<p>Hierarchical heterogeneity is one of the fundamental characteristics of brain networks, but its impact on whole-brain dynamics following structural lesion remains incompletely understood. We first construct homogeneous and heterogeneous whole-brain models based on structural connectivity data. Parameter scanning reveals that, under optimal parameters, the heterogeneous model outperforms the homogeneous model in fitting empirical functional connectivity, providing preliminary validation of its validity. Furthermore, we establish two types of lesion models by randomly removing edges and nodes. Regardless of lesion type, introducing heterogeneity reduces the average loss of synchronization and metastability. The edge deletion model further reveals unique regulatory patterns. Random edge deletion leads to a significant decrease in synchronization and metastability levels, whereas deleting edges across different T1w/T2w myelin-estimate categories induces directional changes in synchronization and metastability. These results highlight the unique regulatory role of heterogeneity in network lesion responses. Node lesion simulations show that node degree is the primary factor predicting the decline in synchronization, with higher-degree node lesions leading to more pronounced decreases in synchrony. In contrast, changes in metastability show no direct correlation with simple structural metrics such as node degree. After controlling for node degree, the partial Spearman correlation coefficient between the clustering coefficient and the magnitude of metastability reduction is significantly negative, suggesting that the residual clustering coefficient may partially predict changes in metastability. In summary, this study provides a theoretical foundation for understanding the role of hierarchical heterogeneity in brain network responses to lesions.</p>