Abstract
A unique trait of cartoon imagery is that it employs abstract symbolic elements, which require learning or culture to understand, in addition to literal iconic elements that resemble features of the real world. Our previous research has demonstrated that more abstract or "cartoonized" iconic images of faces communicate emotion more quickly and efficiently than photorealistic images of faces, and that such heightened communicative value relies on low-level features such as simplicity and contrast. Outstanding questions concern whether iconic facial features (e.g., :)) can be replaced with symbolic ones (e.g., :#) and still be rapidly perceived as being "facelike" with the acquisition of emotional meaning. In the present study we employed a face-sensitive ERP component, the N170, as an index to examine this question. EEG was collected during a probe task in which 23 participants labeled expressions on cartoon faces (happy, sad, neutral, and no emotion) that had either iconic or symbolic features. A control condition employed the same stimuli without eyes, eliminating the facelike configuration. This task was performed before and after a training task in which participants learned that symbolic features represented facial emotions and were trained to criterion. Peak N170 activation was extracted 160-220ms after stimulus onset. Results showed that N170 amplitudes were altered with training for symbolic faces only, such that after the training task they were equivalent to those observed for iconic faces. No changes were observed for iconic stimuli or either type of stimulus in the control condition. These results indicate that simply learning that arbitrary symbols conveyed emotional meaning increased rapid and relatively automatic perception of symbolic faces as "facelike." Follow-up studies explore which aspects of face stimuli, such as the presence of specific features or configural arrangements, are more pliable to symbolic manipulations.
Meeting abstract presented at VSS 2017