Impact: Dynamic off-resonance correction reduces physiological noise and preserves spatial detail, improving overall image quality in high-resolution PCASL perfusion imaging at 7T.

Impact: Integrating transcriptomic, receptor, and connectomic frameworks revealed mitochondrial dysfunction, serotonergic–cholinergic modulation, and an inferior frontal network hub underlying cerebral perfusion abnormalities in methamphetamine dependence, offering mechanistic biomarkers and potential targets for cerebrovascular restoration and addiction treatment .

Impact: Subject-specific T2 priors eliminate hidden modeling biases in blood–brain barrier (BBB) water-exchange imaging. Personalized relaxation parameters enable robust, individualized BBB assessment and allow detection of age and disease related vascular dysfunction in neurological disorders.

Impact: Combined PREFUL-MRI and quantitative CT assessment provides clinical value for preoperative risk stratification and individualized surgical planning in lung cancer patients with COPD.

Impact: QTMnet can be coupled into clinical nasopharyngeal carcinoma practice to predict patient progression free survival time after treatment.

Impact: This study novelly characterizes nonuniform ATT increase across arterial territories and identifies their border zones emergence after age 10. The normative ATT patterns may guide age-specific single-PLD pCASL optimization, enhancing rCBF accuracy and detection of perfusion-vulnerable regions in pediatric populations.

Impact: The proposed rapid B₀ shimming technique significantly improves the robustness of pCASL, unlocking the full potential of ASL’s high sensitivity for enhancing perfusion imaging at 7T.

Impact: This work demonstrates that GM/WM ratios from DSC-MRI are robust biomarkers across 3T and 5T, enhancing the reproducibility of perfusion studies in multi-center trials and facilitating the clinical integration of ultra-high-field MRI.

Impact: Our solution benefits clinicians and researchers encountering inflow artefacts in their 3D PCASL implementations. By mitigating these artefacts, the method enhances diagnostic reliability and image interpretability, supporting broader adoption of ASL in both clinical and research settings.

Impact: We propose ART-Net to identify sources of ASL artifacts, enabling actionable, source-aware quality control. This framework can enhance scan reliability, reduce data loss in multi-site studies, and provide a generalizable template for artifact classification across imaging modalities.

Impact: This work represents a major advance towards making rapid, reliable ASL a clinical reality with Physics-informed deep learning. By enabling robust perfusion quantification in one minute, it significantly improves the practicality and utility of ASL for diagnosing neurological disorders.

Impact: Encoding matrix pseudoinversion allows accurate, fast, and straightforward reconstruction of non-Cartesian Arterial Spin Labelling (ASL) data by linear transformation. Incorporating sensitivity encoding allows accelerated acquisitions without substantially increasing reconstruction time or reducing image quality.

Impact: Accelerated TE-resolved ASL with partition-randomized stack-of-spirals acquisition and subspace reconstruction exploits temporal correlation along the TE dimension and yields T2-decay-free multi-PLD multi-TE ASL images, enabling accurate quantification of cerebral perfusion and BBB permeability within clinically feasible acquisition time.

Impact: We present a deep learning based ASL plane prescription framework that improves planning efficiency and reproducibility. Its consistent performance across datasets supports integration into clinical workflows, advancing automation in cerebrovascular imaging without relying on manual expertise.

Impact: A robust temporal decomposition model is a highly effective method for 4D ASL angiogram denoising, overcoming the signal-destructive artifacts and limitations of common spatial and temporal denoising algorithms.

Impact: We present GigaBlochs, a flexible Bloch simulation framework enabling large parameter space investigations. Simulations of pulsatile flow in Pseudo-Continuous Arterial Spin Labelling (PCASL) reveal an opportunity to reduce SAR without sacrificing SNR by decreasing $B_1$ amplitude during diastole.