Answers are grounded in this paper and cite their sources.
Background Fatigue in sport horses is a multifactorial phenomenon that critically affects performance, injury risk, and welfare, yet its real‐time, field‐based assessment remains an unmet challenge. Objectives To synthesise evidence on the use of wearable sensors and machine learning for fatigue‐related monitoring in sport horses and to identify priorities for future fatigue assessment. Study Design Scoping review. Methods This scoping review was conducted in accordance with PRISMA‐ScR guidelines. Publications from 2015 to 2026 were identified through Web of Science, Scopus, and ScienceDirect. Evidence from 102 publications was synthesised across five sensing modalities—heart rate (HR)/heart rate variability (HRV), thermometry, inertial measurement units (IMUs), surface electromyography (sEMG), and GPS—alongside machine learning applications for feature extraction, state classification, and risk modelling. Results Our synthesis reveals a fundamental disconnect: while each technology captures distinct physiological or biomechanical correlates of load, including autonomic response, thermoregulatory strain, gait stability, neuromuscular activation, and external workload, none provides a standalone signature of fatigue. Only a minority of studies used fatigue as a direct modelling label, with most relying on proxy outcomes such as asymmetry or performance decrement. Main Limitations The included literature was heterogeneous in sensor modalities, study designs, outcome measures, and application contexts. In addition, the limited use of directly defined fatigue outcomes restricted comparison across studies and constrained conclusions regarding fatigue‐specific monitoring. Conclusions The path forward lies not in developing yet another single‐sensor solution, but in establishing a multidimensional, context‐aware framework that integrates complementary data streams, individualised baselines, and interpretable models. Future research should prioritise: (1) operationalised fatigue definitions; (2) longitudinal, within‐horse monitoring designs; (3) robust multimodal fusion architectures; (4) rigorous external validation across equestrian disciplines; and (5) explainable artificial intelligence to bridge the gap between algorithmic output and actionable training decisions.
No takes yet. Share an insight, caveat, or question.
Li et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: