The modeling core (MC) will be responsible forthe implementation of simulation models that will 1) identify the effects of impairments related to deficits at the level of the hypothesized neurocomputational mechanisms, and 2) provide a mechanisfic, theorefically motivated explanat ion of the relationship between these deficits and reading. Core C will relate predictions from simulation models to features of behavioral and neuroimaging data derived from all three Projects. Centralizing this function will provide this core service with rapid access to relevant data, and facilitate close collaborafion and sharing of computational resources with the neuroimaging Core B. In particular, techniques for relating the output of simulation models to patterns of activity observed using fMRI are an area of research that will benefit substantially from collaboration between the two Cores (B, C). A primary function of Core C is to clarify the bases of the hypothesized neurocomputational mechanisms by comparing different ways they can be realized within the computational models and their effects on learning and performance, generating hypotheses to be tested against behavioral and imaging data. The basic methodology involves conducfing experiments with models in which core properties are systematically manipulated. These manipulafions will inifially be motivated by performance on measures thought to implicate hypothesized mechanisms. For example, the RAN task is considered primarily a measure of processing speed, which motivates simulated differences in RAN performance by varying model parameters related to processing speed. We will then relate the model's performance at different levels of processing speed to behavioral tests of reading, and neuroimaging measures of the reading network as described below. It is recognized, however, that the one-to-one correspondence between performance on a particular task and the integrity of a particular neurocomputational mechanism is likely to be an oversimplification. In parallel with efforts in Projects 1-3 to provide better operationa/definitions of these mechanisms in terms of factors that emerge from conslderafion of a fuller behavioral/neural profile of the participants, the models will be applied to improving our theoretical understanding of what the crifical features of the mechanisms are in the context of reading disability.
