A single glance at natural face images generate larger and qualitatively different category-selective spatio-temporal signatures than other ecologically-relevant categories in the human brain

- We advance a novel visual stimulation approach with rapid changes of natural images.
- Face-, body part- and house-selective neural responses are defined objectively.
- Category-selective responses to faces are 2-4 times larger than to limbs or houses.
- First evidence of distinctive category-selective EEG topographical maps in humans.
- Spatio-temporal face-selective responses are more complex than previously described.

Although humans discriminate natural images of faces from other categories at a single glance, clarifying the neural specificity and spatio-temporal dynamics of this process without low-level visual confounds remains a challenge. We recorded high-density scalp electroencephalogram while presenting natural images of various objects at a fast periodic rate (5.88 images/s). In different stimulation sequences, numerous variable exemplars of three categories associated with cortical specialization in neuroimaging - faces, body parts, or houses - appeared every five images (5.88 Hz/5 = 1.18 Hz). In these fast periodic visual stimulation (FPVS) sequences, common low- and high-level visual processes between these categories and other objects are captured at the 5.88 Hz frequency, while high-level category-selective responses are objectively quantified at the 1.18 Hz frequency and harmonics. Category-selective responses differed quantitatively and qualitatively between faces, body parts and houses. First, they were much larger (2-4 times) for faces over the whole scalp. Second, specific and reliable scalp topographical maps of category-selective responses pointed to distinct principle neural sources for faces (ventral occipito-temporal), body parts (lateral occipito-temporal) and houses (dorso-medial occipital). Category-selective EEG responses were found at multiple time-windows from 110 to 600 ms post-stimulus onset. Faces elicited the most complex spatio-temporal profile with up to four selective responses, although body parts and houses also elicited selective responses more complex than previously described. These observations indicate that a single glance at natural face images inserted in a rapid stream of natural objects generates a quantitatively and qualitatively unique category-selective spatio-temporal signature in occipito-temporal cortical areas of the human brain.