Gait Analysis Is Now Detecting Muscular Disease in Dogs Before Symptoms Are Obvious and That Changes Treatment Timing

Golden Retriever Muscular Dystrophy — GRMD — is a naturally occurring canine equivalent of Duchenne Muscular Dystrophy in humans, the most common and most severe form of inherited muscle disease. Dogs with GRMD carry a genetic mutation that disrupts dystrophin production, causing progressive muscle degeneration that begins before birth and produces measurable functional changes from the earliest weeks of life. A study using accelerometry and linear discriminant analysis has developed a gait analysis model that can reliably distinguish GRMD dogs from healthy dogs throughout their development — and detect the effects of treatment with greater sensitivity than previous methods. 🐾

Researchers B. Fraysse, I. Barthélémy, J. Hogrel, and their team performed accelerometric gait analysis twice monthly between the ages of 2 and 12 months on 8 healthy and 20 GRMD dogs, then applied linear discriminant analysis — a statistical approach that identifies the combination of variables that best separates defined groups — to seven accelerometric parameters. The methodological development the paper describes is not only relevant to GRMD research. It represents an advance in how objectively measurable movement data can be used to track disease progression and treatment response in dogs more broadly.

Why Gait Analysis of Muscular Disease Is Harder Than It Sounds

Accelerometry produces continuous, objective movement data — but that data is complex, multidimensional, and difficult to interpret without analytical frameworks that reduce it to clinically meaningful signals. Earlier accelerometric analysis of GRMD dogs had limited sensitivity, partly because the data complexity made it difficult to identify which combinations of parameters most reliably captured the disease phenotype across the developmental timeline.

The three-model development described in the paper illustrates this challenge directly. A simple model selecting genotype — healthy or GRMD — as the only outcome variable produced reasonable discrimination between groups but was not sufficiently representative of how the disease changes across time. Adding age as an additional variable without appropriate handling produced a high misclassification rate of 83.2 percent — the model could not reliably place individual dogs in the correct disease-stage category. Only when age was incorporated both as a dependent and as an explanatory variable simultaneously did the model achieve a misclassification rate below 12 percent across the full developmental period.

The final validated model — Model 3 — was then tested on blinded data from 81 healthy and GRMD dogs not used in model development. In all but one case, the model correctly matched gait phenotype to actual genotype. That external validation on nearly six times the original sample size is the finding that establishes the model’s real-world reliability rather than its performance on the data it was trained on.

What This Means for Preclinical Treatment Research ⚠️

The most directly significant application of the model was the re-analysis of data from a previous study examining immunosuppressive treatment effects on gait quality in GRMD dogs. When Model 3 was applied to that existing dataset, it identified significant treatment effects on gait quality — corroborating the original findings from that study while providing direct statistical analysis with greater sensitivity and more interpretable output than the original method had produced.

This is where the paper’s contribution to the broader field of canine disease research becomes clearest. GRMD dogs serve as the primary large-animal preclinical model for Duchenne Muscular Dystrophy — a human disease affecting approximately 1 in 3,500 male births, currently incurable, and characterised by progressive muscle degeneration leading to loss of ambulation typically in the early teenage years and cardiac or respiratory failure in early adulthood. Therapeutic interventions being developed for Duchenne require preclinical validation in a model that closely mirrors the human disease, and GRMD dogs provide that model with a degree of biological fidelity that rodent models cannot match.

A gait analysis tool that can detect subtle treatment-related improvements in movement quality with higher sensitivity than previous methods accelerates the identification of promising therapeutic candidates. Every improvement in preclinical measurement sensitivity shortens the path from potential therapy to human trial by reducing the sample sizes needed to detect a real effect and the time required to confirm whether an intervention is working.

What Objective Gait Analysis Means for Everyday Dog Owners 🐕

The methodological advances in this paper will primarily matter to veterinary researchers working with GRMD dogs and to the broader Duchenne research community. But the principle it embodies has direct relevance for owners of any dog with a progressive musculoskeletal or neuromuscular condition — and for owners of breeds with elevated risk for conditions that affect movement.

Subjective human assessment of how a dog is moving is systematically less sensitive than objective measurement, for the same reasons discussed throughout this research series in relation to emotional state perception. Owners adapt to gradual change. The dog compensates for mild functional loss through subtle postural adjustments that spread load and reduce observable asymmetry. A dog whose movement has deteriorated meaningfully over six months may still look to their owner as moving “about the same” because the comparison is being made to a recent baseline that was itself already compromised.

Objective gait analysis — whether through veterinary pressure walkways, force platforms, or increasingly accessible consumer accelerometry tools — provides the kind of change-over-time data that subjective assessment cannot reliably produce. For Golden Retrievers, German Shepherds, Labradors, and other breeds with elevated musculoskeletal disease prevalence, having a baseline gait assessment conducted during health and repeating it at regular intervals creates the reference point against which meaningful deterioration can be detected early rather than late.

At Zoeta Dogsoul, seeing the whole animal accurately is the foundation of NeuroBond — and seeing an animal accurately includes seeing how they move, not just how they behave. A dog whose gait has shifted is a dog whose physical experience of the world has changed, and that change shows in every interaction if the owner is paying the right kind of attention. Objective measurement tools make that attention more precise. 🐾

Source: Fraysse, B., Barthélémy, I., et al. (2017). Gait characterization in golden retriever muscular dystrophy dogs using linear discriminant analysis. BMC Musculoskeletal Disorders. Published April 12, 2017.