Automated Inline Normalization Procedure for BOLD-Cerebrovascular Reactivity Using the Resting-State Temporal Shift with Machine Learning

作者：Yihui Zhu, Siddhant Dogra, Xiuyuan Wang, Jon̈athan R. Polimeni, Seena Dehkharghani · 发表于：American Journal of Neuroradiology · 年份：2026 · DOI：10.3174/ajnr.a9267 · 被引用次数：1 · 研究领域：Seismic Imaging and Inversion Techniques、Speech Recognition and Synthesis、Model Reduction and Neural Networks

BACKGROUND AND PURPOSE: Cerebrovascular reactivity (CVR) is commonly used to estimate hemodynamic impairment. Conventional use is best-suited to unilateral vascular disease, such that CVR can be normalized to reference values from the contralateral hemisphere or to the posterior circulation territories; however, major confounds have been identified that leave implementation difficult in more common cases of bilateral disease, even despite common cerebellar normalization. Recently, we reported data-driven identification of candidate healthy voxel signatures learned from contemporaneous imaging data. Here, we introduce an entirely inline, automated approach exploiting the dynamics of resting-state blood oxygenation level-dependent (BOLD) functional MR imaging (rs-BOLD) signal from the BOLD baseline, hypothesizing prediction to within 10% error relative to ground truth healthy-voxel CVR values. MATERIALS AND METHODS: Twenty-two subjects with strictly unilateral intracranial steno-occlusive disease (SOD) underwent 28 CVR studies under pharmacologic provocation using acetazolamide with BOLD-MRI (ACZ-BOLD). Separate affected and unaffected hemispheric masks were segmented to train machine learning models to learn signatures of the unaffected hemisphere using the rs-BOLD baseline, as well as anatomic and vascular parameters. Twenty additional healthy subjects from the Human Connectome Project supplemented training, wherein all voxels were classified as normal. Thirty-two distinct ti...