Neural systems for word meaning modulated by semantic ambiguity
Introduction
A fundamental purpose of language comprehension is the understanding of meanings of words. How does the language system in the brain recognize words with multiple meanings and solve lexical ambiguity? Over the past decade, functional neuroimaging research on brain mechanisms for semantic retrieval has yielded important but contradictory results Bookheimer, 2002, Buckner, 2003, Fletcher and Henson, 2001, Price, 2000). While it has been agreed that the left frontal lobes play a central role in subserving executive aspects of semantic processing involving semantic retrieval, semantic search, and semantic selection between concepts in working memory, it remains unclear as to whether subregions within the lateral frontal cortex contribute to different components in semantic memory.
Previous imaging findings reveal that the inferior prefrontal cortex encompassing Brodmann areas (BA) 45/47 and sometimes a portion of 44 mediates goal-driven semantic retrieval in various paradigms Bokde et al., 2001, Demb et al., 1995, Fiez, 1997, Gabrieli et al., 1998, Gold and Buckner, 2002, Petersen et al., 1989, Poldrack et al., 1999, Roskies et al., 2001, Wagner et al., 2001a, Wagner et al., 2001b, Scott et al., 2003. In parallel to these findings, the left mid-superior frontal cortex is also implicated in the retrieval of lexical semantics (Scott et al., 2003). In tasks that require subjects to make semantic decisions on words, increased brain activity is often seen in left mid-superior regions Brunswick et al., 1999, Fletcher et al., 1996, Mummery et al., 1998, Roskies et al., 2001, Scott et al., 2003. For instance, Scott et al. (2003) asked subjects to decide whether a visually exposed English noun could apply to a human (āambiguous semantic decisionā) and found that, relative to a syllable judgment on English nouns (āunambiguous phonological decisionā), semantic judgments resulted in activity in left superior frontal sites, and this brain activation was dependent on choice reaction time. Thus, these regions seem to be responsible for semantic assessment and meaning selection.
The crucial contribution of left (mid-)superior frontal areas to semantic processing agrees with brain mapping results that show that in a general cognitive system comprising but unlimited to language, the ventrolateral and dorsolateral frontal cortex are assumed to be hierarchically organized. In particular, ventrolateral frontal regions subserve controlled retrieval of representations from posterior cortices and guides active online maintenance and updating of accessed representations Christoff and Gabrieli, 2000, D'Esposito et al., 1998. Dorsolateral regions, on the other hand, mediate goal-directed selection, manipulation, and monitoring of maintained representations (Fletcher and Henson, 2001). Thus, dorsolateral brain systems operate on the products of ventrolateral cortical areas Petrides, 2000, Rowe et al., 2000, Smith and Jonides, 1999, Wagner et al., 2001a.
The present study aims to extend this important line of research by manipulating semantic ambiguity of written words with functional magnetic resonance imaging (fMRI). We employed a word generation task in which subjects covertly produced a word that was semantically related to a viewed target word. Target words used in our study either had precise meanings (ānon-semantic ambiguityā) or had several frequently used meanings such that subjects would encounter a difficulty in semantic retrieval and manipulation when performing the generation task (āhigh semantic ambiguityā). While word generation of the two types of language stimuli requires goal-directed cognitive processes, there are important differences. For words in the non-semantic ambiguity condition, brain activation as indexed by strong BOLD activity will be associated with express and direct semantic retrieval. For words in the high ambiguity condition, there will be a strong demand in semantic retrieval and semantic search because meanings of this type of words may compete with one another in the neurocognitive system. Cognitive research on the processing of ambiguous words indicates that several commonly used meanings of words with lexical ambiguity are activated synchronously and context-independent in an autonomous neural network (reviewed by Simpson, 1994). Thus, we assume that compared with the non-semantic ambiguity condition, semantically ambiguous words would recruit mid-superior frontal sites to serve goal-guided meaning manipulation and selection between maintained representations. Conversely, relative to words of high semantic ambiguity, precise-meaning lexical items would allow us to fractionate brain regions involved in rapid and direct semantic retrieval. Our comparisons between the two kinds of words should control for activation due to the visuo-orthographic and phonological processing in the word recognition system (Price et al., 1997).
Section snippets
Subjects
Eight male volunteers participated in this study. They gave informed consent in accordance with guidelines set by the University of Texas Health Science Center, San Antonio. All subjects were native Chinese (Mandarin) speakers from mainland China, ranging in age from 29 to 39 and living in the US for no more than 6 years.
All subjects were strongly right handed as judged by the handedness inventory devised by Snyder and Harris (1993). We adopted nine items including unimanual tasks (tasks which
Results
Averaged brain activation maps for semantically ambiguous words vs. semantically precise words and semantically precise words vs. semantically ambiguous words are shown in Fig. 2. Table 1 summarizes significant areas of activation for the two comparisons.
Compared with words without lexical ambiguity, semantically ambiguous words provoked very strong brain activity in the left hemispheric sites including mid-superior frontal gyrus (BAs 9, 46, and 10) and the right hemispheric sites involving
Discussion
Previous neuroimaging research on cortical mechanisms for the processing of language indicated that a distributed neural network involving the left inferior prefrontal cortex and the left temporal gyrus are relevant to semantic access and organization Bookheimer, 2002, Fiez, 1997, Price, 2000, though āthe differential roles that these areas play in semantic processing is still a matter of debateā (Price, 2000, p. 353). Research showing the contribution of the left mid-superior frontal cortex to
Conclusions
Unlike phonological encoding of visual words that recruits a concentrated neuroanatomical system, the processing of meanings of words engages a widely distributed neural network, as assumed by Jobard et al. (2003). This neural network is further complicated when one takes word category and concreteness into account Martin and Chao, 2001, Bookheimer, 2002. Our fMRI findings indicate that the neuroanatomical circuitry for semantic representation and analysis is modulated by lexical ambiguity. In
Acknowledgements
This research was supported by a Hong Kong Government Research Grants Council Central Allocation Vote (RGC CAV) group research grant (HKU 3/02C) awarded to LHT. We thank Charles Perfetti, Yonglin Pu, John Spinks, and Jinhu Xiong for their support.
References (48)
- et al.
Functional interactions of the inferior frontal cortex during the processing of words and word-like stimuli
Neuron
(2001) - et al.
Human prefrontal cortex is not specific for working memory: a functional MRI study
NeuroImage
(1998) - et al.
Common prefrontal regions coactivate with dissociable posterior regions during controlled semantic and phonological tasks
Neuron
(2002) - et al.
Evaluation of the dual route theory of reading: a metanalysis of 35 neuroimaging studies
NeuroImage
(2003) - et al.
Semantic memory and the brain: structure and processes
Curr. Opin. Neurobiol.
(2001) - et al.
Functional specialization for semantic and phonological processing in the left inferior prefrontal cortex
NeuroImage
(1999) - et al.
Attentional networks
Trends Neurosci.
(1994) - et al.
Going beyond the information given: a neural system supporting semantic interpretation
NeuroImage
(2003) - et al.
Handedness, sex and familiar sinistrality effects on spatial tasks
Cortex
(1993) - et al.
Prefrontal contributions to executive control: fMRI evidence for functional distinctions within lateral prefrontal cortex
NeuroImage
(2001)
Recovering meaning: left prefrontal cortex guides controlled semantic retrieval
Neuron
Brain regions involved in articulation
Lancet
Functional MRI of language: new approaches to understanding the cortical organization of semantic processing
Annu. Rev. Neurosci.
Explicit and implicit processing of words and pseudowords by adult developmental dyslexics: a search for Wernicke's Wortschatz?
Brain
Functional-anatomic correlates of control processes in memory
J. Neurosci.
Anterior cingulate cortex, error detection, and the online monitoring of performance
Science
The frontopolar cortex and human cognition: evidence for a rostrocaudal hierarchical organization within the human prefrontal cortex
Psychobiology
The neural basis of the central executive system of working memory
Nature
Semantic encoding and retrieval in the left inferior prefrontal cortex: a functional MRI study of task difficulty and process specificity
J. Neurosci.
A new brain region for co-ordinating speech articulation
Nature
Phonology, semantics, and the role of the left inferior prefrontal cortex
Hum. Brain Mapp.
Frontal lobes and human memory: insights from functional neuroimaging
Brain
Brain activity during memory retrieval. The influence of imagery and semantic cueing
Brain
A stereotactic method of anatomical localization for positron emission tomography
J. Comput. Assist. Tomogr.
Cited by (43)
Neural systems for auditory perception of lexical tones
2016, Journal of NeurolinguisticsEffects of representational distance between meanings on the neural correlates of semantic ambiguity
2014, Brain and LanguageCitation Excerpt :Despite the large literature on ambiguity processing in context, to our knowledge only two previous imaging studies have investigated the neural correlates of ambiguity processing outside of a disambiguating context (Chan et al., 2004; Hargreaves et al., 2011). Chan et al. (2004), using a semantic generation task, found that processing ambiguous words relative to unambiguous words was associated with increased activation in bilateral SFG and MFG, whereas processing unambiguous words was associated with increased activation in bilateral IFG, left insula, MTG and ITG. In contrast, Hargreaves et al. (2011), using a semantic categorization task, found that processing ambiguous words was associated with increased activation in left IFG, whereas no increased neural activation was associated with processing unambiguous words.