference is critical—but difficult without the proper tools.

Understanding whether observable temperament traits can be measured would help shift reliance away from subjective interpretation toward a more objective, evidence-based approach. This is where the aforementioned study becomes particularly relevant.

How the study was designed

Researchers in South Korea, led by Youngwook Jung (2026), evaluated 24 dogs using a modified Wesen temperament test. The Wesen test is a structured temperament assessment most commonly used in European working dog and breeding programs. In this test, dogs are exposed to social and environmental challenges, which look at traits such as confidence, sociability, and responses to stress.

While the Wesen test may be unfamiliar to many U.S. veterinarians, the approach is similar to the standardized behavioral assessments used in shelters and by behaviorists, where a dog’s reactions to specific stimuli are used to build a picture of temperament. What was important about this study was not the test itself, but what accompanied it.

Saliva samples were collected before and after testing to measure cortisol and serotonin levels. Salivary sampling was used because it is non-invasive, reducing the risk of additional stress that could affect results. The concentrations of the two hormones were measured and compared to behavioral scores.

A five-fold difference in stress response

The results were significant. What stood out the most was the scale of the cortisol difference between dogs with high and low temperament scores. Dogs in the low-scoring group, tested after assessment, had cortisol levels approximately five times higher than those in the high-scoring group. Dogs that had been assessed as having more stable temperaments showed little change.

These differences were statistically significant (p < 0.01–0.005 across group comparisons), indicating that dogs identified as anxious, fearful, or poorly socialized were experiencing measurable physiological stress.

The serotonin signal

Although based on a smaller sample size, the serotonin findings were also significant. They confirmed that subjective assessments of dog temperaments are actually rooted in measurable physiology.

If a dog had a higher temperament score, it had approximately three and a half times more serotonin than low-scoring dogs. This is consistent with existing evidence that links low serotonin to anxiety, poor emotional regulation, and reduced behavioral inhibition, particularly where there is also aggression.

In the Wesen temperament assessment, serotonin levels showed a clear connection to how dogs performed on the Movement Stability subtest, which looks at how they handle physically unstable or unfamiliar environments, like uneven surfaces or raised platforms. Dogs that stayed relaxed and confident in those situations tended to have higher serotonin concentrations.

What this means in practice

Behavioral assessments like the Canine Behavioral Assessment & Research Questionnaire (C-BARQ), developed by the University of Pennsylvania, were designed to rate temperament traits using a standardized scoring system. This research is part of growing evidence that when a behaviorist uses a test like C-BARQ and finds that a dog shows poor confidence or low sociability, they are not just giving their opinion, they are reporting on what may be measurable differences in stress physiology and brain chemistry. That connection means that these reports can be interpreted by practitioners with greater confidence as indicators of actual underlying tendencies, rather than just tests of situational behavior.

Behavior that is often erroneously labeled as “bad” or ”difficult” is just as likely to be caused by underlying physiological issues. The range of cortisol differences observed in this study is clinically relevant. If a dog has a low temperament score, that does not mean that they are necessarily untrained. Instead, they may be exhibiting higher underlying physiological stress.

The kind of evidence emerging from studies looking at stress and physiological biomarkers in dogs can change how veterinarians discuss welfare and set expectations with clients.

For example, for dogs struggling with stress-related behaviors, training may be helpful, but alone, may not provide the kind of support the dog needs. The dog’s underlying physiological state needs to be addressed.

This research supports an integrated approach to treatment becoming more common in practice, where training is an essential element that, in some cases, benefits the dog most once Elevated cortisol and lower serotonin levels in dogs make them more prone to anxiety and reduced impulse control. This, in turn, makes training and behavioral modification more challenging.

Being able to show evidence that behavior has a physiological component can make conversations with clients around pharmacological intervention easier. In some cases, particularly where anxiety, fear, or aggression are present, medication may make learning and behavior modification easier and more likely to be effective.

Rather than being seen as a last resort, this research helps support recommendations that appropriate medication should form part of a comprehensive behavior treatment plan in many cases.

Understanding the limits of biomarkers

This work by Jung et al. is hugely relevant to the wider study of dog behavior, but it is also important to remember that this area of research is still developing.

A 2025 study on adult pet dogs and assistance dog puppies, published in Scientific Reports, found that salivary measures did not consistently reflect acute stress responses. So, salivary cortisol tests may not fully represent physiological responses accurately in all contexts and those responses are likely to be more complex than a single measure can identify.

It should also be noted that welfare research cautions against relying on cortisol as a stand-alone measure of stress in dogs, as it can be influenced by many different factors and does not necessarily distinguish between different types of arousal, meaning no single biomarker can fully measure an animal’s welfare state. But serotonin has been found to be a reliable biomarker in research on aggression in dogs.

In a 2025 study of working dogs in China, using the C-BARQ temperament assessment, significantly lower serotonin levels were seen in aggressive dogs, especially those displaying defensive aggression. This adds to the evidence that serotonin plays an important role in regulating behavior and supports the use of therapies that target it.

The general practitioner’s role

Behavior is increasingly recognized as having potential medical roots rather than simply being a training . Understanding the physiological basis of temperament strengthens the veterinarian’s role in case triage, treatment planning, and long-term management as part of a professional collaboration.

While PCTs may provide pet parents with behavior and training suggestions or resources, many don’t have the bandwidth to be a one-stop shop in that regard. In such instances, for behavior cases, it may be useful to identify who can fulfill which roles for the family in need.

For example, dog and puppy trainers, behaviorists, and primary veterinarians can support dog parents with issues related to early learning, communication, and management; general practitioners address underlying physiology; and veterinary behaviorists are uniquely positioned to support and guide both. A single approach is likely to be insufficient in complex dog behavior cases.

This study did not introduce a new concept, but it has provided physiological validation for a widely used behavioral tool. It has shown that temperament assessments can detect real differences in how dogs respond to stress, which fills a longstanding gap between subjective assessments, observations, and biology.

Supporting an integrated approach to dog behavior

A snapshot of a dog’s temperament using validated assessment tools reflects underlying biology. If they have poor behavioral scores, distinct patterns of stress hormone activation and lower serotonin levels can be measured, which helps explain why some dogs struggle despite appropriate training and environments. This knowledge is not only useful to the veterinary practitioner when looking at the dog’s general health and potential physical causes of stress, but can also help with educating dog parents and creating treatment plans.

This research helps us better understand that behavioral assessments reflect physiological states and not just learned responses. It clarifies that some dogs can be biologically predisposed to heightened anxiety; in those cases, training alone will not fully address those issues. A team approach, looking at behavior modification and medication as complementary to one another, may result in more successful interventions.

Photo credit: milorad kravic/E+ via Getty Images

Disclaimer: Trends™ content is meant to inform, educate, and inspire by providing an array of diverse viewpoints. Any content published should not be viewed as an official stance, position, or endorsement by the American Animal Hospital Association (AAHA) or its Board of Directors.