Functional differences in key brain networks, including the dorsal attention network (DAN), control network (CN), and default mode network (DMN), have been identified in individuals with obsessive–compulsive disorder (OCD). However, the precise nature of these differences remains unclear. In this study, we further explored these differences and validated previous findings using a novel edge functional connectivity (eFC) approach, which enables a more refined analysis of brain network interaction. By employing this advanced method, we sought to gain deeper insights into FC alterations that may underlie the pathology of OCD. We collected data during movie watching from 44 patients with OCD and 33 healthy controls (HCs). The two-sample t test was used to assess differences in entropy between the DAN, CN, and DMN between groups. The analysis was performed with control for potentially confounding variables to ensure the robustness of the findings. Significant differences in network entropy were found between the OCD and HC groups. Relative to HCs, patients with OCD showed significantly reduced entropy in the DAN and increased entropy in the CN and DMN. The decreased entropy in the DAN and increased entropy in the CN and DMN observed in this study may be related to the core symptoms of OCD, such as attention deficit, impaired cognitive control, and self-referential thinking. These results provide valuable insights into the neurobiological mechanisms of OCD and highlight the potential of network entropy as a biomarker for the disorder. Future research should further explore the relationship between these network changes and the severity of OCD symptoms, as well as assess their implications for the development of treatment strategies.

Automated radiology report generation is essential in clinical practice. However, diagnosing radiological images typically requires physicians 5-10 minutes, resulting in a waste of valuable healthcare resources. Existing studies have not fully leveraged knowledge from historical radiology reports, lacking sufficient and accurate prior information. To address this, we propose a Topic-Keyword Semantic Guidance (TKSG) framework. This framework uses BiomedCLIP to accurately retrieve historical similar cases. Supported by multimodal, TKSG accurately detects topic words (disease classifications) and keywords (common symptoms) in diagnoses. The probabilities of topic terms are aggregated into a topic vector, serving as global information to guide the entire decoding process. Additionally, a semantic-guided attention module is designed to refine local decoding with keyword content, ensuring report accuracy and relevance. Experimental results show that our model achieves excellent performance on both IU X-Ray and MIMIC-CXR datasets. The code is available at https://github.com/SCNU203/TKSG.