Brain network analysis using morphological intrinsic divergence: differences in small-world properties of morphological similarity networks among healthy young adults with different chronotypes.

内容摘要

An individual's chronotype reflects their intrinsic circadian rhythm preference and is closely associated with cognitive function and mental health. However, the relationship between chronotype and whole-brain morphological structural network organization remains unclear. This study aims to explore differences in the topological organization characteristics of morphometric similarity networks (MSNs) among healthy young adults of different chronotypes from a graph theory perspective. We employed a novel Morphometric INverse Divergence (MIND) method, which is more sensitive to subtle morphological differences, to construct individual-level brain MSNs. This method aggregates morphological metrics (cortical thickness, mean curvature, sulcal depth, surface area, gray matter volume) from all vertices within each cortical region to form a regional multivariate distribution. Subsequently, a k-nearest neighbor density algorithm constructs a pairwise distance matrix, and symmetric Kullback-Leibler divergence between regional multivariate distributions quantifies similarity among cortical regions. Using high-resolution Glasser atlas, medium-resolution Destrieux atlas, and low-resolution Desikan-Killiany atlas, MIND networks were constructed for 68 healthy young individuals with early chronotype (EC) and 68 with late chronotype (LC) patterns. We calculated the area under the curve (AUC) for multiple graph-theoretic metrics, including small-world properties, across varying sparsity levels in weighted networks, followed by intergroup comparisons and correlation analyses. Analysis based on the Destrieux atlas revealed that EC participants exhibited significantly higher AUC of Small-World Properties (AUC-SWP) compared to LC participants (P = 0.0045), and this metric showed a significant negative correlation with ChQ-ME scores (rs = -0.2114, P = 0.0135). When using the Desikan-Killiany atlas and the Glasser atlas, the aforementioned intergroup differences and correlations were not detected (P > 0.05). These findings suggest that an individual's chronotype correlates with the topological organization of brain MSNs. This association was detected specifically when using the medium-resolution Destrieux atlas, while was not found with either the lower-resolution Desikan-Killiany atlas or the higher-resolution Glasser atlas under the conditions of this study. This pattern indicates that chronotype-related brain differences may operate at an optimal spatial scale, where brain parcellation strikes a balance between signal integration and anatomical specificity. The results support a model of distributed, subtle morphological alterations that together form a detectable "weak signal" network. This study presented a novel spatial-scale perspective on the relationship between brain structure and circadian rhythms.