Abstract: Reduced rapid-eye-movement (REM) sleep is associated with impaired learning in many species. We developed a reversal learning test (RLT) suitable for field conditions to explore this association in 16 healthy horses. Nocturnal REM-like sleep behavior was recorded five times for 48 h over a six-week period. The horses performed a target training task followed by an RLT using two objects. When the horses reached a predefined frequency of touching the object, the spatial location was reversed. Mean test parameters were statistically analysed using GENLIN models, longitudinal continuous variables were analysed using linear repeated measures models, and dichotomous repeated measures were analysed using GEE models and Kaplan-Meier method. Altogether 15/16 horses completed RLT by reversing three or more times. Most errors occurred before the second and third reversals. Overall, REM sleep duration (mean ± SE) was 46.1 ± 8.7 min. However, there were ten horses that exhibited REM-like sleep for less than 30 min (10.6 ± 2.2 min, range 0.0–28.0 min), while six horses exhibited REM-like sleep for at least 30 min (42.3 ± 4.8 min, range 36.6–65.8 min). Longer REM sleep was associated with a greater number of reversals ( = 0.04), while no relationship was found with error rate. Survival analysis further indicated a significant difference in progression probability between groups, with horses with shorter REM sleep having a 50% chance of progressing after five reversals, compared to six reversals for longer REM sleepers ( < 0.05). We present here a method to test horses’ reversal learning ability on site over a single day. We found that short REM-like sleep duration without clinical signs of sleep disturbances in horses was associated with poorer performance and lower perseverance during the RLT. The online version contains supplementary material available at 10.1038/s41598-025-34463-9.
The Equine Research Bank provides access to a large database of publicly available scientific literature. Inclusion in the Research Bank does not imply endorsement of study methods or findings by Mad Barn.
This research summary has been generated with artificial intelligence and may contain errors and omissions. Refer to the original study to confirm details provided. Submit correction.
Overview
This study investigates how the duration of rapid-eye-movement (REM) sleep behavior in horses affects their performance on a spatial reversal learning test conducted in field conditions.
The researchers developed a test to measure learning flexibility and examined the relationship between REM-like sleep durations and the horses’ ability to adapt to changes in object location during the task.
Background and Purpose
REM sleep is known to play a crucial role in learning and memory consolidation across many species.
Previous research suggests that reduced REM sleep impairs learning, but this association had not been well-examined in horses under practical, field conditions.
The purpose of the study was to develop a feasible, field-side reversal learning test (RLT) to evaluate horses’ cognitive flexibility and to test whether variations in REM-like sleep behavior affect learning performance.
Study Design and Methods
Participants: 16 healthy horses monitored over a six-week period.
Sleep recording: Nocturnal REM-like sleep behaviors measured five times for 48 hours each, to quantify duration and variability.
Task: Horses completed a target training followed by a spatial reversal learning test involving two objects placed in specific locations.
Reversal learning test (RLT): When a horse reached a predefined frequency of touching an object, the location of the two objects was reversed, requiring the horse to adapt and learn the new correct location.
Data analysis:
Generalized Linear Models (GENLIN) assessed mean test parameters.
Longitudinal continuous variables analyzed with linear repeated measures models to handle data over time.
Dichotomous repeated measures (i.e., binary outcomes) assessed using Generalized Estimating Equations (GEE) models and survival analysis with Kaplan-Meier method evaluated progression through reversals.
Results
Completion and reversals:
15 out of 16 horses successfully completed the RLT, reversing the spatial location three or more times.
The majority of errors were made before the second and third reversals, indicating these stages were more challenging.
REM-like sleep durations:
Average REM-like sleep duration across all horses was approximately 46.1 ± 8.7 minutes per night.
Two distinct groups identified:
Short REM group (10 horses): REM-like sleep under 30 minutes on average (mean 10.6 ± 2.2 minutes, range 0–28 minutes).
Long REM group (6 horses): REM-like sleep at least 30 minutes (mean 42.3 ± 4.8 minutes, range 36.6–65.8 minutes).
Relationship between REM sleep and learning performance:
Longer REM sleep durations were significantly associated with completing a greater number of reversals (statistical significance p=0.04).
No significant correlation was found between REM sleep and error rate during the task, suggesting errors were not directly tied to sleep duration.
Survival analysis (Kaplan-Meier) showed horses with shorter REM sleep had a 50% probability of progressing after five reversals, whereas longer REM sleepers reached six reversals at similar probability (p < 0.05), indicating better perseverance and learning adaptability.
Conclusions and Implications
The researchers validated a simple and effective method for field-side assessment of horses’ reversal learning ability in a single day, which could be widely used outside of laboratory settings.
Results demonstrate that even in the absence of clinical sleep disturbance symptoms, horses with shorter REM-like sleep durations showed reduced learning flexibility and lower perseverance in adapting to changing conditions.
This suggests that natural variation in REM sleep behavior can significantly affect cognitive performance in horses.
Findings highlight the importance of REM sleep for cognitive function in equine species, similar to observations in other animals.
Potential practical applications include:
Improving horse welfare by monitoring sleep behavior.
Considering sleep quality when evaluating training and performance.
Additional Information
The study also references supplementary material available in the online version of the article for interested readers.
Cite This Article
APA
Hämäläinen MJ, Brotherus IL, Wigren HM, Kaimio TE, Suomala H, Olbricht AM, Hänninen LT, Mykkänen AK.
(2026).
Effect of horse sleep behavior on performance in a field-side spatial reversal learning test.
Sci Rep, 16(1), 4265.
https://doi.org/10.1038/s41598-025-34463-9
Department of Equine and Small Animal Medicine, Faculty of Veterinary Medicine, University of Helsinki, Helsinki, 00014, Finland. mira.hamalainen@helsinki.fi.
Brotherus, Iina Liisa
Department of Equine and Small Animal Medicine, Faculty of Veterinary Medicine, University of Helsinki, Helsinki, 00014, Finland.
Wigren, Henna-Kaisa Margareta
Molecular and Integrative Biosciences Research Programme, Faculty of Biological and Environmental Sciences, University of Helsinki, Helsinki, 00014, Finland.
SLEEPWELL Research Program, Faculty of Medicine, University of Helsinki, 00014, Helsinki, Finland.
Kaimio, Tuire Eriikka
Vahviste, Vantaa, 01690, Finland.
Suomala, Heli
Department of Equine and Small Animal Medicine, Faculty of Veterinary Medicine, University of Helsinki, Helsinki, 00014, Finland.
Equine Information Centre, Kiuruvesi, 74700, Finland.
Olbricht, Anna-Mari
Equine Information Centre, Kiuruvesi, 74700, Finland.
Hänninen, Laura Talvikki
Department of Production Animal Medicine, Faculty of Veterinary Medicine, University of Helsinki, University of Helsinki, Helsinki, 00014, Finland.
Mykkänen, Anna Kristina
Department of Equine and Small Animal Medicine, Faculty of Veterinary Medicine, University of Helsinki, Helsinki, 00014, Finland.
Conflict of Interest Statement
Declarations. Competing interests: The authors declare no competing interests.
References
This article includes 74 references
Brodt S, Inostroza M, Niethard N, Born J. Sleep—A brain-state serving systems memory consolidation.. 1050–1075 (2023).
Ruckebusch Y, Barbey P, Quillemot P. Stages of sleep in the horse (Equus caballus). Les etats de sommeil Chez Le cheval (Equus caballus).. 658–665 (1970).
Greening L, Downing J, Amiouny D, Lekang L, McBride S. The effect of altering routine husbandry factors on sleep duration and memory consolidation in the horse. 105229 (2021).
Bertone JJ. Sleep and sleep disorders in horses. 123–129 (2015).
Haavisto ML et al. Sleep restriction for the duration of a work week impairs multitasking performance: sleep restriction and multitasking. 444–454 (2010).
Honn KA, Hinson JM, Whitney P, Van Dongen HPA. Cognitive flexibility: A distinct element of performance impairment due to sleep deprivation. 191–197 (2019).
Martin TI, Zentall TR, Lawrence L. Simple discrimination reversals in the domestic horse (Equus caballus): effect of discriminative stimulus modality on learning to learn. 328–338 (2006).
Strobel BK, Schmidt MA, Harvey DO, Davis CJ. Image discrimination reversal learning is impaired by sleep deprivation in rats: cognitive rigidity or fatigue?. 1052441 (2022).
Ineichen C et al. Establishing a probabilistic reversal learning test in mice: evidence for the processes mediating reward-stay and punishment-shift behaviour and for their modulation by serotonin. 1012–1021 (2012).
Rolls ET, Hornak J, Wade D, McGrath J. Emotion-related learning in patients with social and emotional changes associated with frontal lobe damage. 1518–1524 (1994).
Vigouroux L, Jolivald A, Ijichi C, Yarnell K. 89 learning ability and physiological stress response of horses (Equus caballus) undergoing discrimination and reversal learning tasks using a touch screen. 103552 (2022).
Sappington BKF, McCall CA, Coleman DA, Kuhlers DL, Lishak RS. A preliminary study of the relationship between discrimination reversal learning and performance tasks in yearling and 2-year-old horses. 157–166 (1997).
Pannewitz L, Loftus L. Frustration in horses: investigating expert opinion on behavioural indicators and causes using a Delphi consultation. 105818 (2023).