SOURCE RECORD / S0191
基于 EEG 的脑机接口实现实时单指级机器人手控制
EEG-based brain-computer interface enables real-time robotic hand control at individual finger level
- 作者或机构
- Ding Y, Udompanyawit C, Zhang Y, He B
- 发布日期
- 2025-06-30
- 来源类型
- peer_reviewed_human_study
- 技术路线
- non-invasive 128-channel EEG motor imagery
- 功能场景
- robotic device control
- 权利状态
- mirror_allowed / CC BY 4.0
证据摘要
证据支持：实时非侵入式单指级机器人控制的可行性。关键数字：两指运动想象为 80.56%，三指运动想象为 60.61%，均经过一次在线训练和同场微调。局限：参与者是健康、有经验且经过两轮筛选的响应者；模型依赖个人和当天数据；输出是类别决策，并非多指同时独立运动轨迹。
关键数字
80.56% two-finger MI; 60.61% three-finger MI after one online session and same-session fine-tuning
局限
Healthy experienced users; two rounds of responder screening; subject-specific and same-day fine-tuning; task uses class decisions rather than independent simultaneous multi-finger trajectories