Purchase this article with an account.
Joan Liu, Charles Or, Bruno Rossion; Spatiotemporal dynamics of view-invariant face identity perception. Journal of Vision 2018;18(10):1080. doi: 10.1167/18.10.1080.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
In everyday life, humans recognise individual faces at a glance despite substantial variations in their appearance, e.g., due to changes in head orientation. Here, we investigated the neural mechanisms underlying rapid view-invariant face identity perception with electroencephalography (EEG) and a fast periodic visual stimulation paradigm. We measured individual face discrimination responses for upright and inverted faces across 4 ranges of viewpoint variations: 0° (no change), ±30°, ±60°, ±90° in separate 60-s stimulation sequences. In each sequence, images of one face identity were displayed at a rate of F=6 Hz (6 images/s) and interleaved with different face identities at fixed intervals every 7th face (F/7 Hz=0.86 Hz). At every stimulation cycle, faces varied randomly both in size (80-120%) and in viewpoint, within a predefined range. For example, in the ±90° condition, faces varied between –90° (left profile) and +90° (right profile) in steps of 5°. Periodic EEG responses at 6 Hz captured general visual processing of the face stimuli, while those at 0.86 Hz and harmonics captured identity discrimination. In the frequency-domain, all observers (N=17) showed significant face discrimination responses for upright faces, which peaked over bilateral occipito-temporal regions and decreased with increasing viewpoint variation (~50% response reduction between 0° and ±90° viewpoint variations), reflecting robust but view-sensitive processing of face identity. Decomposing the discrimination responses in the time domain revealed that this viewpoint sensitivity was driven by an early (~200–300 ms) component, while a later (~300–600 ms) component maintained stable, significant amplitudes across all viewpoint variations, indicative of view-invariant processing. Note that both components shared similar bilateral occipito-temporal topographies. Face inversion significantly decreased both view-sensitive and view-invariant responses by ~60-80%, pointing towards the involvement of face-specific processes. These findings suggest two distinct temporal stages of view-dependency in high-level representations of face identity in the human visual system
Meeting abstract presented at VSS 2018
This PDF is available to Subscribers Only