Using inertial sensors for assessing movement symmetries in trotting dogs : a feasibility study

Abstract: Lameness or other disturbances in the musculoskeletal system are common reasons why dog owners seek veterinary care. The lameness is most often associated with pain, degenerative changes in joints or muscles, or injury. Evaluation of lameness have traditionally been performed subjectively by means of visual assessment and scoring. Such methods are well established in field of equine medicine where the examiner typically observes the ventro-dorsal movements of the head, the withers and the pelvis during locomotion. A major problem with subjective assessment, however, is that different examiners will arrive at different results. This has been confirmed in studies carried out both within the fields of equine medicine and small animal medicine. Unfortunately, the situation becomes even more challenging when it comes to companion pets such as dogs, as their movements are both faster and smaller in magnitude compared to those of large animals. Computer assisted lameness analysis with motion capture systems or inertial measurement units,IMUs,is becoming more and more common in horse clinics today as a more objective tool to evaluate and assess lameness. Data is collected during exercise and analyzed in a computer program after which movement asymmetries can be assessed. These methods can be helpful both in everyday activities and in more complicated cases, as the asymmetries can be monitored over time. To our knowledge there is no similar commercial tool available for companion pets such as dogs yet. However, in a recent study at SLU which evaluated the use of IMUs on dogs trotting on a treadmill using a Commercial system for horses with a modified software, good results were seen in moderate induced lameness scenarios. More studies are desirable to determine whether or not the method can be used in less controlled forms, e.g., in a hallway or in a corridor at a veterinary clinic. The purpose of the this work has therefore been to investigate whether it is feasible to use standard IMUs, signal processing and software algorithms to reliably assess lameness of grade 2-3 in dogs under conditions prevailing at a regular veterinary clinic. Thirteen clinically sound dogs were included in the study. None of the them had a history of orthopedic conditions or joint surgery. Reversible distal limb disturbances were induced, mimicking supporting limb lameness, in all dogs by placement of cotton wool wads under the paw. Reversible proximal limb disturbances, mimicking swinging limb lameness, were induced by placement of a custom-made weight above the carpusortarsal joint, respectively. Inorderto obtain measurements we utilized a commercial measurement system developed by Delsys Inc. We used four IMUs in parallel located at the head, wither, pelvis and at one of the forelimbs. The data was analyzed offline using different algorithms for lameness assessment. The results indicate that it possible to use standard IMUs to evaluate lameness in dogs, but that more work is needed to robustify th ealgorithms. Compared to horses the physical behavior of dogs and their smaller sizes leads to new challenges that need to be addressed in future research.