关键词: 无创运动想象脑机接口, 信号采集, 编码范式, 解码算法, 临床应用

Abstract: Non-invasive motor imagery-based brain–computer interface (MI-BCI) is a key technology that enables direct human-computer interaction by decoding brain motor imagery activity, and has important application potential in fields such as neurorehabilitation and intelligent control. This paper systematically reviews the current state of key technologies used in non-invasive MI-BCI systems, including electrode technology and signal acquisition, motor imagery task encoding paradigms, and publicly available datasets. It focuses on analyzing electroencephalogram (EEG) decoding methods based on conventional machine learning and deep learning, as well as their performance improvement mechanisms. On this basis, it summarizes the key challenges faced by current non-invasive MI-BCI systems, such as signal non-stationarity, low signal-to-noise ratio, and poor cross-subject generalization capability, and reviews their application progress in clinical scenarios including neurorehabilitation. Finally, it provides an outlook on the development trends of key technologies and the clinical application prospects of non-invasive MI-BCI.