Electrophysiological decoding captures the temporal trajectory of face categorization in infants

The adult human brain rapidly distinguishes between faces at around 170 milliseconds after stimulus onset. In the developing brain, however, the time course of face discrimination is poorly understood. To shed light on this issue, we presented human and nonhuman primate faces to five to thirteen-month-old infants in an event-related electroencephalography experiment. Using time-resolved decoding based on logistic regression we detected above-chance discrimination of human faces from nonhuman faces in a time window starting at around 200 milliseconds, originating from occipito-temporal electrodes. There was no evidence, however, for above-chance discrimination of individual human or individual nonhuman faces. Moreover, using neural network-based decoding, we delivered the proof of principle that face categorization but not individuation can be detected even at the level of single participants. These results indicate that rapid face categorization emerges already in preverbal infants.