Binocular foveation in reading

Abstract

We present a theory of foveation in normal binocular reading. We consider the pervasive, nontrivial binocular fixation disparities (FDs) observed in reading and relate them to the computational problem of resolving retinal disparities in depth perception. We infer that the right eye's fixation being to the right of the left eye's in reading promotes binocular fusion in challenging conditions. We then show a different (nonfusional) processing advantage for the right eye's fixation being to the left of the left eye's in reading conditions in which binocular fusion is assured, by modeling the combined influence of foveal splitting, contralateral preference, ocular prevalence, and fixation disparity. This synthesis of anatomically grounded research in different aspects of visual processing produces a theory of foveation in reading that matches current data and makes testable predictions.

A binocular moving window technique to study the roles of the two eyes in reading

“…This overall pattern agrees with previous studies using EyeLink systems (Jainta et al, 2010;Nuthmann & Kliegl, 2009;Shillcock et al, 2010). The percentage of crossed fixations in normal reading was larger than in Nuthmann and Kliegl (2009) and Jainta et al (2010), but smaller than in Shillcock et al (2010). Comparing results from five studies, Shillcock et al demonstrated that a change in the mean fixation disparity appears as complementary variation between the dominant fixation disparity (crossed in the present data) and alignment, with the nondominant fixation disparity appearing relatively unaffected.…”

“…This overall pattern agrees with previous studies using EyeLink systems (Jainta et al, 2010;Nuthmann & Kliegl, 2009;Shillcock et al, 2010). The percentage of crossed fixations in normal reading was larger than in Nuthmann and Kliegl (2009) and Jainta et al (2010), but smaller than in Shillcock et al (2010).…”

“…Such disparities do not typically involve diplopia. Research continues as to the detailed parameters of this complex behaviour, the causes of the size and direction of binocular disparities (see Kirkby, White, & Blythe, 2011, for review), and their theoretical implications (Shillcock et al, 2010), but the data reported show the size of the disparity ranging widely between less than a character (effective conjointness) and several characters, probably reflecting viewing conditions, individual differences, and experimental procedures. One point of agreement among researchers on binocular disparity in reading is a preference for independent monocular calibration of the eye tracker, although there remain genuine dilemmas regarding calibration procedures (see Nuthmann & Kliegl, 2009, for discussion).…”

“…Before that, binocularity in reading was largely of interest to researchers claiming that poor binocular coordination might contribute to an explanation of dyslexia (e.g., Stein & Fowler, 1985). It emerged that binocular fixations are often not conjoint in nonreading tasks involving vergence (e.g., Cornell, MacDougall, Predebon, & Curthoys, 2003;Enright, 1998), and it further became clear that this disparity between the two eyes' fixation points extended to reading on a two-dimensional screen (Blythe et al, 2006;Jainta, Hoormann, Kloke, & Jaschinski, 2010;Juhasz, Liversedge, White, & Rayner, 2006;Kirkby, Blythe, Drieghe, Benson, & Liversedge, 2013;Kliegl, Nuthmann, & Engbert, 2006;Nuthmann & Kliegl, 2009;Shillcock, Roberts, Kreiner, & Obregón, 2010; see Kirkby, Webster, Blythe, & Liversedge, 2008, for a review of binocular coordination during reading and nonreading tasks). The two eyes frequently do not fixate on exactly the same point in a line of text; sometimes the left eye may fixate to the left of the right eye, which has been termed an uncrossed fixation disparity, or to the right, a crossed 1 fixation disparity.…”

“…Specifically, we present a novel binocular version of the moving window technique (McConkie & Rayner, 1975), and we report data from a basic experiment allowed by this technique. The particular experiment was motivated by a published theory of binocularity (Shillcock et al, 2010), part of which involves the claim that under some circumstances binocular disparity can be adaptive. (The alternative view is that binocular disparity is just error and noise.)…”