Abstract
Category selectivity in the visual system is frequently observed with fMRI, either as local or distributed patterns of activity for specific object categories like faces, tools, body parts, or letters. Experience can modify activity in object processing areas, as demonstrated by the engagement of face-selective areas after subordinate-level individuation training with novel objects. However, the specific role of experience per se in determining any given pattern of category selectivity is weak because little empirical work has been concerned with measuring the neural changes resulting from different types of experience. Here, we trained two groups of participants in two different learning regimens that required either subordinate-level individuation or basic-level categorization of a set of novel objects (Ziggerins). Individuation training involved learning to categorize the Ziggerins at a subordinate level quickly, similar to how people discriminate faces during person identification. Categorization training involved learning to rapidly recognize at the basic level Ziggerins that were presented in spatially organized arrays with coherent styles, similar to how people process letters when reading a text. The two regimens resulted in different patterns of changes in fMRI responses. Local activity in the fusiform gyrus increased after individuation training and was correlated with the magnitude of configural processing for Ziggerins measured behaviorally. In contrast, categorization training caused more distributed changes, with increased recruitment of the medial portion of the ventral occipito-temporal cortex relative to more lateral areas. These results demonstrate that objects with the same geometry can be processed in qualitatively different ways (e.g., focal vs. distributed patterns of specialization) because of the different recognition demands of the different training regimens using the same set of objects. We suggest that the role of prior experience in determining responses in the visual system may have been underestimated because training experience is rarely manipulated explicitly in other studies.
This research was supported by grants from the James S. McDonnell Foundation, the National Science Foundation (to I.G. and T.J.P) and the National Eye Institute (to I.G.).