A Think Again Media Group publication
Sport Debunk

Do horses only sleep standing up?

Horse REM sleep: what the overnight recordings show

By Phil Springer, Editor · Published · Sport Debunk

Horse REM sleep: what the overnight recordings show

Horse REM sleep needs a resting place on the ground, even though a horse can obtain other sleep while standing. The distinction appears directly in overnight electrical recordings. In a 2026 study led by Lisa Zimmer at the University of Veterinary Medicine Vienna, horses in REM lay on their sides or rested on their chests with their heads supported on the ground. Lying completely flat on the side was not the only workable posture. Support for the head and neck mattered because this sleep stage comes with very low muscle tone.

Zimmer's team studied ten teaching-herd horses over four consecutive nights in each of two housing arrangements: a communal stable and adjacent individual stalls. A noninvasive recording mask collected electrical signals, while video documented posture. The researchers manually classified 30-second segments rather than trusting automated software designed for human sleep. They combined the electrical patterns with the synchronized video and excluded segments that could not be interpreted. This linked a visible resting position to a measured sleep state instead of treating every motionless horse as asleep.

The recordings separated states that can look deceptively similar from a distance. Intermediate non-REM sleep occurred while horses stood or lay down. Slow-wave sleep could also occur standing, with the head and neck lowered. REM was identified using characteristic eye movements together with electrical activity and posture. In the individual-stall setting, REM was lowest on the first night and increased across subsequent nights. The authors cautioned that previous experiences and differences between the settings could have contributed. The pattern supported acclimation without establishing one housing arrangement as universally superior.

Linda Greening and Sebastian McBride's 2022 review supplies the broader physiological context. The stay apparatus allows standing rest, while the loss of muscle tone associated with REM makes bodily support essential. Their synthesis distinguishes sternal recumbency, resting on the chest, from lateral recumbency, resting on the side. Neither the amount of standing nor the amount of lying is a perfect substitute for measuring sleep stages. Chest-down rest can include more than one stage, and a horse on the ground need not be in REM throughout that interval.

Space on that ground was manipulated directly in Joan-Bryce Burla and colleagues' 2017 experiment. Thirty-eight horses in eight groups experienced four conditions, each lasting 11 days. The researchers varied the bedded area from none to half, equal to, or one-and-a-half times the Swiss minimum dimensions, with hard rubber mats in the nonbedded areas. They observed the final 72 hours of each condition. With no bedding, lying bouts were rare and most horses spent less than 30 minutes per day recumbent. Fifteen lay down for less than 30 minutes across the entire three-day observation period.

Providing bedding changed how the horses used the resting space. They strongly favored it over rubber mats or firm ground. Increasing the available area also changed the social pressure around rest: lower-ranking horses experienced more interrupted lying bouts in the smaller bedded areas. This was an experiment on recumbency, rather than a direct electrical measurement of horse REM sleep. Its value lies in identifying a physical and social constraint on the posture that makes REM possible. A resting surface can exist yet remain difficult for particular horses to use.

Light has also been tested experimentally, with more restrained results than a simple rule about bright and dark stables would suggest. In Greening and colleagues' 2025 PLOS ONE crossover study, ten riding-school horses experienced two lighting systems for four weeks each. One used timed, blue-enriched white light by day and dim red light at night; the other used fluorescent light switched on and off manually. Video behavior showed no detected difference between the systems in total sleep, standing sleep or chest-down and side-down recumbency. Day and night did affect several behaviors in both conditions. These particular systems were not interchangeable with every possible nighttime disturbance.

A separate Finnish study published in Scientific Reports in 2026 connected rest with a field learning task. Sixteen horses were observed for REM-like sleep behavior and trained to touch objects for food; the rewarded spatial position was then reversed. Fifteen completed at least three reversals. Longer REM-like sleep was associated with completing more reversals, but not with a lower error rate. The authors interpreted the difference as potentially involving perseverance or motivation. Behavioral estimates and an association between variables do not establish that adding a fixed number of sleep minutes will produce a particular training result.

Across these experiments, the useful distinction is between standing still, lying down and entering a particular sleep stage. Horse REM sleep requires an opportunity for supported recumbency, while electrical recordings identify when that opportunity becomes REM. Bedding, usable space and adaptation to surroundings influence the opportunity itself. The upright doze remains a normal part of equine rest, but it cannot perform every job in the night's sleep.

True or False?

Horses occasionally have to lie all the way down to get proper REM sleep.

Show the answer

The Record

Sure, they can doze upright like four-legged fence posts, and they do it plenty. Horses also occasionally have to lie all the way down to get proper REM sleep. Standing and lying down serve different parts of their sleep.

Send this question to a friend

Read next

By clicking “Send me the real story” you agree to Sport Debunk’s Terms of Use and Privacy Policy.