Approaches to the Rehabilitation of “Phonological Assembly”: Elaborating the Model of Nonlexical Reading

There are a variety of ways that models of normal cognition can be applied to rehabilitation of patients with cognitive deficits, and many of them are illustrated in this volume. This chapter concerns a somewhat different type of application. It focuses on the contribution that attempting to carry out a treatment study can make to the development and expansion of the models themselves. We will try to demonstrate, using the outcome of an attempt to teach a severely aphasic/dyslexic patient to “sound out” print, that thinking about how to improve patients’ deficits forces one to consider carefully a set of issues that are often ignored or downplayed in cognitive neuropsychology. In order to design a set of therapy tasks directed at specific components of a cognitive model, it is essential to have some idea about the answers to the following questions: What sorts of representations and processes does the model postulate? If the model is schematised using the typical boxes and arrows format, this question asks that details about the contents of the boxes and the nature of the arrows be specified. Can the components of the model (i.e. the boxes and arrows) be analysed with respect to their potential recovery through treatment? How do isolated components of a specific model (the boxes) interact with or exploit other specific cognitive processes from other domains? To what extent do isolated components of specific models require that more general cognitive capacities are intact?