Parallel graded attention in reading: A pupillometric study

Joshua Snell,Sebastiaan Mathôt,Jonathan Grainger,Jonathan Mirault

doi:10.1038/s41598-018-22138-7

Joshua Snell, Sebastiaan Mathôt + Show 2 more

Open Access

https://doi.org/10.1038/s41598-018-22138-7

Copy DOI

Abstract

There are roughly two lines of theory to account for recent evidence that word processing is influenced by adjacent orthographic information. One line assumes that multiple words can be processed simultaneously through a parallel graded distribution of visuo-spatial attention. The other line assumes that attention is strictly directed to single words, but that letter detectors are connected to both foveal and parafoveal feature detectors, as such driving parafoveal-foveal integrative effects. Putting these two accounts to the test, we build on recent research showing that the pupil responds to the brightness of covertly attended (i.e., without looking) locations in the visual field. Experiment 1 showed that foveal target word processing was facilitated by related parafoveal flanking words when these were positioned to the left and right of the target, but not when these were positioned above and below the target. Perfectly in line with this asymmetry, in Experiment 2 we found that the pupil size was contingent with the brightness of the locations of horizontally but not vertically aligned flankers, indicating that attentional resources were allocated to those words involved in the parafoveal-on-foveal effect. We conclude that orthographic parafoveal-on-foveal effects are driven by parallel graded attention.

Highlights

There are roughly two lines of theory to account for recent evidence that word processing is influenced by adjacent orthographic information
We predicted that flankers positioned to the left and right of the target would have a stronger impact on target processing than flankers positioned above and below the target, in line with the idea that attention would be mainly distributed along the horizontal axis during reading
Within the alternative line of reasoning – i.e., that foveal letter detectors may be connected to parafoveal feature detectors3 – it is possible that the letter detectors are mainly connected to parafoveal feature detectors in the horizontal dimension rather than the vertical dimension

Summary

Introduction

There are roughly two lines of theory to account for recent evidence that word processing is influenced by adjacent orthographic information. The other line assumes that attention is strictly directed to single words, but that letter detectors are connected to both foveal and parafoveal feature detectors, as such driving parafoveal-foveal integrative effects Putting these two accounts to the test, we build on recent research showing that the pupil responds to the brightness of covertly attended (i.e., without looking) locations in the visual field. There was no orthographic overlap between targets and flankers, implying that the idea of parafoveal feature detectors influencing foveal letter detectors cannot account for these particular findings It must be noted that while these higher-order (e.g. syntactic, semantic) parafoveal-on-foveal effects show up in artificial reading tasks such as the flanker paradigm, they are not expressed in more natural measures of reading speed, such as fixation durations in sentence reading3,7,8 ‒ possibly because higher-order information is not integrated across words during normal reading. We report a novel methodology that builds on recently obtained evidence that the pupillary light response can reflect visual attention

Objectives

Methods

Results

Conclusion