Language Disorders in the Brain: Distinguishing Aphasia Forms with Recurrent Networks

Contents

This paper attempts to identify certain neurobiological constraints of natural language processing and examines the behavior of recurrent networks for the task of classifying aphasic subjects. The specific question posed here" is: Can we train a neural network to distinguish between Brocaphasics, Wernicke aphasics and a control group of normal subjects on the basis of syntactic knowledge? This approach could aid diagnosis] classification of potential language disorders in the brain and it also addresses computational modeling of language acquisition.