Elsevier

Brain and Language

Volume 115, Issue 2, November 2010, Pages 113-120
Brain and Language

Task effects in the mid-fusiform gyrus: A comparison of orthographic, phonological, and semantic processing of Chinese characters

https://doi.org/10.1016/j.bandl.2010.08.001Get rights and content

Abstract

The left mid-fusiform gyrus is repeatedly reported to be involved in visual word processing. Nevertheless, it is controversial whether this area responds to orthographic processing of reading. To examine this idea, neural activity was measured using functional magnetic resonance imaging in the present study while subjects performed phonological, semantic, and orthographic tasks with Chinese characters under equivalent task difficulties. One region in the left mid-fusiform gyrus exhibited greater activity during the orthographic task than during the phonological and semantic tasks, which did not differ, suggesting that this region is involved in orthographic processing to a greater extent than phonological or semantic processing. In addition, a region in the right mid-fusiform gyrus exhibited a similar effect. This right mid-fusiform activity may relate to the use of pictorial Chinese characters.

Introduction

Many studies have noted the importance of the left mid-fusiform gyrus for the successful performance of reading tasks (Benson and Geshwind, 1969, Beversdorf et al., 1997, Booth et al., 2002a, Cohen and Dehaene, 2004, Cohen et al., 2002, Dehaene et al., 2002, Dehaene et al., 2001, Dehaene et al., 2004, Price et al., 1996, Price et al., 1994, Warrington and Shallice, 1980). However, whether the left mid-fusiform gyrus (BA37) is involved in orthographic processing of reading is still a controversial issue.

This idea was first supported by pure alexia cases in which patients with lesions in this region showed preserved ability in production and comprehension of oral speech but severe impairments in visual word reading (Benson and Geshwind, 1969, Beversdorf et al., 1997). Intact comprehension of auditory words in pure alexia patients suggests that damage to this region does not affect phonological or semantic processing leading some researchers to associate the damage with orthographic representations (Cohen et al., 2004, Warrington and Shallice, 1980). In line with this idea, functional neuroimaging studies have demonstrated greater activity in this area in response to visual words than auditory words (Booth et al., 2002a, Dehaene et al., 2002). In addition, this region exhibited greater activation for words and orthographically legal pseudowords than non-words that do not have regular orthographic structure (Cohen et al., 2002, Dehaene et al., 2002). Critically, however, the difference between these materials is not only orthographic. For example, both words and pseudowords can be pronounced, while non-words cannot. In addition, words have meanings that non-words do not have. In short, these findings do not necessarily mean that the left mid-fusiform gyrus is involved in orthographic processing per se, as they can be explained by the possibility that the area is involved in the mapping of orthographic representations to phonological and/or semantic representations.

Indeed, findings from existing patient studies suggest that the left mid-fusiform gyrus is not necessary for orthographic representation. Some pure alexia patients with damage in the left mid-fusiform gyrus exhibited better performance during lexical decision, semantic categorization, and written word/picture verification tasks than during naming tasks (Bub and Arguin, 1995, Coslett and Saffran, 1989, Hillis et al., 2005, Sakurai et al., 2006). Since orthographic processing is a necessary step for lexical decision and semantic access, successful lexical decision and semantic access indicates that the orthographic representation is intact. Thus, lesions in the left mid-fusiform gyrus are not necessarily associated with impaired orthographic representation. On the contrary, data from these patients suggest that this region is involved in the phonological access that is required by a naming task.

Further support for this idea comes from a study in which Chinese subjects were trained to recognize the visual forms, the phonologies, and the meanings of words in an artificial language (Xue, Chen, Jin, & Dong, 2006). The artificial language was created by arbitrarily pairing Korean characters with phonologies and meanings, such that there was no phonological or semantic regularity in any of the pairings. The training procedure was divided into three parts—visual training, phonological training, and semantic training. In the first two weeks of the study, subjects were trained on the visual forms; in the following two weeks, they learned to link the visual forms to their (arbitrary) phonologies; in the last two weeks, they learned to link the visual forms and phonologies to their (arbitrary) meanings. Importantly, activity in the left mid-fusiform gyrus was significantly greater after phonological training than before (immediately after visual training), although it is possible that greater visual familiarity lead to this increase, since phonological training included additional visual exposure. Nonetheless, results from this study provide support for the role of the left mid-fusiform in phonological access.

Interestingly, results from the study by Xue et al. (2006) also indicate a role for semantic access in the left mid-fusiform gyrus. In particular, activity in the left mid-fusiform was greater after semantic training than before (immediately after phonological training), suggesting that learning the meanings of visual forms may also be important for left mid-fusiform involvement. Further support for this idea comes from studies observing priming modulation that depends on the semantic relationship between prime and target words (Devlin et al., 2006, Klaver et al., 2007, Raposo et al., 2006, Wheatley et al., 2005). Therefore, in addition to evidence that the left mid-fusiform is involved in phonological access, evidence also suggests a role for the left mid-fusiform in accessing the meanings associated with visual forms.

In sum, previous findings are not sufficient to support the role of the left mid-fusiform in orthographic processing, as they can be interpreted by the involvement of this region in phonological and/or semantic processing, which has been suggested by some studies (e.g., Hillis et al., 2005, Xue et al., 2006). Whether this region is indeed involved in orthographic processing in addition to its potential roles in phonological and semantic processing is still a question that needs further investigation.

One way to investigate this question is to compare activity for words, pseudowords and non-words. Although, as mentioned before, because non-words and pseudowords lack the meanings and/or pronunciations that real words have, they should not involve any semantic and/or phonological processing, this assumption is problematic and thus the comparison among words, pseudowords and non-words is not efficient for distinguishing orthographic, phonological and semantic processing. In the view of connectionist models of reading, after the orthographic processing of sublexical components, both words and pseudowords can activate a group of phonological as well as semantic representations of their orthographically and phonologically related real words (McClelland and Rumelhart, 1985, Seidenberg and McClelland, 1989). From this point of view, reading of pseudowords can involve both phonological and semantic processing. As such, evidence that pseudowords without semantic information evoked greater activation than non-words in the left mid-fusiform (e.g., Cohen et al., 2002, Dehaene et al., 2002) does not necessarily rule out the possibility that semantic processing occurred, because the meanings of real words that are orthographically and phonologically similar to those pseudowords can be activated. In addition, because the generation of correct responses benefits from prior experience, for words, the final correct representations are picked more rapidly among these spread-activated representations and produce less activation per stimulus, while the spread activation of representations by pseudowords is solved less rapidly due to the lack of prior experience, and produces more activation (Fiez & Petersen, 1998). In other words, reading of pseudowords may not only involve phonological and semantic processing, but also produce even greater neural activation than reading of words due to the phonological and semantic processing. Interestingly, using the comparison between words and non-words, Liu et al. (2008) reported a somewhat opposite pattern of results to those mentioned earlier that the left mid-fusiform exhibited greater activation for non-words than words in a study using Chinese characters. Nevertheless, because the artificial characters used in this study still possessed at least partials of the real word radicals and these unreadable sublexical components were located in novel positions, these artificial characters required more difficult location invariant orthographic processing than real characters. Therefore this result still supports the orthographic processing idea. However, it is also possible that, just as English pseudowords, the partial radicals in artificial characters might potentially activate semantic and phonological information of the real words that share these radicals and produce greater neural activation than words attributing to these semantic and phonological processes. If this is the case, this finding does not necessarily rule out a possible role of this region in phonological and semantic processing.

Another way to tease apart the three types of processing is to compare between orthographic, phonological, and semantic tasks. If the left mid-fusiform gyrus is involved in orthographic processing, this region should exhibit greater activation during the orthographic task than the other two, because there is greater orthographic demand in orthographic task. Evidence from previous studies, that is either greater activity during orthographic task than during phonological task, or equal activity during these two tasks in the left mid-fusiform, suggests that this region is not involved in phonological processing to a greater extent than orthographic processing (Bitan et al., 2005, Booth et al., 2002a, Booth et al., 2003, Cohen et al., 2004).

Nevertheless, some of these studies did not match the difficulty between tasks that possibly confounded this finding (Bokde et al., 2005, Booth et al., 2002a, Booth et al., 2003, Cohen et al., 2004, Grady et al., 1996, Gur et al., 1988, Holcomb et al., 1996). For example, in Cohen et al. (2004), the response times during the letter detection task were much longer than those during the phoneme detection task. It has been reported that phonological and semantic information is often automatically activated by written words regardless of the task being performed (Folk, 1999, Lukatela and Turvey, 1994a, Lukatela and Turvey, 1994b, Van-Orden, 1987, Van-Orden et al., 1988). Automatic activation of phonological and/or semantic information in the letter detection task may evoke some, or in this case, even greater neural activity in phonological-related and/or semantic-related regions due to the longer overall processing time, because increased task difficulty and cognitive load often associated with the increase of brain activity (Braver et al., 1997, Carpenter et al., 1999, Grasby et al., 1994, Jonides et al., 1997). Hence, without equivalent task difficulty, it is difficult to conclude that the greater activation of the left mid-fusiform gyrus in the orthographic task is necessarily associated with orthographic processing. Rather, this region might be involved in auto-activation of phonological and/or semantic processing, but not related to orthographic processing at all. On the other hand, similarly, when the phonological or semantic tasks are more difficult, greater activity in these tasks than the orthographic task may reflect greater engagement of the regions in the orthographic processing that is required for phonological and semantic processing.

Moreover, because English relies on (mostly) regular mappings between orthography and phonology, phonological rhyming tasks with English words, such as those used by Booth et al., 2002a, Booth et al., 2003, could be performed using orthographic rather than phonological information. For example, “cat” and “mat” are both phonologically and orthographically similar. As such, alphabetic languages such as English do not provide the strongest test of the left mid-fusiform’s involvement in phonological processing. Indeed, a character-based language without regular mappings between orthography and phonology, such as Chinese, may provide a better test. In Chinese, the basic writing unit, a character, represents a word as a whole, and there is no grapheme-phoneme mapping. Although for about 25% Chinese characters, the orthographic elements give consistent phonological information for the whole characters (Booth et al., 2006), the orthography of most Chinese characters does not encode phonological information at all because none of their elements are phonologically related to their pronunciation. For example, the whole character shown in Fig. 1 is pronounced “FA”, while the right element cannot be pronounced on its own, and the left element is pronounced “QU”. Thus, Chinese may be a particularly useful language for discriminating the role of the left mid-fusiform gyrus in orthographic and phonological processing. Previous studies examining orthographic and phonological tasks with Chinese characters have not observed activation differences in the left mid-fusiform gyrus (Dong et al., 2005, Liu et al., 2006), consistent with results from studies using English. Nevertheless, Dong et al. (2005) did not directly compare two tasks, and there was no behavioral data reported in Liu et al. (2006), so it is unclear whether the difficulty was matched between tasks.

Using Chinese characters without orthography-phonology mapping, the present study was designed to examine whether the left mid-fusiform gyrus is indeed involved in orthographic processing in addition to its potential roles in phonological and semantic processing. Accordingly, activity in the left mid-fusiform was measured using functional magnetic resonance imaging (fMRI) and compared during orthographic, phonological and semantic tasks that varied in their orthographic, phonological, and semantic demands but elicited similar accuracy levels and response times. If the left mid-fusiform plays a role in orthographic representation, activity in this region should be greater during the orthographic task than the others. If the left mid-fusiform does not play a role in orthographic processing, activity in this region should be greater during phonological and semantic tasks than the orthographic task or at least equal among three tasks.

Section snippets

Subjects

Seventeen subjects were recruited. All subjects were native Chinese Mandarin readers. Subjects were screened using a detailed questionnaire to ensure that they had no history of neurological or psychiatric problems. In addition, all subjects were right-handed and had normal or corrected-to-normal vision. Informed consent was obtained from each subject in accordance with the guidelines and approval of the Rice University Institutional Review Board. Data from five subjects were excluded because

Performance

Response times and error rates for orthographic, phonological and semantic tasks were compared via a one-way repeated measures ANOVA. No main effect was observed in either the response times, F(2, 22) = 1.16; p = .33 (orthographic task: 1083 ms vs. phonological task: 1077 ms vs. semantic task: 1016 ms), or error rates

Discussion

Under equivalent task difficulties, results from the present study revealed greater activity during the orthographic task than the phonological and semantic tasks in both left and right mid-fusiform regions, suggesting that the mid-fusiform gyrus is involved in orthographic processing to a greater extent than phonological or semantic processing.

Very importantly, no difference among orthographic, phonological and semantic tasks was observed at the behavioral level in the present study. Under the

References (79)

  • J. Dien

    A tale of two recognition systems: Implications of the fusiform face area and the visual word form area for lateralized object recognition models

    Neuropsychologia

    (2009)
  • Y. Dong et al.

    Neural mechanisms underlying the processing of Chinese words: An fMRI study

    Neuroscience Research

    (2005)
  • I. Gauthier et al.

    Levels of categorization in visual object recognition studied with functional MRI

    Current Biology

    (1997)
  • A.E. Hillis et al.

    The roles of the visual word form area in reading

    NeuroImage

    (2005)
  • H.H. Holcomb et al.

    Brain metabolism patterns are sensitive to attentional effort associated with a tone recognition task

    Biological Psychiatry

    (1996)
  • K. Jordan et al.

    Women and men exhibit different cortical activation patterns during mental rotation tasks

    Neuropsychologia

    (2002)
  • C. Kao et al.

    The inversion effect in visual word form processing

    Cortex

    (2010)
  • P. Klaver et al.

    Functional dissociations in top–down control dependent neural repetition priming

    NeuroImage

    (2007)
  • W. Kuo et al.

    Orthographic and phonological processing of Chinese characters: An fMRI study

    NeuroImage

    (2004)
  • C. Liu et al.

    The visual word form area: Evidence from an fMRI study of implicit processing of Chinese characters

    NeuroImage

    (2008)
  • F.M. Miezin et al.

    Characterizing the hemodynamic response: Effects of presentation rate, sampling procedure, and the possibility of ordering brain activity based on relative timing

    NeuroImage

    (2000)
  • C.J. Moore et al.

    Three distinct ventral occipitotemporal regions for reading and object naming

    NeuroImage

    (1999)
  • C.J. Price et al.

    The myth of the visual word form are

    NeuroImage

    (2003)
  • C.J. Price et al.

    Cortical localisation of the visual and auditory word form areas: A reconsideration of the evidence

    Brain and Language

    (2003)
  • A. Raposo et al.

    Repetition suppression and semantic enhancement: An investigation of the neural correlates of priming

    Neuropsychologia

    (2006)
  • Y. Sakurai et al.

    Fusiform type alexia: Pure alexia for words in contrast to posterior occipital type pure alexia for letters

    Journal of the Neurological Sciences

    (2006)
  • A.Z. Snyder

    Difference image vs. ratio image error function forms in PET realignment

  • L.H. Tan et al.

    The neural system underlying Chinese logograph reading

    NeuroImage

    (2001)
  • S.L. Thompson-Schill et al.

    A neural basis for category and modality specificity of semantic knowledge

    Neuropsychologia

    (1999)
  • K.J. Worsley et al.

    Analysis of fMRI time-series revisited—again

    NeuroImage

    (1995)
  • G. Xue et al.

    Language experience shapes fusiform activation when processing a logographic artificial language: An fMRI training study

    NeuroImage

    (2006)
  • E. Zarahn et al.

    Empirical analyses of BOLD fMRI statistics. I. Spatially unsmoothed data collected under null-hypothesis conditions

    NeuroImage

    (1997)
  • T. Allison et al.

    Face recognition in human extrastriate cortex

    Journal of Neurophysiology

    (1994)
  • A. Belger et al.

    Dissociation of mnemonic and perceptual processes during spatial and nonspatial working memory using fMRI

    Human Brain Mapping

    (1998)
  • D.F. Benson et al.

    The alexias

  • D. Beversdorf et al.

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

    Clinical Neuropathology

    (1997)
  • T. Bitan et al.

    Shifts of effective connectivity within a language network during rhyming and spelling

    Journal of Neuroscience

    (2005)
  • J.R. Booth et al.

    Modality independence of word comprehension

    Human Brain Mapping

    (2002)
  • J.R. Booth et al.

    The relation between brain activation and lexical performance

    Human Brain Mapping

    (2003)
  • Cited by (0)

    View full text