Elsevier

Brain and Language

Volume 92, Issue 3, March 2005, Pages 251-261
Brain and Language

A critical boundary to the left-hemisphere advantage in visual-word processing

https://doi.org/10.1016/j.bandl.2004.06.105Get rights and content

Abstract

Two experiments explored boundary conditions for the ubiquitous left-hemisphere advantage in visual-word recognition. Subjects perceptually identified words presented directly to the left or right hemisphere. Strong left-hemisphere advantages were observed for UPPERCASE and lowercase words. However, only a weak effect was observed for AlTeRnAtInG-cAsE words, and a numerical reversal of the typical left-hemisphere advantage was observed for words in a visual prototype font (a very unfamiliar word format). Results support the theory that dissociable abstract and specific neural subsystems underlie visual-form recognition and fail to support the theory that a visual lexicon operates in the left hemisphere.

Introduction

One of the most well-established findings in neuropsychology is that visual words are processed more effectively when they are presented directly to the left cerebral hemisphere than to the right (e.g., Babkoff, Faust, & Lavidor, 1997; Beaumont, 1982; Bradshaw & Nettleton, 1983; Bub & Lewine, 1988; Burgund & Marsolek, 1997; Chiarello, 1985, Chiarello, 1988; Ellis, Young, & Anderson, 1988; Eviatar, Menn, & Zaidel, 1990; Fiset & Arguin, 1999; Hines, 1978; Jordan, Redwood, & Patching, 2003; Koenig, Wetzel, & Caramazza, 1992; Krueger, 1975; Lambert & Beaumont, 1983; Lavidor & Ellis, 2001; Lavidor, Ellis, & Pansky, 2002; Leiber, 1976; Liu, Chiarello, & Quan, 1999; Schmuller & Goodman, 1979; Young, Ellis, & Bion, 1984; Young & Ellis, 1985; for a review, see Chiarello, Liu, & Shears, 2001). The most common interpretation of this phenomenon is that lexical access and language processing in general are accomplished more effectively in the left hemisphere than in the right, but this explanation is not without competition. Exploring the boundary conditions of the consistently demonstrated left-hemisphere advantage may help to determine what underlies the effect. A useful variable for exploring such boundary conditions may be the familiarity of visual formats for word presentation. A word presented in an unfamiliar visual format can be associated with the same phonological and semantic information as it would when presented in a more familiar form, but the visual processing may differ depending on the formats. Could the left-hemisphere advantage in word recognition be eliminated with the use of an unfamiliar visual format?

One prediction is that the left-hemisphere advantage should be eliminated when words are presented in an unfamiliar format. By unfamiliar format, we mean when words are displayed in such a manner that they take the shape of unfamiliar wholes (i.e., forms that—in terms of their entire holistic configurations—are unlikely to have been viewed before). This can create a situation in which processing in the right hemisphere is relatively more advantaged than when the stimuli are not unfamiliar wholes. This prediction comes from an abstract/specific neural subsystems theory.

Dissociable neural subsystems appear to operate asymmetrically to underlie visual-form recognition (Marsolek, 1995, Marsolek, 1999, Marsolek, 2004; Marsolek, Kosslyn, & Squire, 1992; for a review, see Marsolek & Burgund, 1997). An abstract-category subsystem recognizes the visual-form category to which an input belongs (e.g., the visual-form category for band/BAND) and operates more effectively than a specific-exemplar subsystem in the left cerebral hemisphere. In contrast, a specific-exemplar subsystem recognizes the visual exemplar to which an input form corresponds (e.g., the exemplar for “band” which is different from the exemplar for “BAND”) and operates more effectively than an abstract-category subsystem in the right cerebral hemisphere.

According to this theory, left-hemisphere advantages are observed typically in visual-word processing studies using familiar stimuli because an abstract subsystem operates effectively in the left hemisphere and most word processing tasks require the participant to recognize the abstract category of an input (e.g., what word is it or is it a word?) rather than the specific exemplar to which it corresponds (e.g., which specific exemplar is it?). Rarely do we need to recognize an exemplar (e.g., “band” in lowercase 12-point Times font) in word processing tasks or in everyday reading situations, and rarely do cognitive experiments require participants to do so (for an exception, see Burgund & Marsolek, 1997). Also according to this theory, the use of unfamiliar stimuli may alter the situation. Left-hemisphere advantages may not be observed in visual-word processing studies using unfamiliar stimuli because novel visual whole forms can be processed effectively by a specific subsystem, which may counteract any tendency for a left-hemisphere advantage caused by the goals of the experimental task.

Some of the evidence in support of this theory comes from visual-form classification studies. Marsolek (1995) conducted experiments in which subjects first viewed unfamiliar letter-like forms (stick figures) presented in the central visual field and learned to classify these forms into eight different categories. Then, they were asked to classify laterally presented forms into the newly learned categories. Subjects classified the previously unseen central tendencies (prototypes) of the categories more efficiently when they were presented directly to the left hemisphere than when they were presented directly to the right. An abstract subsystem should excel at storing prototypes for categories in particular because prototypes contain the features shared by many members of a category but not features that are distinctive to different exemplars in a category. In addition, subjects classified the forms that were previously presented during the learning phase more efficiently when those forms were presented directly to the right hemisphere than when they were presented directly to the left. A specific subsystem should excel at storing the previously presented exemplars in particular because it differentiates specific exemplars effectively.

These results support the theory that dissociable neural subsystems are involved. In divided-visual-field experiments, when stimuli are presented directly to one hemisphere, subsystems in that hemisphere are given advantages in timing and in the quality of the information received over subsystems in the other hemisphere (as measured via neuronal firing, e.g., Gross, Rocha-Miranda, & Bender, 1972, and amplitudes of functional magnetic resonance signals, Tootell, Mendola, Hadjikhani, Liu, & Dale, 1998, following contralateral vs. ipsilateral visual stimulation). Thus, if the characteristic processing of one hypothesized subsystem (storing prototypes effectively) is observed when subsystems in the left hemisphere are advantaged by the visual input, and if the characteristic processing of another hypothesized subsystem (storing exemplars effectively) is observed when subsystems in the right hemisphere are advantaged by the visual input, then the two subsystems appear to operate at least relatively independently.

Similar results were obtained in another study when familiar stimuli (single letters) were used, but the familiarity of the visual format varied. Bryden and Allard (1976) presented letters printed in ten different fonts to the right or left visual field. Most of the letters were identified more accurately when presented directly to the left hemisphere than to the right, but the letters printed in two of the fonts yielded the opposite result. The two fonts that led to a right-hemisphere advantage differed from the other fonts in being much more script-like and atypical. In other words, a left-hemisphere advantage was observed when letters were presented in prototypical fonts (sans serif), whereas a right-hemisphere advantage was observed when letters were presented in the relatively less familiar or less prototypical of the fonts (serif).

Other evidence more directly indicates that a specific subsystem in the right hemisphere stores novel visual wholes effectively (Marsolek, Schacter, & Nicholas, 1996). In repetition priming experiments, participants first read centrally presented word pairs (one word above the other) and then completed word–stems presented beneath (complete) context words in the left or right visual field. Word–stem completion priming that was specific to a letter-case match between prime words and test stems was found only when the context word was the same word that had previously appeared above the primed completion word and the test items were presented directly to the right hemisphere. Priming that was not letter-case specific did not depend on context or hemisphere of direct stimulus presentation. Thus, a subsystem that stores the visually distinctive information needed to differentiate lowercase and uppercase versions of the same word appears to operate more effectively in the right hemisphere than in the left. Also, letter-case specific priming was dependent on visual context, therefore this specific visual-form subsystem apparently stores word pairs as single novel whole representations.

Most important for present purposes, these results indicate that left-hemisphere advantages have been observed for visual forms that are very typical for their shape categories, such as prototypes or forms that are visually very similar to prototypes (e.g., the letter “a” in a common font). In contrast, left-hemisphere advantages have not been observed (and sometimes right-hemisphere advantages are observed) for visual forms that are very atypical for their shape categories, such as forms that are distinctive or dissimilar to the prototypes (e.g., a letter in an unfamiliar serif font), or when novel visual whole representations are stored. By the abstract and specific subsystems theory, a left-hemisphere advantage may not be observed for processing visually unfamiliar word forms.

A different prediction concerning asymmetries in processing visual words comes from the theory that the left hemisphere contains a visual lexicon that is not present in the right hemisphere (Arguin, Bub, & Bowers, 1998; Bowers, 1996; Jordan et al., 2003; Miozzo & Caramazza, 1998). For example, Jordan et al. (2003) hypothesize that stimuli presented directly to the left hemisphere are advantaged by direct access to lexical representations that allow word recognition. Words presented directly to the right hemisphere are disadvantaged by having to cross commissures to the left hemisphere for recognition to occur. In addition, neuroimaging and neuropsychological results have long suggested that areas in the left hemisphere play very important roles in visual-word recognition (Beauregard et al., 1997; Beversdorf, Ratcliffe, Rhodes, & Reeves, 1997; Damasio & Damasio, 1983; Petersen, Fox, Snyder, & Raichle, 1990; Reuter-Lorenz & Baynes, 1992). By the visual lexicon theory, left-hemisphere advantages should always be observed for visual-word recognition, no matter what kinds of word stimuli are presented.

Recently, Polk and Farah (2002) used functional magnetic resonance imaging to investigate the “visual-word-form area” of the brain. This is an area in the left ventral visual stream that has been activated by words and word-like stimuli in previous studies (Petersen et al., 1990). In the new experiment, subjects passively viewed different types of stimuli including words, pseudowords, and consonant strings as well as alternating-case versions of words and pseudowords. They observed (as Petersen et al. had before) that this area was more activated by words and pseudowords than by consonant strings. More important, no difference in activation was observed between words and pseudowords presented in alternating-case format vs. words and pseudowords presented in pure-case formats. This indicates that the left-hemisphere visual-word-form area is not sensitive to the perceptual familiarity of the stimuli. If this area underlies a visual lexicon that is needed for visual-word processing, then alternating-case words should yield the same kind of left-hemisphere advantage in divided-visual-field experiments that lowercase and uppercase words yield, despite the fact that alternating-case words are visually unfamiliar forms.

Alternating case (or case mixing) has been a very useful method of examining visual-word recognition in behavioral studies (Besner, 1983). Several recent studies have directly measured the effect of case alternation on lateralized word presentations (Fiset & Arguin, 1999; Jordan et al., 2003; Lavidor & Ellis, 2001; Lavidor et al., 2002). First, Jordan et al. have used the word-superiority effect (Reicher, 1969; Wheeler, 1970) to demonstrate that alternating-case words are processed more effectively when presented directly to the left hemisphere than to the right. In that study, a word, a pseudoword, or a consonant string was presented in uppercase, lowercase, or alternating case briefly in the left or right visual field. After the string disappeared, a row of dashes appeared in the center of the screen to correspond with the letters of the flashed word. Above and below one of the dashes, two different letters were presented. The subject then had to decide which of those two letters had been presented in that position in the previous letter string. The typical word-superiority effect was observed, in that participants were more accurate when words were presented than when consonant strings were presented. In addition, this word-superiority effect was greater when strings were presented directly to the left hemisphere than to the right, and this asymmetry effect was observed for alternating-case, uppercase, and lowercase words. This finding supports the left-hemisphere lexicon theory, given that a hallmark of the visual lexicon is that it includes knowledge of how letters can be sequenced to form valid words, and given that top–down processing from word representations to letter representations presumably underlies the word-superiority effect.

However, it is not clear whether the top–down processing underlying the word-superiority effect takes place from visual word representations to letter representations. One could argue that top–down processing from post-visual word representations (e.g., phonological or semantic word representations) can support the effect, given the highly interactive nature of different processing subsystems of the brain (e.g., for an argument that phonological recoding is involved, see Carr & Pollatsek, 1985). If so, evidence from the Reicher–Wheeler task may not strongly constrain theories of the visual processing of word forms in areas of extrastriate visual cortex.

Lexical decision tasks have been used in other studies to examine the effect of alternating case on lateralized word presentations. In these studies, word/nonword decisions were made after letter strings were presented directly to the left or right hemisphere. Fiset and Arguin (1999) reported that case alternation detrimentally affected performance following direct left-hemisphere presentations only. However, from this brief report, it cannot be verified whether a left-hemisphere advantage (over the right) was observed for the alternating-case stimuli. This experiment also lacked an uppercase condition to fully counterbalance the experiment (see Note 1 in Jordan et al., 2003). Other researchers compared uppercase, lowercase, and alternating-case performance in the lexical decision task and reported more complete results (Lavidor & Ellis, 2001; Lavidor et al., 2002). Most important, the typical left-hemisphere advantage was observed for lowercase words (magnitudes of 41 and 39 ms in Experiments 1 and 2, respectively) and for uppercase words (magnitudes of 40 and 44 ms, in Experiments 1 and 2, respectively), but it was unclear whether significant effects extended to alternating-case words (Lavidor et al., 2002). For alternating-case words, a numerical left-hemisphere advantage was found (magnitudes of 12 ms in both Experiments 1 and 2), but the reported results did not allow a test of whether these effects were significant; the researchers were interested in different questions. To the extent that the left-hemisphere advantage for processing alternating-case words may be reliable in this study, the results support the left-hemisphere lexicon theory, in that they indicate that the left-hemisphere advantage for processing words is observed even for unfamiliar visual-word forms (alternating-case words).

However, evidence from the lexical decision task may not be the most useful for testing theories of visual-word processing. Mayall and Humphreys (1996) compared performance with mixed-case words and pure-case words on several word tasks and demonstrated that case mixing had more of a cost on the lexical decision task than on either the word naming or semantic classification tasks. These findings support the idea that a “familiarity discrimination mechanism” may be involved in lexical decision (Besner, 1983). Because alternating case disrupts the outline shape of a word form, this manipulation reduces the familiarity of the overall form. Besner has argued that a crude estimate of this kind of familiarity of the overall form can be used to perform a lexical decision, independently of an identification of which word was presented. If so, the processing in this task may not reflect the processing that takes place when a visual-form input is recognized or identified per se, and results from other tasks are needed to test the main question of whether a left-hemisphere advantage is observed for unfamiliar forms of words (e.g., alternating-case words).

Therefore, in this study, the perceptual identification task was used to assess word recognition performance. In the perceptual identification task, words are presented very briefly, and participants are asked to visually identify and write down each word. Unlike the lexical decision task, this task requires identification of the word, and a crude estimate of familiarity of the overall form would not suffice for accurate performance. In addition, unlike the Reicher–Wheeler task, perceptual identification involves simple recognition of a word, and interpretation of results does not rely on assumptions about the particular representations involved in a top–down process that takes place after identification of the word.

Section snippets

Experiment 1

In this experiment, participants perceptually identified visual words presented directly to the left or right hemisphere (briefly in the left or right visual field). The words were lowercase, UPPERCASE, or AlTeRnAtInG-cAsE. The hypotheses were as follows. If the typical left-hemisphere advantage in visual-word processing is observed for uppercase and lowercase words but not for alternating-case words in a significant interaction effect, the abstract/specific subsystems theory would be

Experiment 2

As in Experiment 1, participants perceptually identified visual words presented directly to the left or right hemisphere (briefly in the left or right visual field). The words were lowercase, uppercase, or prototype-font words. The hypotheses were as follows. If the typical left-hemisphere advantage in visual-word processing is observed for uppercase and lowercase words, but not for prototype-font words, the abstract/specific subsystems theory would be supported. However, if the left-hemisphere

General discussion

The goal of this study was to explore boundary conditions for an important general finding in the neuropsychology of reading. First, the results from Experiment 1 indicated that the typical left-hemisphere advantage in visual-word processing was weak for alternating-case words. More important, the results from Experiment 2 indicated an elimination (and numerical reversal) of the ubiquitous left-hemisphere advantage in visual-word processing when unfamiliar word forms were presented. These

Acknowledgments

This work was supported in part by the Center for Cognitive Sciences, University of Minnesota, and the National Institute of Health and Human Development (HD-07151). We sincerely thank Marivelisse Rodriguez, Craig Roelke, and Charla Weiss for valuable assistance with data collection and analysis. Portions of this report will be presented at the Annual Meeting of the Cognitive Neuroscience Society, San Francisco (2004).

References (53)

  • D. Hines

    Visual information processing in the left and right hemispheres

    Neuropsychologia

    (1978)
  • A.J. Lambert et al.

    Imageability does not interact with visual field in lateral word recognition with oral report

    Brain and Language

    (1983)
  • M. Lavidor et al.

    MiXeD case effects in lateralized word recognition

    Brain and Cognition

    (2001)
  • M. Lavidor et al.

    Case alternation and length effects in lateralized word recognition: Studies of English and Hebrew

    Brain and Cognition

    (2002)
  • L. Leiber

    Lexical decisions in the left and right cerebral hemispheres

    Brain and Language

    (1976)
  • C.J. Marsolek et al.

    Computational analyses and hemispheric asymmetries in visual-form recognition

  • R.C. Oldfield

    The assessment and analysis of handedness: The Edinburgh Inventory

    Neuropsychologia

    (1971)
  • J. Schmuller et al.

    Bilateral tachistoscopic perception, handedness, and laterality

    Brain and Language

    (1979)
  • D.D. Wheeler

    Processes in word recognition

    Cognitive Psychology

    (1970)
  • A.W. Young et al.

    Left hemisphere superiority for pronounceable nonwords, but not for unpronounceable letter strings

    Brain and Language

    (1984)
  • A.W. Young et al.

    Different methods of lexical access for words presented in the left and right visual field

    Brain and Language

    (1985)
  • M. Arguin et al.

    Extent and limits of covert lexical activation in letter-by-letter reading

    Cognitive Neuropsychology

    (1998)
  • J.G. Beaumont

    Studies with verbal stimuli

  • M. Beauregard et al.

    The neural substrate for concrete, abstract, and emotional word lexica: A positron emission tomography study

    Journal of Cognitive Neuroscience

    (1997)
  • D. Besner

    Basic decoding components in reading: Two dissociable feature extraction processes

    Canadian Journal of Psychology

    (1983)
  • D.Q. Beversdorf et al.

    Pure alexia: Clinical-pathologic evidence for a lateralized visual language association cortex

    Clinical Neuropathology

    (1997)
  • Cited by (16)

    • Separability of abstract-category and specific-exemplar visual object subsystems: Evidence from fMRI pattern analysis

      2015, Brain and Cognition
      Citation Excerpt :

      Evidence for multiple dissociable object representation systems comes mostly from repetition priming experiments using divided visual field presentations (e.g., Marsolek, 1999; Marsolek & Burgund, 2003, 2005; see also related evidence from unilateral auditory presentations, González & McLennan, 2007, 2009) and fMRI (e.g., Koutstaal et al., 2001; Simons, Koutstaal, Prince, Wagner, & Schacter, 2003). However, convergent support for the dissociability of these systems also has been obtained using other methods, such as neuromodulatory evidence (Burgund, Marsolek, & Luciana, 2003), neuropsychological dissociations (Beeri, Vakil, Adonsky, & Levenkron, 2004; Vaidya, Gabrieli, Verfaellie, Fleischman, & Askari, 1998), interhemispheric transfer of visual information (Marsolek, Nicholas, & Andresen, 2002), visual word identification (Deason & Marsolek, 2005), asymmetries in relevant amygdala activations (McMenamin & Marsolek, 2013), and assessments of encoding-related activity in relation to subsequent memory performance (Garoff, Slotnick, & Schacter, 2005). Parallel functional arguments and repetition priming evidence for lateralized form-specific versus form-abstract subsystems also have recently been extended beyond the processing of individual objects to the processing of complex visual scenes (Epstein & Morgan, 2012; Stevens, Kahn, Wig, & Schacter, 2012).

    • Hemispheric differences in the organization of memory for text ideas

      2012, Brain and Language
      Citation Excerpt :

      The left hemisphere appears to retain the meanings of words in memory, whereas the right hemisphere appears to retain more information about a word’s physical form. For example, divided VF studies have shown that the right hemisphere is more sensitive than the left hemisphere to changes in font and letter case (Burgund & Marsolek, 1997; Deason & Marsolek, 2005; Lavidor & Ellis, 2001; Marsolek, Kosslyn, & Squire, 1992). Findings such as these have led some researchers to claim that the right hemisphere processes words as physical objects, encoding a representation in memory that is veridical and holistic, whereas the left hemisphere processes words more abstractly, encoding a representation that is more conceptual and categorical (Deason & Marsolek, 2005; Marsolek, Schacter, & Nicholas, 1996).

    • Communication between the cerebral hemispheres in dyslexic and skilled adult readers

      2009, Revista de Logopedia, Foniatria y Audiologia
      Citation Excerpt :

      Just how much visual word recognition can be achieved within the right hemisphere is a matter of active debate. Some theorists believe that the normal right hemisphere is capable of recognising many written words, even if it is limited in its ability to represent their meanings and sounds (Shillcock, Ellison and Monaghan, 2000; Deason and Marsolek, 2005). An alternative, and probably more popular view is that visual word recognition is conducted mainly within the left hemisphere (Cohen, et al., 2002; Ellis, 2004; Whitney, 2001; Zaidel, Clarke and Suyenobu, 1990).

    View all citing articles on Scopus
    View full text