A tale of two recognition systems: Implications of the fusiform face area and the visual word form area for lateralized object recognition models