In order to study the task-based functional connectivity of the whole brain in schizophrenia patients and healthy controls during task switching, the NSP(nested-spectral partition) method was used to quantify the degree of integration and segregation of the whole brain network and explain differential remodeling patterns of brain networks in schizophrenia patients and healthy controls. Results show that changes in whole-brain functional connectivity patterns in schizophrenia are driven by lower network integration and higher segregation, suggesting global and network-wide local information transfer during task switching in schizophrenia patients efficiency was lower than that of healthy controls. Furthermore, the association between task switching costs (reaction time and switching costs) and task state modularity was analyzed using simple linear correlation. Lower whole-brain modularity and lower salience network modularity in schizophrenia were found to be associated with larger switch costs but not with mean reaction times across the task. In contrast, higher cortical network modularity in healthy control patients was associated with faster task reaction times but not with shorter switch costs. Taken together, the results suggest that quantifying functional network configurations across different task states can provide information about cognitive control performance.