Towards a computational model of Dyslexia

Abstract

Dyslexics are diagnosed for their poor reading skills. Yet, they characteristically also suffer from poor verbal memory, and often from poor auditory skills. We now hypothesize that dyslexia can be understood computationally as a deficit in integrating prior information with noisy observations. To test this hypothesis we analyzed performance in two tones pitch discrimination task using a two-parameter computational model. One parameter captures the internal noise in representing the current event and the other captures the impact of recently acquired prior information [1]. We found that dyslexics' perceptual deficit can be accounted for by inadequate adjustment of these components: low weighting of their implicit memory in relation to their internal noise (Figure ​(Figure1).1). Using ERP measurements we found evidence for dyslexics' deficient automatic integration of experiment's statistics (Figure ​(Figure2).2). Taken together, these results suggest that dyslexia can be understood as a well-defined computational deficit. Figure 1 Estimated parameters of the Implicit Memory Model. Estimated values of η (weighting of implicit memory) as a function of estimated values of σ (percentage of internal noise) of Controls (blue) and Dyslexics (red). The optimal weighting ... Figure 2 Grand Average ERPs to the two-tone stimulation. A. Controls B. Dyslexics. Trials are sorted according to the trial type, Bias+ (where the impact of previous trials improves performance) and Bias- (where the impact of previous trials impairs performance). ...