A review of multimodal neuroimaging fusion methods and their clinical applications to brain diseases

Abstract

With the rapid development of neuroimaging technology and related data processing methods, multimodal neuroimaging has been widely used in research fields such as neuroscience and clinical diseases. In this paper, the current development of multimodal neuroimaging fusion algorithm and its application in the diagnosis and treatment of brain diseases were reviewed. The definitions, applications, and advantages of the three levels of multimodal neuroimaging fusion, i.e. early fusion, late fusion, and intermediate fusion, were introduced and analyzed. The commonly used multi-modal neuroimaging algorithm basing on signal source separation method and deep multi-modal learning was introduced. The application of multimodal neuroimaging technology in the diagnosis and treatment of severe brain diseases such as schizophrenia and Alzheimer's disease was further discussed. Finally, the existing challenges and future research directions of multimodal neuroimaging methods and applications were summarized. Key words: Multimodal fusion; Neuroimaging; Magnetic resonance imaging; Deep multimodal learning; Neurological diseases