Brain potentials and integration of external and internal features into face representations.

In this ERP study we analyzed how different orders of presentation of external and internal features influence the integration of facial components into face gestalts. Participants carried out a face-feature matching task in which, in each trial, external (E) and internal (I) facial features were presented separately and in sequence, followed by a complete unfamiliar face (matching or mismatching). For the E-I group of participants the order of presentation was external features, internal features and then the complete face. The I-E group viewed the internal features first. Mismatch effects in complete faces were more conspicuous and lasted from about 300 to 600 ms with the order E-I, while those with the order I-E were scarcely observable, and significant only from 450 to 470 ms. The external features tended to elicit a larger P150, while the N170 was preferentially associated with the internal ones. A P360, probably indicating stimulus relevance, was present in both groups for external features, while it was associated with internal features only in the I-E group. These results suggest that in the E-I order the binding of facial features into a single face representation occurs according to a stepwise process which facilitates the integration. In turn, in the I-E order the processing related to the two sets of features appears more dissociated and is of a more componential nature. Moreover, we propose that the external features may be especially relevant for object categorization, while internal features would be more closely related to subsequent configural mechanisms.