Modeling categorization of scenes containing consistent versus inconsistent objects.