This study focused on understanding how nonhuman primates eat by breaking down their behavior into two main phases: searching for food and actually eating it. To do this, researchers developed nine different ways to observe these behaviors using both manual methods and advanced computer analysis.

They tested these methods on three rhesus macaques, giving them different types of food and monitoring their hunger levels. The experiments were conducted in a special cage designed for studying eating behaviors, using both natural and artificial food.

Results showed that certain measures, like how often the macaques approached the food and how long they spent eating, were most useful for understanding their hunger levels. Another measure, combining how actively they sought food with how long they ate, helped explain how the taste of the food affected their behavior.

Using visual plots, the researchers could clearly see how hunger and food taste influenced the macaques' actions. They found that using deep learning for analysis was not only very accurate but also much faster than traditional manual methods, speeding up their research significantly.

Overall, this study introduced effective new ways to study how nonhuman primates eat, providing valuable insights into how hunger and food preferences shape their behavior.