Category learning is accomplished by several different brain systems that are specialized for the acquisition of different forms of information used to make category judgments. As discussed previously, nonverbal rules may be stored in corticostriatal loops. Exemplar-based categories, which form the basis of our semantic knowledge of the world, appear to be stored in the cortex of the temporal lobe. This is particularly true for categories we learn about through visual experience with the world (e.g., animals and plants) and thus may be thought of as the output of the ventral visual stream, or "what" pathway. Categories of actions may be represented in frontal regions associated with movement planning. Data from fMRI studies have revealed that categories such as houses, faces, and chairs activate distinct regions of posterior temporal cortex (Fig. 4). Perhaps categories that are defined more conceptually than visually (animals that live in the jungle and things you can buy at a supermarket) are represented more anteriorly in the temporal lobe. Patients with anterior and lateral temporal lobe damage exhibit a breakdown in semantic knowledge including deficits in sorting items into categories that had been learned previously. This phenomenon is seen most readily in patients with semantic dementia, but it also occurs with the diffuse temporal lobe damage seen in Alzheimer's disease.
The existence of patients with category-specific agnosias has been used as support for the idea that stored semantic knowledge in the brain is organized to some extent by category. Patients have been described with specific deficits in identifying members of such categories as faces, animals, tools, or foods as the result of brain damage. It may be that knowledge about these different categories is stored separately in distinct
brain regions. In this view, damage to one region could disrupt knowledge of animals, for example, but not knowledge of other objects.
The view that there are category-specific regions of the brain has been challenged by the fact that different categories may have different processing requirements and thus brain specificity may reflect modules for different types of processing. For example, animals are composed of basically the same parts in the same configuration (eyes, legs, tail, etc.), and they are identified by the relative sizes and shapes of those parts. In contrast, tools generally contain parts that are present in some other tools but not others, and they are identified by the location of the parts relative to each other. A broom and a brush are differentiated by the relative position of the handle and bristles. Thus, brain regions associated with identifying animals may actually be important for processing the metric properties of the parts of items. "Tool areas'' could actually be involved in processing the relationships between parts. It may be that knowledge about different category members is stored in a distributed fashion throughout the brain, but that specific cortical areas are involved in analyzing particular stimulus properties that are confounded with category membership.
Was this article helpful?