Visual Processing in Indian Children withDyslexia: an Eye-Tracking StudySuvarna ChintaIIIT-Hyderabadsuvarna.rekha.11@research.iiit.ac.inCo-author: Harini SampathEye movement studies provide valuable insightin cognitive processes involved in reading.However, there are no eye movement studiesin Indian languages. The E-Z reader modelpredicts that eye movements in typical readersare characterized by linear saccadic movementswith few regressions and fixation points. Studieshave shown that the preferred visual location(PVL) and optimal viewing position (OVP) tendsto be initial and centre of a word in alphabeticallanguages. In this study, we explore if thispattern holds for Indian dyslexic children. Werecorded eye movements of dyslexic (N = 18)and typically developing (N = 16) children duringthree tasks - passage reading, auditory-word andpicture identification tasks.We computed wpm, regression percentage andfirst fixation for text analysis, and first fixationduration, gaze duration and total fixation timefor word level analysis. We also conducted PVLanalysis, which yielded word specific measuresfor the OVP. Experimental design was 2(group)X 3(tasks), and in the PVL and OVP analysis weemployed 2(group) X 2(text, word) X 3(wordzone).MANOVA were computed for theseparameters. Dyslexic readers performed moreaccurately in picture identification compared toauditory- word identification task. In paragraphreading, mean fixation duration was significantly(p<0.01) higher for dyslexic (mean 12.28 + 0.98SD) compared to typical readers (mean 5.46 +2.64 SD). Eye movements in the dyslexic readerwas characterized by multiple regressions anddo not correspond to the E-Z reader model. ThePVL and OVP found to be initial and last-zoneof a word in dyslexic readers, initial and centrein typical readers. In line with this our resultssupport dominant para-foveal processing
