Can modern digital gait analysis systems rival a trained human eye? Dr Thomas Witte discusses their role in lameness detection


When I first started carrying out equine gait analysis during my veterinary degree, around 15 years ago, our techniques were still very much in their infancy.

The technology was cumbersome and so sensitive that only highly skilled technicians could get accurate results, and even then horses had to be measured in a controlled environment moving in a straight line through a predefined space. Force platforms, basically glorified high-precision weighing scales, were used to gauge the forces under a horse’s feet and infrared cameras and reflective markers attached to surface landmarks measured the resultant movement.

These techniques were similar to those used in filmmaking at that time. While the same technology still exists and has been incredibly fruitful in moving forward our understanding of equine locomotion and musculoskeletal pathology, this approach was never going to enable widespread measurement of large numbers of horses.

As with many areas, it was technological breakthroughs and spins-offs from other fields that allowed this work to advance. Inertial measurement units that calculate movement from acceleration are used in navigation systems. As the technology became smaller and cheaper, researchers started deploying inertial measurement for gait analysis.

Ultimately, the technology has become so small and lightweight that it is found in virtually every smartphone or smartwatch, and offers a non-invasive way to measure movement under real-world conditions. We first used inertial sensors on galloping racehorses to measure energy efficiency and soon started to see applications in the measurement of movement symmetry and lameness detection.

Making strides

The movement of a trotting horse should be completely symmetrical. To be more specific, when moving in a straight line the two halves of the stride cycle should result in the same mirror-image movement pattern.

As vets we rely on this fact, with any asymmetry at trot pointing to a potential lameness. The amount of asymmetry in vertical movement is used to grade and localise pain, and to monitor responses to diagnostic testing and treatment. Anyone who has studied a lame horse will have seen the familiar “head nod” of a forelimb lameness or “hip hike” of a hindlimb lameness.

With an obvious lameness in a single limb, observation by eye is very reliable at grading the degree of lameness and seeing changes. However, when more than one limb is involved or the lameness is more subtle, the human eye can sometimes be misled.

The two main failings of subjective visual assessment are low resolution and bias.

The human eye can only resolve differences in symmetry of movement to a certain level, so small differences can be hard to discern. Even more important than assigning a grade to the lameness on a single occasion is the ability to judge when lameness has changed between two examinations. The change may represent a response to nerve or joint blocks, which can help us pinpoint the origin of lameness, or to therapy such as surgery or other treatment intervention.

High-quality research into surgical and medical treatments has always relied on gait analysis to confirm treatment effectiveness, but modern miniature wireless sensors bring the benefits of definitive accurate measurement to many more horse owners than ever.

Bias is the second challenge that we work hard to avoid when assessing lameness in horses by eye. There are a range of situations in which bias is a risk. For example, if we find something on palpation of a limb we may subconsciously be influenced when watching the horse move, a reason why most vets assess movement before palpating. Similarly, it has been shown that simply knowing that a given lame horse has undergone nerve blocking changes a vet’s perception of the degree of lameness he is exhibiting.

Finally, if you’re keeping your fingers crossed that your treatment has been effective or that your favourite horse is no longer lame, this can also
lead to a biased interpretation of the movement.

Wearable technology

This is where new wireless wearable motion sensors come in. They offer the ability to measure symmetry of movement non-invasively, without recourse to an expensive lab set-up, bringing cutting-edge and precise diagnostics to the masses. Importantly, the best systems have been carefully validated against the more involved — and more expensive — lab-based gait analysis systems.

Current systems range from very simple smartphone apps, which use the phone’s built-in sensor technology to compute movement, to custom-designed sensor systems smaller than a matchbox, which can provide a greater range of results and have applications beyond lameness alone. One such system in regular use at the Royal Veterinary College uses five or more sensors and is able to measure lameness as well as the range of movement of the horse’s back.

These wireless gait analysis systems, also known as inertial measurement units (IMU) or inertial sensors, can be attached with double-sided tape.

Sensors are placed at key points on the horse’s body — the poll of the head is essential for forelimb lameness and the pelvis for hindlimb lameness. Some systems have additional sensors on the legs to measure the timing of the stride, while others can do this from the movement of the pelvis.

Ultimately, these systems all produce a very similar output. They yield various metrics of lameness which can be used to guide veterinary interventions, and can even measure horse-rider interaction. Importantly, because the sensors are non-invasive they stay in place for the duration of the examination, allowing for repeated measurements.

Although the sensors eliminate the subjective and sometimes error-prone human eye, as with any diagnostic test, vets continue to play the most important role in interpreting the results.

Symmetry of movement is just one of many features of how a horse moves, so the additional findings of physical examination will always contribute to the conclusions drawn from the results. In fact, knowing which leg the horse is lame on and to what degree is only the first step on the path to achieving a precise diagnosis and putting a therapeutic plan in place.

It will inevitably need to be followed by appropriate imaging techniques. H&H

H&H, 27 July 2017