Abstract
A popular hypothesis holds that people with dyslexia have reduced sensitivity to visual motion, possibly originating in abnormalities in the magnocellular pathway. Recently, it has been proposed that dyslexics perform poorly on behavioral psychophysics in general for reasons unrelated to sensitivity, including suboptimal decision-making strategies. We revisit the question of sensitivity to visual motion in poor readers using the drift diffusion model as a tool to tease apart sensitivity from other aspects of decision-making. Seventy five children ages 8–12 with varying levels of reading skill participated in a motion direction discrimination task in which the coherence of the random dots varied from 6% to 48%. Accuracy and reaction times were collected from each participant on 240 trials. Psychometric functions fit to the accuracy data replicated the trend that dyslexic have, on average, slightly higher thresholds than their typically reading peers, although this finding was not significant (⊠ = 0.053, SE = 0.34, p =0.136; estimated Cohen’s d = 0.5). Reaction time measures revealed a larger effect: median reaction time was negatively correlated with reading ability (⊠ = −0.21, SE = 0.055, p < 0.001), and a significant interaction of reading ability and response (correct or incorrect) revealed a tendency for poor readers to make relatively faster errors than strong readers (⊠ = −0.15, SE = 0.055, p = 0.007). After estimating the parameters of the drift diffusion model for each subject, we found not only that the quality of evidence amassed from dot motion is indeed lower in poor readers (lower drift rate), but also that dyslexics employ a less consistent decision boundary. Thus, some dyslexics may indeed be less able to extract evidence from moving stimuli, but a roughly equal fraction of variance in reading skill can be explained by suboptimal decision making factors unrelated to the stimulus.