Split fovea theory and the role of the two cerebral hemispheres in reading: A review of the evidence

Abstract

Split fovea theory proposes that when the eyes are fixated within a written word, visual information about the letters falling to the left of fixation is projected initially to the right cerebral hemisphere while visual information about the letters falling to the right of fixation is projected to the left cerebral hemisphere. The two parts of the word must be re-united before the word can be recognised. Bilateral projection theory proposes instead that visual information is projected simultaneously to both hemispheres provided that it falls within the fovea (defined as the central 2–3°). On this more traditional account, no interhemispheric transfer would be required in order to read a word presented within the fovea. We review the evidence in support of split fovea theory and consider some of the objections that have been raised. We argue that a split fovea affects the reading of words at fixation, something that must be recognised and accounted for by cognitive, computational and neural models of reading.

Introduction

The essence of split fovea theory (SFT) is that the fovea is anatomically and functionally divided down the middle, with all visual information that originates to the left of fixation projecting initially to the right cerebral hemisphere while all visual information that originates to the right of fixation projects first to the left cerebral hemisphere. In contrast, the longer-established bilateral projection theory (BPT) proposes that while information presented in the left and right visual fields outside the fovea projects to the right and left hemispheres respectively, foveal information (usually taken to be the central 2–3°) is projected simultaneously to both hemispheres. That would include all the letters in a centrally fixated word that falls within the fovea.

When skilled readers are processing connected text, some words are skipped, but most are fixated. When a word is fixated, the fixation point typically falls about one-third of the way into the written word (O’Regan and Jacobs, 1992, Rayner, 1998). If the BPT is correct then, under normal reading conditions, the whole of a word that is contained within the fovea will be projected to visual areas in both hemispheres, including the language-dominant hemisphere from which word recognition can proceed. Word recognition will include access to the word's meaning (semantics) and its spoken form (phonology). If, however, the SFT is correct, then those letters in a centrally fixated word that fall to the right of fixation will be projected initially to visual areas in the left hemisphere while those letters that fall to the left of fixation will be projected first to visual areas in the right hemisphere. The two parts of the word will need to be brought together for identification, which would presumably involve the transfer of those letters projected to the non-dominant hemisphere across the corpus callosum to the language-dominant hemisphere. In readers with left hemisphere language dominance, which is the majority of readers, that would involve transfer of letters that fell to the left of fixation from the right hemisphere to the left hemisphere, where they could be re-united with the letters that fell to the right of the original fixation. If the SFT provides a better account of foveal processing of words and other visual stimuli, then researchers will need to explore the implications for cognitive, computational and brain-based models of reading, including how and where the left and right parts of a fixated word are re-united, and what problems might arise as a result of deficient callosal transfer (cf. Brysbaert, 1994, Brysbaert, 2004, Ellis, 2004, Ellis, 2009, Henderson et al., 2007, Lavidor and Walsh, 2004; Monaghan & Shillcock, 2008; Shillcock et al., 2000, Whitney, 2001, Whitney and Cornelissen, 2008).

There are three main lines of evidence relevant to the issue of whether the human fovea is or is not split, and whether any putative split affects visual word recognition. The first line of evidence comes from research involving patients with loss of vision in one or other visual hemifield (hemianopia). SFT predicts that when vision is lost in one visual hemifield, that loss will extend right up to the midline, even within the fovea. In contrast, because the BPT proposes that the fovea projects to both hemispheres, the BPT predicts that the fovea should be spared in hemianopic patients in whom only one visual field is damaged. The second line of evidence relevant to whether the human fovea is split or not concerns so-called ‘split-brain’ patients whose cerebral hemispheres have been surgically disconnected, either wholly or partially. SFT predicts that the problems such patients experience in transferring visual information between hemispheres as a result of the severing of callosal fibres should mean that one hemisphere will be unable to access visual information falling in the opposite visual field, even when that input falls within the fovea. The BPT theory makes the opposing prediction that because the fovea projects in its entirety to both hemispheres, each hemisphere will retain full access to foveal information even when extrafoveal information can no longer be transmitted between hemispheres. The third and final line of evidence comes from studies of word recognition in normal, healthy readers which explore parallels between the processing of whole words in the left or right visual fields (beyond the fovea) with the processing of the left and right halves of centrally fixated words. We will review each of these lines of evidence in turn, paying attention to the various points raised by Jordan and Paterson (2009) in their critical assessment of research in this area, an assessment which comes down in favour of BPT over SFT. Other reviews of the evidence, including more extensive coverage of older research, can be found in Brysbaert, 1994, Brysbaert, 2004, Ellis (2004), Leff (2004), and Lavidor and Walsh (2004). We should note that while our primary concern is with the implications of a split fovea for understanding reading processes, some of the crucial studies in the three different areas employed simple, nonverbal stimuli such as small dots or geometric figures. The importance of those studies for discriminating between BPT and SFT is such, however, that we will include them in this review.