Strength of resting state functional connectivity and local GABA concentrations predict oral reading of real and pseudo-words

Lisa C Krishnamurthy,Daphne Greenberg,Robin D Morris,Dina M Schwam,Bruce Crosson,Venkatagiri Krishnamurthy,Kenneth R Pugh,Douglas L Rothman

doi:10.1038/s41598-019-47889-9

Abstract

Reading is a learned activity that engages multiple cognitive systems. In a cohort of typical and struggling adult readers we show evidence that successful oral reading of real words is related to gamma-amino-butyric acid (GABA) concentration in the higher-order language system, whereas reading of unfamiliar pseudo-words is not related to GABA in this system. We also demonstrate the capability of resting state functional connectivity (rsFC) combined with GABA measures to predict single real word compared to pseudo-word reading performance. Results show that the strength of rsFC between left fusiform gyrus (L-FG) and higher-order language systems predicts oral reading behavior of real words, irrespective of the local concentration of GABA. On the other hand, pseudo-words, which require grapheme-to-phoneme conversion, are not predicted by the connection between L-FG and higher-order language system. This suggests that L-FG may have a multi-functional role: lexical processing of real words and grapheme-to-phoneme processing of pseudo-words. Additionally, rsFC between L-FG, pre-motor, and putamen areas are positively related to the oral reading of both real and pseudo-words, suggesting that text may be converted into a phoneme sequence for speech initiation and production regardless of whether the stimulus is a real word or pseudo-word. In summary, from a systems neuroscience perspective, we show that: (i) strong rsFC between higher order visual, language, and pre-motor areas can predict and differentiate efficient oral reading of real and pseudo-words. (ii) GABA measures, along with rsFC, help to further differentiate the neural pathways for previously learned real words versus unfamiliar pseudo-words.

Highlights

Reading is a culturally invented activity that is predominantly and explicitly an acquired higher order cognitive skill requiring complex learning over years of education, exposure, and practice
We evaluated this composite neurocognitive model by using a multi-modal approach of resting state functional connectivity (rsFC) Magnetic resonance imaging (MRI) and gamma-amino butyric acid (GABA) and GLX spectroscopy to identify resting brain networks that are associated with oral reading of previously learned words and unfamiliar words
In this study we demonstrated the capability of resting physiological measures to predict single word oral reading abilities

Summary

Introduction

Reading is a culturally invented activity that is predominantly and explicitly an acquired higher order cognitive skill requiring complex learning over years of education, exposure, and practice. From a clinical standpoint, resting state functional connectivity (rsFC) data provides critical evidence for why the network is not establishing an efficient reading circuit, especially since children and adult readers show different resting network profiles[1] These previous reports establish the groundwork for the utility of MR-based rsFC measures in examining the resting physiology in typical and struggling readers, which has been used as a biomarker across multiple studies despite the inherent variability in fMRI data that may reduce the detection power of brain differences across cohorts[14,15,16]. Investigating the anterior cingulate cortex (ACC), Horowitz-Kraus et al.[17] identified that dyslexic kids have a negative association between measures of executive function and choline, as well as executive function and myo-inositol These studies are important, and establish the utility of understanding neurochemical differences between typical and disabled readers.

Objectives

Methods

Results

Conclusion