Abstract:

Most previous studies only consider face adaptation when the adaptors are faces. No face adaptation effects are observed when the adaptors and test faces do not belong to the same category. This condition implies that face adaptation is category-selective, suggesting that the adaptation occurs at the sensory processing level. The present study attempts to verify the existence of cross-category face adaptation and examines the perceptual results, conditions, and the mechanism of cross-category face adaptation. A total of 43 college students participated in the study. Among the participants, 12 participated in Experiment 1, 12 in Experiment 2, and 19 in Experiment 3. Experiment 1 explored whether cross-category face adaptation and the effect of duration (50, 100, 200, 400, 800, 1600, and 3200 ms) of adaptors exist. Participants were presented with photographs of the objects with gender feature as adaptors, then they identified the gender of a series of faces that were chosen from a morphing spectrum between some male and female faces (the tasks in Experiments 2 and 3 are similar to Experiment 1), whereas behavioral and electrophysiological aftereffects were recorded as indicators of cross-category face adaptation. In Experiment 2, we explored whether the cross-category aftereffects in Experiment 1 were caused by adaption to the feature based on gender. Photographs with items associated with gender, the corresponding names of these items, and the words to describe gender (i.e., “male” and “female”) were used as adapters. In Experiment 3, we explored whether awareness of adaptors affects cross-category face adaptation by varying degrees of attention load on adaptors (high, low, or no loads). In Experiment 1, all conditions, except for 50 ms duration of adaptors demonstrated aftereffects, and the aftereffects were inverted U-shaped that reached their maximum when the duration of adaptors was 400 ms. Results of Experiment 2 showed that all three kinds of adapters produced cross-category face adaptation. Therefore, cross-category adaptation occurs when the adaptors and test faces share the same properties in terms of the nature of task. Results of Experiment 3 showed that the magnitude of adaptation effect was greater for the no attention than low attention load conditions, which consequently was greater than the high attention load condition. The adaptation effects between the baseline and high attention load conditions were not significant. Therefore, awareness of adaptors is necessary for cross-category adaptation. In summary, the current study proved that high-level cross-category face adaptation could occur when properties or features of adaptors can be automatically inferred from the adaptors, and the adaptors and test faces share the same properties (in this study, it was gender) in terms of the nature of task.

Key words: feature association, cross-category, face adaptation, aftereffects