Association vs dissociation and setting up appropriate criteria for object agnosia

Abstract

We are grateful to Geskin and Behrmann (Citation2017) for carrying out their exhaustive review and, in doing so, bringing the issues discussed in the review and the accompanying commentaries to the fore. Twenty years ago, only a handful of congenital prosopagnosia (CP) cases had been reported, so the number of cases in the review is a testament to the rapid growth of CP research. The review is motivated by a question that received attention even in the earliest CP research (McConachie, Citation1976): what is the relationship between face and object recognition? Their review indicates that around 80% of CPs tested have deficits with object recognition, whereas about 20% have face-specific deficits. These figures led the authors to suggest that face and object recognition are likely to depend on a common mechanism and that face-specific mechanisms are unlikely to exist. In our commentary, we discuss two points that challenge this conclusion. First, because the inferential value of cases showing a dissociation between two abilities is greater than the value of cases showing an association between the same two abilities, we feel the substantial proportion of CPs with face-specific deficits actually provides support for a model where face and object recognition depend on different processes. Second, we believe that the criteria used to classify participants as impaired with objects are too liberal, and we show below that they can lead to misidentification of object deficits even in a substantial proportion of a control sample.