The effect of different training experiences on object recognition in the visual system

Abstract

Functional specialization occurs in the visual object recognition system in terms of specific behavioral phenomena assoicated with different object categories, and their selective activations of different neural regions of the ventral occipito-temporal (VOT) cortex. Several factors have been suggested to explain such category selectivity, yet little empirical work has probed the individual contributions of a single type of factors. The current study adopted a novel-object training paradigm inspired by Gauthier et al. (1997, 1998, 2002) to probe the role of recognition experience. Two groups of participants went through training procedures that mirror either face perception or letter perception with the same set of novle objects (Ziggerins). Face-like training involved learning to categorize the Ziggerins at a subordinate level, similar to how people discriminate among faces that share a similar spatial configuration in person identification. Letter-like training involved learning to recognize, at the basic level, Ziggerins presented in spatially organized clusters with coherent styles, similar to how people process letters when reading a text. The effects of face-like training were similar to what had been found for face perception. There was a drop in the basic-level advantage and an increase in holistic and configural processing. Besides, activity in the fusiform face area increased after training, with the amount of increase correlated with the magnitude of the configural effect. In contrast, letter-like training led to a larger basic-level advantage, and increased activity in the medial parts of the VOT cortex. The lack of changes in the letter-selective regions suggested that the letter-like training may better be characterized as basic-level categorization training. Results of the current study provided support for the role of recognition experience in determining category selectivity in the object recognition system. The training paradigm used also demonstrated its potential to be powerful tool for the study of category selectivity and perceptual learning in general.