N-Back Working Memory Task in Chinese-English Bilinguals Engages Different Subregions Within Broca's Area: A Functional MRI Study

• In Event: 45.060 - Phonological and Morphological Training Evidence Among Children and Adults: Bilingual Approach to and Empirical Evidence From Behavioral, Developmental, and Neurophysiological Studies

Abstract

1. Objectives or Purposes
Baddeley and colleagues argued that working memory consists of a domain-general central executive system, and domain-specific slave system used to rehearse presented information. In support of this model, neuropsychological studies indicate that damage to persylvania language areas can produce specific deficits in verbal working memory.

2. Perspectives or Theoretical Framework
However, attempts to use fMRI and PET to find domain-specific regions that support verbal versus non-verbal working memory have produced discrepant results. These studies have typically varied the type of information that subjects are asked to recall (e.g., letters versus locations). In this current study, we ask participants to recall only verbal information, but vary the stimuli so that the utility of a phonologically-based rehearsal strategy is minimized.

3 Methods, Techniques, or Modes of Inquiry
Prior imaging findings do not consistently dissociate verbal vs. non-verbal working memory areas. Two types of stimulus (non-homophonic character and homophone character) sets were created, and two working memory tasks (3-back and 0-back and delayed serial recall DSR fixation) were studied. 8 subjects, 1.5 T scanner, 2-short spiral, TR-2000, 26 slices, 3.8mm voxels. Data corrected for movement, detrended, mean normalized, smoothed, and warped to common reference brain.

4. Data sources, Evidence, Objects, or Materials
First, we collapsed data across the homophone and non-homophone conditions. Second, data from Chinese-English bilingual subjects who performed DSR & Nback tasks with English words. CE bilingual reference brain and statistical map warped to the Chinese reference brain. Overlap maps created using AFNI. Stimulus-dependent activation evaluated via paired-tests comparing percent change of the homophone vs. non-homophone activation from baseline in each brain region.
5. Results and/or Conclusions or Warrants for Arguments/Point of View
We observed activation within the left angular gyrus that was greater for non-homophone than homophones. This is consistent with neuro-psychological findings implicating temporo-parietal cortex in phonological processing. However, the region is not usually observed in Nback tasks. Both C-E bilinguals and Chinese only subjects showed activation in Broca’s area, and the performance of Chinese subjects is disrupted during concurrent articulation for homophone and non-homophone

6. Scientific or Scholarly Significance of the Study
Neuro-cognitive theories of working memory functioning in reading suggests that WM task (N back tasks) encourages the E-C bilinguals to engaged in phonological processing. In contrast, Chinese only subjects may perform these WM task based upon visual perceptual or orthographic representation. Furthermore, there are functionally different subregions within Broca’s area (dorsal BA 44 vs. ventral BA 45). These subregions have different temporal patterns of activation and sensitivity to stimulus manipulation among Chinese-English bilinguals after years of language experiences in contrast to the Chinese monolinguals.

Authors

• Qun Guan, University of Science and Technology Beijing

• Corrine Durisko, University of Pittsburgh

• Julie Fiez, University of Pittsburgh