Are representations of unfamiliar faces independent of encoding modality?

It is well documented that both featural and configural information are important in visual face recognition. Less is known, however, about the nature of the information underlying haptic face recognition and whether or not this information is the same as in vision. In our experiments we found better within modal than crossmodal face recognition performance suggesting that face representations are largely specific to each modality. Moreover, this cost in crossmodal performance was found to be independent of differences in exploratory procedures across the modalities during encoding. We found that crossmodal face perception was most efficient when configural information of the facial features was preserved suggesting that configural information is shared across modalities. Our findings suggest that face information is processed in a similar manner across vision and touch but that qualitative differences in the nature of the information encoded underlies efficient within modal relative to crossmodal recognition.