Exploring the Dimensions of Object Recognition

Abstract

The structural organization of object representation in the inferotemporal cortex (IT) has been debated in the field of cognitive science for a long time. Early accounts explain object recognition abilities in terms of separated and independent modules. However, more recent studies suggest that an object space constructed from orthogonal dimensions underlies object representations in the brain. It is also possible that object representations are modular but organized according to the axes of an object space. The main dimensions of interest were those representing animacy and stubby-spiky. Alternatively, other dimensions and factors previously proposed might matter for object recognition, such as a common object recognition factor, the dimension of real-world size or whether an object is a face or not. The aim of this thesis was to assess those possible dimensions with behavioral measures, using two different tasks and both novel and familiar objects of various categories. Our results indicated that a common object recognition factor, or o-factor, underlies inter-individual differences in object discrimination, rather than specific dimensions of an object space. Whether objects are faces, novel or familiar is important for object recognition in addition to this o-factor. Furthermore, the axes of an object space might be a possible organizing principle of object representations. In addition, the importance of size seems to depend on whether an object is animate or inanimate. Further research is needed to determine the contribution of these factors and dimensions to object recognition.