Abstract
How does the visual system integrate object parts into coherent global visual representations? We investigated this question using faces, a category of stimuli which is perceived more holistically than other visual objects. Holistic face perception was tested by means of frequency-tagging (Regan & Heron, 1969). Participants saw a whole face whose left and right halves flickered at different frequencies (at 4 or 5 Hz, counterbalanced across different stimulation blocks of 90 seconds). In two other conditions, the face halves were spatially separated by 0.29 or 1.74 visual degrees. Electroencephalogram (EEG) was recorded (128 channels) and transformed by means of Fast Fourier Transform (FFT). EEG signal showed posterior responses at fundamental frequencies of stimulation (4 and 5 Hz), and at several interaction terms (e.g., 4 + 5 = 9 Hz) at occipito-parieto-temporal sites, with a right hemisphere advantage. EEG power at fundamental frequencies increased significantly as soon as half faces were moved apart from each other, independently of the amount of spatial separation. Most interestingly, EEG power of the first order interaction term (9 Hz) showed the opposite pattern of behaviour than that of the fundamental frequencies, decreasing when the face halves were separated. To control for interpretation of the interaction term as reflecting a border effect, we conducted a second experiment with a condition in which half faces were vertically misaligned but shared the same amount of border as in the normal whole face condition. The results were replicated: increase of the fundamental frequencies and decrease of the first interaction term in the vertically misaligned condition. Overall, these results show that single face parts are less strongly represented when they are integrated into a coherent whole stimulus. Moreover, the presence and behaviour of interaction terms in the frequency-tagging method provide evidence for integration of facial parts into a holistic representation.
Belgian National Fund for Scientific Research (Fonds de la Recherche Scientifique -FNRS).