A Preliminary Study Exploring the Relationship Between Exercise Regimes and Sleep-Related Behavior in Stabled Horses

Exercise is a recognized time cue for circadian rhythms, influencing sleep architecture and quality across diverse human populations. While its timing and intensity affect sleep outcomes in humans, little is known about how exercise impacts sleep in horses, particularly those with early morning or late evening workloads, as is common in working or competition horses. This study sought to explore this relationship in stabled horses experiencing changes to their workload patterns. Five geldings housed at Hartpury Equestrian Centre were observed across three exercise phases (normal, increased, and return to normal) using an ABA design. Sleep-related behaviors were recorded overnight using closed-circuit television (CCTV) equipment and analyzed using a validated ethogram with continuous focal sampling. Inter-observer reliability tests confirmed consistent behavioral scoring, and data were analyzed using nonparametric statistics to compare behavioral time budgets, as well as postures, across phases. Horses spent the majority of their time awake, followed by sleep and rest, across all exercise phases. Increased exercise led to significantly longer sleep bouts (Phase B: 10.79 min vs. Phase A: 5.33 min, P < 0.05) and more frequent rest episodes. Sleep postures also changed, specifically sternal recumbency, where Phases B and C were statistically different from Phase A. Results also indicated a residual effect of exercise on sleep and rest patterns, with sleep patterns remaining altered even after a return to normal exercise (Phase C: 8.98 min), indicating prolonged recovery effects. Findings demonstrate that exercise-induced changes in sleep architecture persist beyond the intervention period, highlighting the importance of recovery planning in equine management and supporting the need for larger-scale studies to disentangle the relationship between sleep and exercise to optimize equine performance and well-being.