What is it about?
Hyperactive or impulsive behavior in dogs — similar to ADHD in humans — can be hard to diagnose because vets currently rely on owners' descriptions, which can be subjective. We developed a simple, objective tool that uses a ceiling-mounted camera and artificial intelligence to analyze a dog's movement patterns during a short visit to the vet. By tracking how the dog explores a room — its speed, turning patterns, and how much space it covers — our system generates a score that reflects the level of ADHD-like behavior. In dogs receiving treatment, this score decreased as their behavior improved. Behavioral veterinarians who reviewed our approach said it could help them monitor treatment progress and explain results to pet owners using clear, visual data. This is the first step toward making behavioral assessments in dogs as objective and measurable as a heartbeat or temperature reading.
Featured Image
Photo by Rafaëlla Waasdorp on Unsplash
Why is it important?
This is the first study to propose a fully objective, video-based machine learning method for assessing ADHD-like behavior in dogs — replacing subjective owner questionnaires with quantifiable movement metrics. At a time when pet mental health awareness is rising and AI tools are transforming clinical practice, our approach offers three timely advances: (1) it uses low-cost, widely available technology (a ceiling-mounted camera + open-source ML) rather than specialized equipment, making it feasible for routine veterinary clinics; (2) it leverages growing evidence that dogs share genetic and neurochemical underpinnings of ADHD with humans, positioning canine behavioral assessment as a translational bridge for neurodevelopmental research; and (3) it was co-designed with behavioral veterinarians to ensure clinical utility, not just algorithmic accuracy. The resulting H-score provides an objective benchmark for diagnosing severity, monitoring treatment response, and communicating outcomes to owners—potentially reducing misdiagnosis, improving welfare, and accelerating cross-species discovery in behavioral neuroscience.
Perspectives
Working on this paper was a deeply rewarding experience because it brought together three passions I hold dear: advancing machine learning methods, improving animal welfare, and fostering international scientific collaboration. As a researcher based in Saint Petersburg, contributing to a project with colleagues from Israel and the UK reminded me that science has no borders—especially when the goal is to help vulnerable beings who cannot advocate for themselves. What excites me most is seeing how a relatively simple setup — a ceiling camera and open-source algorithms — can generate objective insights into a problem that has traditionally relied on subjective reports. I hope this work encourages more AI researchers to look beyond benchmark datasets and engage with real-world clinical constraints: limited data, privacy considerations, and the need for interpretable outputs that clinicians can trust. Finally, I hope this paper inspires early-career researchers — especially those in engineering or computer science — to seek out interdisciplinary problems. The most impactful innovations often happen at the intersection of fields, where domain expertise meets computational creativity. If this work sparks even one new collaboration between a vet and a data scientist, I will consider it a success.
Dr. Aleksandr Sinitca
Sankt-Peterburgskij gosudarstvennyj elektrotehniceskij universitet LETI
Read the Original
This page is a summary of: Objective Video-Based Assessment of ADHD-Like Canine Behavior Using Machine Learning, Animals, September 2021, MDPI AG,
DOI: 10.3390/ani11102806.
You can read the full text:
Contributors
The following have contributed to this page
