Visual expertise is developed through extensive practice [1], which enables individuals to classify objects with greater specificity and abstraction [2]. In the case of radiological expertise, radiologists can effectively detect abnormalities and identify deviations that may indicate pathological conditions, with the ability to recognize lesions serving as its foundation [[3], [4], [5]]. This expertise intertwines visual behavior with higher-level cognitive processes such as decision-making, memory retrieval, and attention [6,7]. Neural correlates of visual expertise involve multiple brain systems, leading to widespread neural activation beyond the visual cortex [8].

Numerous studies have explored the neural correlates of radiological expertise. While each study contributes valuable insights, a clearer understanding emerges when these findings are synthesized. Consistently, the bilateral middle temporal gyrus (MTG), middle frontal gyrus (MFG), left superior frontal gyrus (SFG) and lateral occipital cortex (LO) have been identified as key regions involved in radiological expertise. For instance, Haller reported selective activations in the MTG, MFG, and SFG when comparing radiologists to laypersons, highlighting the role of these regions in visual processing and cognitive control [3]. Additionally, Bilalić found that posterior MTG and MFG distinguish radiologists from medical students, with posterior MTG playing a key role in knowledge retrieval and application, which are crucial for effective clinical decision-making [9]. Wang further suggested that functional connectivity alterations between the SFG and MFG reflect how cognitive functions, such as attention and semantic knowledge, interact with visual processing to enhance radiological expertise [10]. LO plays a role in the adaptation to new diagnostic tasks [5]. In summary, the brain regions mentioned above act their central role in the acquisition of radiological expertise, particularly in the domains of visual processing, attention, and memory retrieval. However, despite these findings, a gap remains in translating this knowledge into practical applications for radiological education. This gap stems from the lack of clarity regarding the mechanisms underlying these neural changes, particularly in the context of short-term training and brain reorganization. To address this, further research utilizing longitudinal designs is essential to explore the dynamic neural changes during the acquisition of radiological expertise, bridging the gap between theory and practice in radiological education.

Resting-state functional magnetic resonance imaging (rs-fMRI) captures the brain's intrinsic activity in the absence of external stimuli [11]. This state reveals functional adaptations linked to expert performance and shows how training modifies brain function without external stimuli [12]. Degree centrality (DC) measures a node's connectivity within the brain's network, with higher values indicating more central nodes [13]. It quantifies the connectivity between a specified node and other brain regions in resting-state data, providing insights into its role in information transfer [14]. We chose DC because it is sensitive to changes in brain network connectivity, making it well-suited for detecting training-induced changes in neural representations related to radiological expertise [[14], [15], [16]]. Unlike other measures, DC is highly reproducible, does not require predefined regions of interest, and highlights the centrality and integration of key brain areas involved in cognitive processing and expertise development [17]. Support vector machines (SVM) classify subjects pre- and post-training by finding a hyperplane [18,19]. SVM effectively uncovers complex patterns in high-dimensional fMRI data, making it ideal for identifying biomarkers of training-induced neural mechanisms [18,20,21]. However, little attention has been given to DC in research exploring the neural mechanisms of radiological expertise.

This study innovatively focuses on the impact of short-term radiological interpretation training on the resting-state brain, specifically examining the changes in DC using rs-fMRI. The research targeted radiologists who underwent four weeks of intensive rotation in the X-ray department. To capture subtle shifts in the importance of different brain regions associated with radiological expertise, we employed the recursive feature elimination-support vector machine (RFE-SVM) method. We hope this study will help understand rapid brain remodeling induced by short-term learning, revealing how the brain adapts to new skill acquisition and providing insights into the accumulation process of long-term learning effects.