Measurement of Canine Ataxic Gait Patterns Using Body-Worn Smartphone Sensor Data.

Area of Science:

• Veterinary Neurology
• Biomechanical Engineering
• Machine Learning in Animal Health

Background:

• Ataxia in dogs is characterized by impaired coordination and gait disturbances.
• Current diagnosis relies on subjective neurological examinations and scales.
• Objective, quantitative methods for ataxia assessment are needed.

Purpose of the Study:

• To investigate the utility of smartphone inertial sensors for detecting canine ataxia.
• To evaluate the accuracy of machine learning algorithms in classifying ataxic versus healthy canine gaits.

Main Methods:

• Collected inertial sensor data (accelerometer, gyroscope) from smartphones placed on dogs' backs during walking.
• Analyzed 770 walking sessions from 55 healthy dogs and 23 dogs with ataxia.
• Applied machine learning techniques, including K-nearest neighbors, for gait analysis and classification.

Main Results:

• The K-nearest neighbors algorithm achieved 95% accuracy in distinguishing between healthy and ataxic dogs.
• Smartphone-based inertial sensing effectively captured gait alterations associated with ataxia.
• The study demonstrated the feasibility of using consumer electronics for objective veterinary diagnostics.

Conclusions:

• Smartphone inertial sensors offer a promising, non-invasive tool for canine ataxia detection.
• This technology has the potential to aid in early diagnosis and monitoring of treatment efficacy for canine ataxia.
• Future development of smartphone applications could facilitate widespread use in veterinary practice.