NEURAL APPROACH TO SKILLED AND DYSLEXIC READING

Arizona State University-Tempe Campus, Tempe AZ

Investigators

Linked publications & trials

Abstract

The proposed research uses neurally-inspired models as a basis for understanding perception of print, the crux of reading ability. The mathematical basis of these models is dynamic systems theory; the basis for comparison is recurrent feedback. The framework is applied in widely used laboratory reading tasks with skilled, developmentally dyslexic and acquired dyslexic readers (stroke patients). The experiments build on previous demonstrations that phonology plays a key role in reading. Experiment 1 tests for "feedback" effects of phonology for skilled readers in a naming task. Feedback effects are found when a spelling that "could have" been presented (e.g., you "could" spell HEAP with two Es as in DEEP), influences performance on the word (HEAP) that is presented (i.e., performance is poorer in simple reading tasks when a Word's phonology can feed back multiple spellings). Experiment 1's design is a 2 (feedback consistent vs. inconsistent) x 2 (feedforward consistent vs. inconsistent) design that also tests for interaction effects with traditional feedforward consistency. Experiment 2 extends the feedback x feedforward design of Experiment 1 to lexical decision and adds a third between- subject factor, nonword lexicality (illegal vs. legal vs. pseudohomophone foils), that tests for strategy effects. Experiment 3 tests whether developmental dyslexics confuse word and nonword homophones based on their phonology (e.g.1 ROWS or ROZE falsely recognized as ROSE). Experiment 4 tests whether frequency effects of homophones (ROSE) are mediated by sound-alike mates (ROWS) in categorization performance by developmental dyslexic subjects. Experiment 5 extends the feedforward x feedback consistency effect to naming and lexical decision by acquired dyslexic (stroke) patients. Experiments 6 and 7 use a novel time-series method to test for on-line effects of strategic control. Manipulations of strategic control are operationalized as continuous changes, during the course of an experiment, in task demands or stimulus contexts. Continuous changes in strategic factors can result in abrupt, qualitative changes in behavior. In Experiments 6 and 7, the manipulations of strategic control are effectively manipulations of task difficulty. These experiments are prototypes for a much broader application of this new method. They define a starting point for studies of on-line "cognitive dynamics."