Experience -dependent dissociations in perception and neural organization: Phenomena and underlying mechanisms.

Abstract

The architecture of cognition is widely acknowledged to be modular; different parts of the architecture are specialized for different functions. This dissertation addresses two questions about this modular organization. First, can experience lead to the development of new functional modules that are specialized for acquired cognitive functions? Second, if it can, how might it do so? I present two studies suggesting that experience can indeed lead to new functional modules and a third study investigating one potential mechanism. The first study supports the hypothesis of experience-dependent functional specialization at the behavioral level, by demonstrating that letter recognition is functionally dissociable from digit recognition. It is well known that letters are detected more efficiently in the context of digits than in the context of other letters, but some researchers have argued that this dissociation is an artifact that arises because letters are physically more similar to other letters than they are to digits. If so, one would expect this so-called alphanumeric category effect to remain relatively constant when the characters are inverted, because the inverted letters would still presumably be more similar to each other than to the inverted digits. I show that letters and digits that fail to show the category effect when inverted, nevertheless exhibit a robust effect when turned upright. This result falsifies the physical similarity hypothesis and suggests that letter recognition and digit recognition are indeed dissociable. This dissociation presumably must be experience-dependent, because the distinction between letters and digits is arbitrary and culturally defined. The second study presents evidence for experience-dependent functional specialization at the neural level, by demonstrating that typing, a learned skill typically acquired during adolescence, is neurally dissociable from matched visual and manual controls tasks. Using functional magnetic resonance imaging, I found areas in posterior parietal and dorsolateral prefrontal cortices that were selectively active for typing, but not for matched visual and motor tasks. This result suggests that an acquired cognitive motor skill can be subserved by specialized neural tissue. The third study tested a mechanistic hypothesis about how experience could lead to neural specialization for an acquired function. I tested the hypothesis that co-occurrence statistics in the environment interact with correlational learning mechanisms in the brain to produce neural segregation of learned skills. Using functional MRI, I compared patterns of brain activation before and after participants trained extensively with two sets of stimuli presented in spatiotemporal isolation from each other which, according to the co-occurrence hypothesis, should lead to neural segregation of the two sets. I failed to confirm this prediction. Together, these investigations suggest that experience can indeed lead to the development of new functional modules, both at the behavioral and the neural level, but they leave open the question of how it does so.