Morphological changes in the cerebellum during aging: evidence from convolutional neural networks and shape analysis

作者：Yu Wang, Teng Ye, Tianci Liu, Yuchun Tang, Wenjia Liang, Wenjun Wang, Zhuoran Li, Qing Xia, Feifei Xu, Shuwei Liu · 发表于：Frontiers in Aging Neuroscience · 年份：2024 · DOI：10.3389/fnagi.2024.1359320 · 被引用次数：16 · 研究领域：Vestibular and auditory disorders、Advanced Neuroimaging Techniques and Applications、Morphological variations and asymmetry

The morphology and function of the cerebellum are associated with various developmental disorders and healthy aging. Changes in cerebellar morphology during the aging process have been extensively investigated, with most studies focusing on changes in cerebellar regional volume. The volumetric method has been used to quantitatively demonstrate the decrease in the cerebellar volume with age, but it has certain limitations in visually presenting the morphological changes of cerebellar atrophy from a three-dimensional perspective. Thus, we comprehensively described cerebellar morphological changes during aging through volume measurements of subregions and shape analysis. This study included 553 healthy participants aged 20-80 years. A novel cerebellar localized segmentation algorithm based on convolutional neural networks was utilized to analyze the volume of subregions, followed by shape analysis for localized atrophy assessment based on the cerebellar thickness. The results indicated that out of the 28 subregions in the absolute volume of the cerebellum, 15 exhibited significant aging trends, and 16 exhibited significant sex differences. Regarding the analysis of relative volume, only 11 out of the 28 subregions of the cerebellum exhibited significant aging trends, and 4 exhibited significant sex differences. The results of the shape analysis revealed region-specific atrophy of the cerebellum with increasing age. Regions displaying more significant atrophy were predominantly l...