运动表象
脑-机接口
解码方法
神经解码
脑磁图
计算机科学
脑电图
大脑活动与冥想
节奏
神经影像学
任务(项目管理)
人工智能
神经科学
语音识别
模式识别(心理学)
心理学
物理
管理
经济
电信
声学
作者
Xu Wang,Jianjun Meng,Yu Zheng,Yuxuan Wei,Fan Wang,Han Ding,Yan Zhuo
标识
DOI:10.1088/1741-2552/adf011
摘要
Abstract Objective. The limited spatial resolution inherent in electroencephalography (EEG), a widely-adopted non-invasive neuroimaging technique, combined with the intrinsic complexity of performing unilateral lower-limb motor imagery (MI), restricts decoding accuracy. To address these challenges, we propose a paradigm based on action observation-guided rhythmic motor execution (AO-ME) and motor imagery (AO-MI), designed to simplify task demands and enhance decoding performance. Magnetoencephalography (MEG) serves as the data acquisition method, leveraging its superior spatiotemporal resolution. Approach. Spatiotemporal and spectral features were characterized at the sensor level, and source imaging techniques were employed to examine cortical activation patterns. Ensemble task-related component analysis (eTRCA) facilitated decoding of unilateral tasks. And multiple decoding algorithms were employed to validate the effectiveness of the proposed paradigm. Main results. Robust lateralized neural responses were observed, exhibiting low-frequency phase-locked components that distinctly reflected the task frequency and its second harmonic within sensorimotor, parietal, and occipital cortices. Moreover, significant contralateral suppression of the sensorimotor rhythm was observed. Decoding accuracies reached 95.22 ± 4.75% for AO-ME and 88.66 ± 8.52% for AO-MI across twenty participants based on the phase-locked features using eTRCA. Significance. Collectively, our findings demonstrate that the proposed paradigm provides an effective approach for eliciting robust, distinguishable neural responses, enabling high decoding performance of unilateral lower-limb movements. This work offers new insights into the underexplored domain of lower-limb MI and highlights the paradigm’s potential for brain–computer interface applications.
科研通智能强力驱动
Strongly Powered by AbleSci AI