Metacontrast masking as a measure of change detection: Children with poor reading fluency show impaired change detection for both letters and shapes.

Developmental dyslexia, a specific learning difficulty in reading, manifests as effortful decoding of words and as such is commonly associated with reduced phonemic awareness. However, its underlying cause remains elusive, with magnocellular visual processing, temporal auditory processing, visual attentional deficits and cerebellar dysfunction all gaining some traction. More recent theories have concerned visual attention span, measuring the parallel attentive capacity of the sensory visual system. However the VA span task as implemented requires reports, both conscious recall and recognition of letters, that activate many cortical areas beyond sensory visual cortex. Change detection, in contrast, does not require the conscious recognition of items, but simply awareness that the stimulus has changed, or not, again testing visual attention in a parallel fashion, but avoiding the complications of higher order cognitive processes. Thus, we investigated change detection in 33 good and poor readers with ages of around 10 yr, using a gap paradigm. Groups of 4 letters or 4 shapes were presented for a fixed time (0.7 s), followed after a 0.25 s gap, by a second similar group, each item surrounded by an annular frame filled with dynamic random noise of variable contrast. Detection performance was manipulated by varying the contrast of these meta-contrast mask frames, yielding a threshold contrast of the frames at which participants could just detect change. In two separate experiments, letters and rectangular shapes were used as target items, in order to test whether previous findings of superior change detection in good compared with poor readers was a result of greater automaticity in letter recognition of the good readers. The results indicate that the good readers were able to detect change at higher levels of masking distraction for both the letter and shape targets, indicating that this difference is not specifically related to to the training of graphemic or lexical information but more likely reflects a difference in alerting or pre-recognition stages of visual processing. Together, the results provide further support of the notion that there is a low level attentional performance difference between dyslexic and normal reading children. Thus, the results further bring transient spatial attention directly into the spotlight as an ability critical for learning to read.