Elsevier

NeuroImage

Volume 31, Issue 2, June 2006, Pages 513-519
NeuroImage

Rapid Communication
Connectivity–behavior analysis reveals that functional connectivity between left BA39 and Broca's area varies with reading ability

https://doi.org/10.1016/j.neuroimage.2005.12.040Get rights and content

Abstract

Correlations between temporal fluctuations in MRI signals may reveal functional connectivity between brain regions within individual subjects. Such correlations would be especially useful indices of functional connectivity if they covary with behavioral performance or other subject variables. This study investigated whether such a relationship could be demonstrated in the context of the reading circuit in the brain. The method proved sufficiently powerful to reveal significant correlations between the reading abilities of subjects and the strength of their functional connection between left Brodmann's area 39 and Broca's area during reading. This suggests that the disconnection of the angular gyrus previously reported for dyslexic readers is part of a larger continuum in which poor (but nonimpaired readers) also show reduced connectivity to the region. In addition, it illustrates the potential power of paradigms that examine correlations between behavior and functional brain connections.

Introduction

The neural circuitry involved in reading has been investigated extensively using neuroimaging techniques and lesion studies. Specific language areas have been implicated repeatedly, including left inferior frontal gyrus and superior and middle temporal cortex. In addition to the classic language areas, regions in the occipito-temporal and extrastriate cortex, as well as left Brodmann's area 39 (often referred to as left angular gyrus, although it also includes the posterior superior aspect of the middle temporal gyrus), have been consistently reported to be implicated in reading (Beauregard et al., 1997, Constable et al., 2004, Fiez et al., 1999, Fiez and Peterson, 1998, Pugh et al., 2001).

Neuroimaging activation studies map out brain regions involved in specific cognitive processes, and there is considerable interest in examining the connectivity between those brain areas. Diffusion tensor imaging methods allow the examination of anatomical connectivity, while measurements of interregional correlations in activity patterns (as inferred from functional imaging data) allow assessment of the state-dependent functional connectivity between regions. Functional connectivity does not itself imply a direct, causal influence of one region over the other (otherwise known as effective connectivity), as it is correlational in nature. However, it can provide important insights into the functional relationships between brain areas.

Previous studies have examined the functional connectivity of the angular gyrus during reading. Using positron emission tomography, Horwitz et al. (1998) examined correlations in regional cerebral blood flow between a locus in left BA39 and other parts of the brain during reading tasks and found that the correlations were markedly reduced in dyslexic readers as compared to healthy control subjects. A functional magnetic resonance imaging study (fMRI) then extended this finding by reporting disrupted correlations between activations of left Brodmann's area (BA39) and activations in posterior reading areas in dyslexic readers specifically for reading tasks involving phonological processing (Pugh et al., 2000). These two studies assessed functional connectivity between brain regions via interregional correlations computed across subjects.

An alternate method for looking at functional connectivity in the brain that has gained popularity more recently is to examine interregional temporal correlations within subjects over time (Biswal et al., 1995, Hampson et al., 2002, Lowe et al., 1998, Xiong et al., 1999). This is the approach used in this study. It has the potential advantage that the strength of functional connection between two regions is assessed for each subject individually and thus theoretically can be related to behavioral variables such as task performance (or other subject variables such as age). To simplify discussion, this particular type of brain–behavior analysis will hereafter be referred to as connectivity–behavior analysis, with the intent that connectivity be understood to refer to the strength of specific interregional functional covariations in the brain. If behavioral measures and brain connectivity can be successfully related, such connectivity–behavior analyses could prove extremely useful for investigating a variety of brain functions and/or dysfunctions.

In this work, we hypothesized that correlations in the fluctuations in blood oxygenation level-dependent (BOLD) MRI signals from different parts of the reading circuit could be detected reliably in individual subjects in a reasonable time for scanning. In addition, we hypothesized that the magnitudes of those correlations would covary with behavioral measures of reading ability. A specific aim was to determine whether functional connectivity with left BA39 is related to reading ability in healthy readers. Originally, to allow the most straightforward comparison with previous studies of connectivity between brain areas involved in reading, the intent was to examine correlations between left BA39 and the rest of the brain. However, due to difficulties in functionally localizing this region (interestingly, it only activated reliably in the more skilled readers), maps of functional connectivity to Broca's area were computed instead. The strength of the correlation between reading ability and functional connectivity between Broca's area and left BA 39 was then measured across subjects.

Section snippets

Subjects

Nineteen native English-speaking, literate, right-handed adults with no reported psychiatric diagnoses or history of reading or learning disability participated in the study. Subjects ranged in age from 18–48. There were ten men and nine women. All subjects gave informed consent in accordance with a protocol reviewed and approved by the Human Investigations Committee of the Yale School of Medicine.

Reading tests

Four subtests of the WIAT-II were administered to subjects in a separate session from the scan.

Results

As mentioned in the Introduction, the block design sentence-reading task did not consistently activate left BA39 in all subjects. This was surprising at first, but examination of the data revealed that subjects that had low scores on the reading tests were generally the individuals who had little or no activation of BA39. Indeed, activation of this region was highly correlated with reading ability. The correlations between left BA39 activation and reading ability were significant for three of

Discussion

The correlations found between left BA39 activation and reading ability were pronounced. As a result, we could not functionally localize left BA39 in the poor readers. However, the strong correlation of activation in this region with reading ability is an interesting discovery. Angular gyrus/BA39 activation has been previously reported to be diminished in dyslexic subjects (Shaywitz et al., 1998) across a range of reading tasks. However, to our knowledge, activation of the angular gyrus/BA39

Acknowledgments

We thank Michael Westerveld for helping with the selection of the reading skill tests used, Ken Pugh for information regarding reading circuitry, Jed Meltzer for informative discussions about statistical analyses, Einar Mencl for sharing his software, and Karen Martin for technical assistance. This material is based upon work supported by the National Science Foundation under Grant No. BCS-0121926, and by NIH NS38467, NIH NS EB00473, and NIH EB00461 (JCG).

References (19)

There are more references available in the full text version of this article.

Cited by (155)

  • Resting-state functional connectivity and reading subskills in children

    2021, NeuroImage
    Citation Excerpt :

    With respect to reading, previous studies using a RSFC approach have shown that networks of functional connectivity align closely between task-based and task-free fMRI data. Using a seed-based approach, Hampson et al. (2006) observed consistent patterns of functional connectivity from the left inferior frontal gyrus in adults, both while they completed a sentence reading task and while they rested quietly in the scanner. The correlations observed from the seed region, the left inferior frontal gyrus, included positive correlations with areas adjacent to the inferior frontal gyrus, and positive correlations extending bilaterally to the superior and middle temporal gyri, the medial occipito-temporal cortex, middle frontal gyrus, and angular gyrus.

View all citing articles on Scopus
View full text