Deviant functional connectivity patterns in the EEG related to developmental dyslexia and their potential use for screening

Abstract

Developmental dyslexia (DD) is a common learning disorder with potential neural origins. While EEG-based brain activation measures combined with machine learning have shown promise for DD screening, these approaches often lack validation on independent participants, a crucial step for practical application. This study developed an EEG-based screening approach and investigated the neural correlates of DD in Chinese children. EEG was recorded from 130 children (82 DD, 48 typically developing; 7–11 years) during resting state (eyes-open, eyes-closed) and verbal working-memory tasks. After artifact rejection, signals were segmented and converted to functional-connectivity (FC) measures across delta, theta, alpha, and beta bands using Pearson correlation coefficients (PCC), phase-locking value (PLV), and a rho (RHO) measure. Segments were split into two non-overlapping samples to ensure participant-level independence: Sample 1 for training and five-fold cross-validation of a convolutional neural network, and Sample 2 for held-out cross-sample evaluation. Balanced accuracy (BA) served as the primary outcome, with significance assessed by permutation testing. Beta band matrices were most informative: the eyes-open beta-band RHO achieved the highest within-sample performance (BA = 97.47 %), and the eyes-closed beta-band PLV yielded above-chance cross-sample performance (BA = 64.99 %, permutation p < .001). Given the reduced cross-sample BA, further validation in larger and more diverse cohorts and other language systems is needed to establish the model’s generalizability. Discriminative FC patterns revealed that children with DD exhibited reduced temporal-parietal and central connectivity but increased frontal-central connectivity, likely reflecting compensatory mechanisms. Within the DD group, stronger FCs showed significant negative correlations with Chinese word reading accuracy and fluency. These results highlight functional network abnormalities in Chinese children with DD and offer preliminary evidence for EEG-based screening. However, current performance remains exploratory and insufficient for deployment without further refinement.