Face detection dissociates from face recognition: evidence from ERPs and the naso-temporal asymmetry

Abstract

Neuropsychological data indicate that face processing could be distributed among two functionally and anatomically distinct mechanisms, one specialised for detection and the other aimed at recognition (de Gelder & Rouw, 2000; 2001). These two mechanisms may be implemented in different interacting regions of the human brain with sub-cortical structures (SC) dedicated to face detection and cortical structures (FFA and LOC) dedicated to face recognition. This model is consistent with recent data from developmental studies showing that the temporal hemifield of the retinotectal system may be involved in face preference in neo-natal observers. The present study used electrophysiological measurements (high density EEG) to test the hypothesis that neuro-anatomical correlates of a primitive face detection mechanism could be found. Participants were 10 normal adults and two developmental prosopagnosics. They were shown B/W photo-realistic or schematic images of faces and objects presented for 30, 50 or 100 ms. Monocular stimulus presentations were used with the stimulus presented in the periphery in order to stimulate either the temporal or nasal hemifield. We focussed on early extrastriate activities (occipital P1 and occipito-temporal negativities). Our results provide evidence for a temporal-nasal asymmetry for faces at around 150 ms. The same pattern obtains in normal adults and in developmental prosopagnosics but only for shorter durations.