Abstract
Recent behavioral studies have shown that an individual human face can be recognized rapidly and efficiently by means of two main sources of information: diagnostic three-dimensional (3D) shape and two-dimensional (2D) surface reflectance (texture and color) information (O'Toole et al., 1999; Lee & Perrett, 2000; Jiang et al., 2006; Russell et al., 2006, 2007). However, the time-course of the respective contribution of these two sources of information during the faster recognition of individual faces remains unclear.
Here we aimed at clarifying the temporal characteristics of neural representations of shape and reflectance information diagnostic for individual faces. We used a 3D morphable model (Blanz & Vetter, 1999) to generate pairs of face stimuli that were either identical, varied in 3D shape only, in 2D surface reflectance only, or in both 3D shape and 2D surface reflectance. During a face identity adaptation paradigm in event-related potentials (ERPs) (Jacques, d'Arripe & Rossion, 2007), sixteen participants discriminated individual faces in these 4 kinds of pairs, in which the first (adapting) face was presented for several seconds (∼ 3000 ms) rapidly followed (ISI ∼250 ms) by a target face (200 ms).
While participants‘ behavioral matching performances were as accurate and fast for diagnostic 3D shape as for 2D surface reflectance, the time-course of brain activation following the target face indicated that shape was diagnostic of facial identity at about 160 ms (during the face-sensitive N170 time window) well before reflectance, especially in the right occipito-temporal cortex. Both kinds of information combined at about 250 ms, leading to equally large effects as reflected several hundreds of milliseconds later in the observers‘ behavior. These observations indicate that diagnostic information about a particular individual's face accumulates faster in the occipito-temporal cortex for 3D shape than 2D reflectance diagnostic cues.