Monkeys lack humanlike behavioral and neural inversion costs in face identity recognition

There is long-standing confusion in the scientific literature as to whether macaque monkeys show, like humans, a large decrease of performance for recognizing the identity of faces presented upside-down. Clarifying this issue is important because performance decrease for inverted faces is a hallmark of our species’ expertise at face identity recognition (FIR), and the macaque monkey is widely considered as the reference animal model to clarify the neural mechanisms of this function. Here I report a theoretically-oriented critical review of behavioral and neural studies regularly cited as having tested the effect of inversion on FIR in macaque monkeys. I show that the confusion originates essentially from studies in macaques showing a mere behavioral or neural preference for upright over inverted faces independently of identity recognition, or confounding effects of stimulus inversion with training and stimulation constancy. Considering studies that properly tested the effect of inversion on FIR makes it clear that there is no behavioral or neural cost of such stimulus manipulation in macaques. This conclusion is congruent with a wide source of neurofunctional evidence indicating that the macaque monkey is not a valid animal model for human face identity recognition.