Abstract
Fast periodic visual stimulation can be used to rapidly and objectively measure individual face discrimination (Rossion & Boremanse, 2011), with an optimal frequency rate of about 6 Hz (Alonso-Prieto & Rossion, 2012). Here we defined the tuning function for fine-grained individual face discrimination with a rapid oddball paradigm (Liu-Shuang et al., 2012). High density EEG (128-channels) was recorded in 10 participants presented with 40-second sequences of faces appearing at a rate of 5.88 Hz (sinusoidal contrast modulation, Figure 1). Each stimulation sequence contained a series of 4 repetitions of an identical original face (A, 0% morph) followed by a morphed face (B) (i.e., AAAABAAAABA…). The morphed face (B) differed from the original face (A) from 0% (same identity) to 100% (different identity), in 10% steps, for a total of 44 sequences presented in random order. EEG signal-to-noise ratio (SNR) at the identity-change frequency (every 5 stimuli, or 5.88 Hz/5 = 1.18 Hz) and its harmonics (2F/5 = 2.35 Hz, 3F/5 = 3.53 Hz, 4F/5 = 4.70 Hz) was used as a neural index of individual face discrimination (Liu-Shuang et al., 2012). Significant SNR increases were found over the right occipito-temporal cortex at the identity-change frequency and its harmonics for stimuli that differed by more than 40% of identity information, with no such evidence before this threshold (Figure 2). A complementary behavioural experiment supported the presence of a shift in the tuning function for fine-grained individual face discrimination even though participants’ results were more variable than in the EEG experiment. Overall, these findings indicate that the coding of individual faces in the occipito-temporal cortex is non-linear with respect to physical similarity between faces. They also highlight the power of the fast periodic visual stimulation approach in EEG to characterize individual face perception at a fine level.
Meeting abstract presented at VSS 2013