Assessing equine prospective memory in a Y-maze apparatus.

Eight horses were tested in a Y-maze to see if they would approach the side on which they had seen and heard food being delivered into a food trough. The horses were tested in five experimental contexts: immediate release (IR), delayed release 3s (3DR), 6s (6DR), 9s (9DR) and 12s (12DR) after food delivery. Individual subjects performed five IR trials, followed by five each of 3DR, 6DR, 9DR and 12DR trials. A re-rest was performed 1 week later. Data were analysed by log-linear analysis of frequencies. The results showed that the horses were capable of achieving the correct choice during IR trials, were successful during the DR trials and improved marginally during the re-test. The findings suggest that horses could have a trainable form of prospective memory and benefit from incremental memory-training exercises. However, there is a risk of overestimating equine cognitive ability and unnecessary reinforcement delays are likely to lead to welfare and training setbacks within equitation.