Background: Brainstem tumors are rare and result in enduring behavioral issues, posing challenges for patients and surgeons. Despite similarities in symptoms to those resulting from cerebellar injuries, limited research has investigated cerebellar changes in patients with brainstem tumors. Methods: Here, we employed U-Net-based segmentation to analyze the cerebellum in 26 regions, assessing individual volume deviations. Additionally, we developed a behavior prediction model using the total outlier count (tOC) index and brain volume as features. Findings: Our normative model results revealed that most patients exhibited negative deviation in most cerebellar regions, particularly in lobule Right V, Left I-III, Right X, and Left X. The results found a significant association between a higher tOC and severe social problems (r = 0.31, p = 0.001) and withdrawal behavior (r = 0.28, p = 0.001). Further analysis showed that smaller cerebellar regions correlated strongly with more pronounced social problems (r = 0.27, p = 0.007) and withdrawal behavior (r = 0.25, p = 0.015). Notably, lobules such as Right X, V, IV, and VIIB, along with Left IX, VIII, and X exerted the greatest influence on social problems, whereas regions including Left V, Right IV, and Vermis VI and VIII were most impactful on withdrawal behavior. Interpretation: In summary, our study revealed cerebellar atrophy patterns in patients with brainstem tumors, emphasizing the role of both anterior and posterior cerebellar lobes in social problems and withdrawal behavior. This research sheds light on the brain mechanisms underlying complex behavioral disorders in brainstem tumor patients. Funding: This study was supported by the STI2030-Major Projects 2021ZD0200201 and Beijing Municipal Public Welfare Development and Reform Pilot Project for Medical Research Institutes (grant ID: JYY202X-X).