Wearable sensor shown to specifically quantify pruritic behaviors in dogs

Area of Science:

• Veterinary Medicine
• Animal Behavior
• Wearable Technology

Background:

• Traditional activity monitors in dogs measure overall activity, which can inaccurately reflect pruritus.
• Previous methods relied on owner observation, lacking objective data for conditions like orthopedic disease or post-surgical recovery.
• Distinguishing pruritic behaviors from general activity is challenging with current tools.

Purpose of the Study:

• To validate a wearable sensor system (Vetrax®) for specifically identifying canine pruritic behaviors like scratching and head shaking.
• To compare sensor data with expert-observed video annotations to differentiate behaviors.
• To develop and test a computer algorithm for accurate behavior prediction.

Main Methods:

• Collected sensor and time-stamped video data from 361 dogs (normal and pruritic).
• Independent blinded observers annotated video data for specific behaviors.
• Developed a computer algorithm to interpret sensor data and differentiate behaviors.
• Validated the system's predictive accuracy using test subject data.

Main Results:

• The system achieved 72.16% sensitivity and 99.78% specificity for head shaking.
• Scratching detection showed 76.85% sensitivity and 99.73% specificity.
• Overall accuracy for detecting scratching and head shaking was 99.24% and 99.56%, respectively.

Conclusions:

• The validated system accurately and objectively quantifies scratching and head shaking in dogs.
• While minor behaviors may be missed (sensitivity), detected instances have high confidence.
• This technology serves as a valuable tool for objective pruritus assessment in veterinary clinical and research settings.