Multimodal non-invasive approaches for early Alzheimer's disease detection: a review of neuroelectrophysiological and neuroimaging techniques.

Alzheimer's disease (AD) is a progressive neurodegenerative disorder characterized primarily by a gradual decline in cognitive function and specific pathological changes in the brain. In recent years, although various neuroelectrophysiological and neuroimaging techniques have greatly advanced the mechanistic study of abnormal brain function in AD, an integrative discussion of these technologies remains fragmented. This paper primarily summarizes and interactively analyzes the research progress of several non-invasive neuroimaging and neuroelectrophysiological techniques-event-related potential (ERP), electroencephalogram (EEG), transcranial magnetic stimulation-electroencephalogram (TMS-EEG), functional near-Infrared spectroscopy (fNIRS), magnetoencephalography (MEG), structural magnetic resonance imaging (structural MRI) and functional magnetic resonance imaging (fMRI)-to depict a panoramic view of AD pathology from a microscopic to a macroscopic scale from a multimodal perspective. It further compares the advantages and limitations of various technologies for detecting early AD biomarkers, emphasizing the synergistic value of multimodal integration in capturing changes in dynamic functional and structural brain networks. Additionally, we explore the potential of these technologies in clinical translation, particularly when combined with machine learning and deep learning approaches, to enhance the accuracy of early diagnosis and the depth of mechanism analysis. Through the above discussion, this review aims to provide new insights for the early identification of AD and advance our understanding of the neural mechanisms underlying AD.