When your dog's 'bad' behavior is actually their body chemistry

Doges Editorial · 2026-05-17 · 5 min read

A new study from South Korea measured cortisol and serotonin in 24 dogs before and after a structured temperament assessment. The anxious dogs had five times the cortisol of the confident ones — and it changes how we should think about 'difficult' dogs.

Two dogs enter a veterinary waiting room. The first stands rigid by the door, pressed flat against the wall, eyes wide and flicking, refusing to move. The second bounces toward the reception desk like it is being personally welcomed back after a long absence. Both dogs are about the same age, same size. Neither has any physical complaint. They are just — different.

Every vet, every trainer, every regular at a dog park has seen this difference. The question that a new study, published in PLoS ONE and analyzed this week by the American Animal Hospital Association, has now begun to answer is: what is actually different, chemically, between those two dogs?

The experiment in South Korea

Researchers led by Youngwook Jung at a South Korean institution evaluated 24 dogs using a modified Wesen temperament test — a structured European assessment that exposes dogs to a series of social and environmental challenges. The test looks at things like confidence, sociability, and how a dog responds to stress. It is more common in working dog and breeding programs in Europe than in the average vet clinic, but the underlying approach is similar to the standardized shelter assessments used in the US.

Before and after the test, the researchers collected saliva samples from each dog. They chose saliva for a specific reason: blood draws would add another stressor, potentially contaminating whatever they were trying to measure. Salivary sampling is non-invasive. The dogs gave it willingly.

From those samples, the researchers measured two things: cortisol, the body's primary stress hormone, and serotonin, which plays a central role in mood, impulse control, and emotional regulation.

Five times the cortisol

The scale of the difference surprised even the researchers. Dogs that scored poorly on the temperament test — the anxious ones, the fearful ones, the dogs that shrank from novel stimuli or reacted with alarm to things the high-scoring dogs barely noticed — had cortisol levels roughly five times higher after the assessment than the stable, confident dogs did.

Five times. Not "somewhat elevated." Five times. The difference was statistically significant across group comparisons (p < 0.01 to 0.005). This was not noise in the data. The dogs with the worst scores on the behavioral assessment were, simultaneously, the dogs whose bodies were flooding with stress hormones.

Behavior that is often erroneously labeled as 'bad' or 'difficult' is just as likely to be caused by underlying physiological issues.

— AAHA Trends, analyzing the Jung et al. study (2026)

What serotonin adds to the picture

The serotonin findings built on the cortisol story. The high-scoring dogs had approximately three and a half times more serotonin than the low-scoring dogs. This tracked with what behavioral research has shown elsewhere: low serotonin is associated with anxiety, reduced emotional regulation, and heightened reactivity — particularly in dogs that display defensive aggression.

One of the more telling connections emerged from the Movement Stability subtest — a section of the Wesen assessment that looks at how dogs handle physically unfamiliar or unstable environments: uneven surfaces, raised platforms, footing that shifts underfoot. Dogs that stayed calm on the platform had higher serotonin levels. Dogs that stiffened or froze had lower ones. Their bodies were already having a different experience of the world.

The dog that won't walk past certain corners

If you have walked a high-anxiety dog, you know what this looks like in daily life. The dog that plants its feet at a specific corner and will not move. The one that shakes for forty minutes after another dog walked past on the other side of the street. The one that is fine on Tuesday mornings and completely undone by Friday afternoons, for no visible reason.

The instinct has often been to frame this as a training problem: the dog lacks exposure, or confidence, or has been inadvertently reinforced for avoidance. And training — particularly reward-based, patient, low-pressure training — can absolutely help. But this study adds an important qualifier. If the dog's baseline cortisol is chronically elevated and its serotonin is low, training is working against a physiological headwind. The body keeps producing the experience of threat whether or not the environment contains one.

The case for a different conversation

What this study offers, in practical terms, is a better vocabulary for discussing these dogs with veterinarians. If a dog consistently scores poorly on behavioral assessments, and if salivary biomarkers reveal the hormonal correlates of that poor scoring, then medication — specifically therapies targeting serotonin, like SSRIs or certain anti-anxiety medications — becomes a clinical tool rather than an admission of defeat.

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

— AAHA Trends, on implications of the Jung et al. findings

This is not an argument that every anxious dog needs drugs. The researchers themselves note that salivary cortisol can be influenced by many factors, that context matters, and that a single biomarker is not a complete picture of welfare. A 2025 study in Scientific Reports found that salivary measures did not consistently reflect acute stress responses across all contexts. The science is still developing.

What this changes

What the study does shift is something more attitudinal. The dog who freezes in the waiting room, who trembles when the trash truck passes, who cannot settle no matter how long the walk — that dog may not be making a choice. Their body chemistry may be generating the experience of danger faster than any training can neutralize it.

That distinction matters. It matters for how owners talk about their dogs, how trainers design programs, and how veterinarians triage the behavioral cases that walk through their doors. The dog who won't sit in heel position is a training problem. The dog whose cortisol is five times elevated after a standard behavioral assessment is a physiological one — and probably needs both a good trainer and a good vet working together.

The dog in the waiting room, pressed flat against the wall, is not being difficult. They are telling you everything they can about what is happening inside them. This study is beginning to confirm that what they are telling you is true.