Neuroimaging for dementia has made remarkable progress in recent years, shedding light on diagnostic subtypes of dementia, predicting prognosis and monitoring pathology. This review covers some updates in the understanding of dementia using structural imaging, positron emission tomography (PET), structural and functional connectivity, and using big data and artificial intelligence. Progress with neuroimaging methods allows neuropathology to be examined in vivo, providing a suite of biomarkers for understanding neurodegeneration and for application in clinical trials. In addition, we highlight quantitative susceptibility imaging as an exciting new technique that may prove to be a sensitive biomarker for a range of neurodegenerative diseases. There are challenges in translating novel imaging techniques to clinical practice, particularly in developing standard methodologies and overcoming regulatory issues. It is likely that clinicians will need to lead the way if these obstacles are to be overcome. Continued efforts applying neuroimaging to understand mechanisms of neurodegeneration and translating them to clinical practice will complete a revolution in neuroimaging.

近年来，痴呆症的神经影像学取得了显着进展，阐明了痴呆症的诊断亚型，预测预后和监测病理。这篇综述涵盖了使用结构成像，正电子发射断层扫描（PET），结构和功能连接以及使用大数据和人工智能对痴呆症的理解的一些最新进展。神经影像学方法的进步使人们可以在体内对神经病理学进行检查，从而提供了一套生物标记物，用于了解神经退行性病变以及在临床试验中的应用。此外，我们重点介绍了定量磁化率成像作为一种令人兴奋的新技术，它可能被证明是一系列神经退行性疾病的敏感生物标记。将新颖的成像技术转化为临床实践面临挑战，特别是在开发标准方法和克服法规问题方面。如果要克服这些障碍，临床医生可能需要带头。继续努力应用神经影像学来了解神经退行性机制并将其转化为临床实践将完成神经影像学的一场革命。