November 09, 2016

Do I Know You? The Use of Facial Recognition by Capuchins

In a study recently published in Animal Behaviour, capuchin monkeys used facial recognition to discriminate faces of other capuchin monkeys across several degrees of familiarity.
By JoAnna Pendergrass, DVM
In a study recently published in Animal Behaviour, researchers at Georgia State University’s Language Research Center (LRC) used the match-to-sample (MTS) paradigm to demonstrate capuchin monkeys’ ability to discriminate faces of other capuchin monkeys across several degrees of familiarity.
 
Primates live in complex societies that make recognizing and remembering individuals an important advantage. With several modes of recognition available, including olfaction and auditory clues, primates most frequently rely on vision to recognize individuals. Human and nonhuman primates look at faces to collect social information such as emotional state. Given the similarities in the use of facial recognition in humans as in nonhuman primates, it is possible that primates share a common neural mechanism for facial recognition.
 
Several nonhuman primate species, including capuchins, have shown the ability to discriminate individual conspecific faces (faces within the same species). However, the mechanism for this ability is not yet clear.
 
Face processing studies often use the MTS paradigm, in which a study subject clicks on a sample stimulus on a screen, then views several other stimuli. The subject then selects the stimuli that matches the sample stimulus. Interestingly, real-life exposure to a particular species can help a study subject correctly match a stimulus to a sample stimulus.
 
Authors of human face processing studies have reported on the “familiarity effect,” a behavioral effect describing the ability to recognize familiar faces compared with unfamiliar faces. This effect has been studied in chimpanzees and orangutans. However, facial recognition studies, particularly those that evaluate face processing of familiar and unfamiliar individuals, are rare in New World monkeys, capuchins among them.
 
The current study used two social groups of capuchin monkeys (“group 1” and “group 2”); these groups were formed from the split of one large social group. Each group had vocal and visual access to each other when outdoors; full vocal access and limited visual access were available between groups when indoors. From these groups, eight subjects (three adult males, one subadult male, four adult females) were studied. All subjects were previously trained to use a computerized joystick and the MTS paradigm to perform various cognitive tasks, but had no prior computerized training with social stimuli.
 
Face stimuli were color photos of male and female capuchins (ten photos per capuchin), taken from different head positions and gaze orientations. The photos were divided into training and test stimuli. Training stimuli contained photos of 10 capuchins from a third group at the LRC, to whom groups 1 and 2 did not have visual access. Test stimuli contained photos of 15 capuchins: five from group 1, five from group 2, and five from an unfamiliar group of capuchins from Scotland. Test stimuli were used to assess three degrees of familiarity to the study subject: in-group (same social group), out-group (neighboring social group), and unfamiliar (Scotland group).
 
Study participation was voluntary. The eight subjects that chose to participate sat in individual testing chambers with a computer. After a sample stimulus appeared on the center of the computer screen, four other randomly-chosen comparison stimuli appeared on the screen. Subjects used a modified joystick to choose the one comparison stimulus that matched the sample; the comparison stimulus was either identical to the sample or a different photo of the sample. Only correct choices were rewarded with a treat.
 
Each test session lasted about two hours. Subjects could perform up to 1000 trials per session, but did not perform more than one session per day.
 
Authors analyzed the data using two-way mixed design analysis of variance.
 
Subjects were significantly better at recognizing the faces of in-group and out-group individuals than those of unfamiliar individuals. However, when comparing the ability to discriminate between in-group and out-group individuals, authors observed no significant difference in performance.
 
Authors identified sex differences in the ability to discriminate between faces, but these differences were not statistically significant. For example, when compared with female capuchins, male capuchins more accurately discriminated male faces. Male and female capuchins were equally able to discriminate female faces.
 
Study results support evidence of the familiarity effect in a New World primate species. Due to the fact that study subjects were better able to discriminate familiar individuals than unfamiliar individuals, authors noted that, for capuchins, “familiarity plays a significant role in the discrimination of faces by aiding recognition in familiar individuals.”
 
Given that the study subjects discriminated the faces of in-group and out-group individuals equally well, authors suggested that capuchins pay close attention to the faces of individual members of neighboring groups.
 
​Dr. JoAnna Pendergrass received her doctorate in veterinary medicine from the Virginia-Maryland College of Veterinary Medicine. Following veterinary school, she completed a postdoctoral fellowship at Emory University’s Yerkes National Primate Research Center. Dr. Pendergrass is the founder and owner of JPen Communications, LLC.

Sign up to receive the latest news in veterinary medicine.

Latest Issue

Client Education

American Veterinarian