Attention Load Modulates Functional Brain Connectivity Reorganization in Children with High Autistic Traits: Evidence from Phase and Amplitude Coupling Metrics.

Abstract

Whether emotional processing difficulties in autism spectrum disorder (ASD) stem from deficits in recognizing emotional stimuli or from limitations in flexibly reorganizing brain networks under increasing cognitive demands remains an open question. To our knowledge, this study is among the first to investigate the effects of graded attentional load (low/moderate/high) on EEG-based functional connectivity and brain network organization during task-irrelevant emotional prosody processing in children with high autistic traits (n = 31), by integrating phase-based (iPLV, wPLI) and amplitude-based (AEC) metrics together with graph-theoretic indices within a unified analytic framework. Connectivity matrices from a sliding-window approach were decomposed into edge-level, distributional, and graph-theoretic features and load-dependent changes in connectivity were assessed with statistical analyzes and classification models. The results suggested that attentional load did not uniformly increase or decrease connectivity but rather produced a multi-layered pattern of reorganization that varied as a function of frequency band, connectivity metric, and network topology. Phase-based connectivity in the beta band was decreased with increasing load, while amplitude envelope coupling (AEC) in the theta and alpha bands was increased under high load. This dual-mechanism pattern was most pronounced under the high-load condition, which yielded the greatest number of nominally significant features (uncorrected p<0.05; none survived FDR correction), the largest theta-AEC effect size (Kendall's W up to 0.22), and the best-performing classification model among those tested (accuracy = 72.6%, F1 = 0.725, AUC = 0.764). In conclusion, this exploratory study suggests that, in children with high autistic traits, increasing cognitive demands may be associated with band-specific network reorganization characterized by attenuated phase-based coordination alongside enhanced amplitude-based coupling; these patterns did not survive correction for multiple comparisons, have not been compared with those of typically developed children, and should be treated as hypothesis-generating rather than confirmatory. If replicated, such findings could eventually inform how environmental stimulation and cognitive load are structured in special education settings.