On the feasibility of using motor imagery EEG-based brain-computer interface in chronic tetraplegics for assistive robotic arm control: a clinical test and long-term post-trial follow-up
On the feasibility of using motor imagery EEG-based brain-computer interface in chronic tetraplegics for assistive robotic arm control: a clinical test and long-term post-trial follow-up
复制标题
DOI：
10.1038/sc.2012.14
复制
发表时间：
2012-08-01
期刊：
影响因子：
2.2
通讯作者：
Popescu, F.
中科院分区：
文献类型：
--
作者：
Onose, G.;Grozea, C.;Popescu, F.
研究设计：对九名慢性脊髓损伤后 (SCI) 四肢瘫痪患者进行调查和长期临床试验后随访（访谈/通讯）。目的：评估使用基于脑电图的脑机接口 (EEG-BCI) 通过机械臂帮助四肢瘫痪患者伸手/抓握的可行性。设置：物理和（神经肌肉）康复医学、心脏病学、神经外科TEHBA 和 UMPCD 的临床部门与“Brain2Robot”（由欧盟委员会资助的同名玛丽居里卓越团队组成，由弗劳恩霍夫研究所-FIRST 主办）合作，于 2008 年下半年进行。 方法：入组患者接受了 EEG-BCI 初步培训和机器人控制课程。统计需要多重线性回归和聚类分析。后续定制调查问卷包括患者对其 EEG-BCI 控制能力的认知，持续至初始实验后 14 个月。结果：EEG-BCI 性能/校准阶段分类准确度平均为 81.0%；对于完成至少一次反馈培训课程的 7 名受试者，反馈培训课程的平均准确度为 70.5%； 9 名受试者中有 7 名 (77.7%) 表示感觉到对光标的控制； 3 名（33.3%）受试者认为他们也在通过运动想象来控制机器人。没有发生明显的副作用。 BCI 性能与 beta (13-30 Hz) EEG 频谱功率密度呈正相关（系数 0.432，标准化系数 0.745，P 值=0.025）；另一个可能的影响是感觉AIS评分（范围：0分钟到224最大，系数-0.177，标准化系数-0.512，P = 0.089）。结论：发现EEG-BCI驱动的机电一体化设备对慢性四肢瘫痪患者进行自我救助的潜力有限但确实存在，这主要与β范围内的谱密度正相关（随之增加）和AIS感觉评分负相关。脊髓 (2012) 50, 599-608; doi：10.1038/sc.2012.14； 2012 年 3 月 13 日在线发布
Study design: Survey and long-term clinical post-trial follow-up (interviews/correspondence) on nine chronic, post spinal cord injury (SCI) tetraplegics.Objective: To assess feasibility of the use of Electroencephalography-based Brain-Computer Interface (EEG-BCI) for reaching/grasping assistance in tetraplegics, through a robotic arm.Settings: Physical and (neuromuscular) Rehabilitation Medicine, Cardiology, Neurosurgery Clinic Divisions of TEHBA and UMPCD, in collaboration with 'Brain2Robot' (composed of the European Commission-funded Marie Curie Excellence Team by the same name, hosted by Fraunhofer Institute-FIRST), in the second part of 2008.Methods: Enrolled patients underwent EEG-BCI preliminary training and robot control sessions. Statistics entailed multiple linear regressions and cluster analysis. A follow-up-custom questionnaire based-including patients' perception of their EEG-BCI control capacity was continued up to 14 months after initial experiments.Results: EEG-BCI performance/calibration-phase classification accuracy averaged 81.0%; feedback training sessions averaged 70.5% accuracy for 7 subjects who completed at least one feedback training session; 7 (77.7%) of 9 subjects reported having felt control of the cursor; and 3 (33.3%) subjects felt that they were also controlling the robot through their movement imagination. No significant side effects occurred. BCI performance was positively correlated with beta (13-30 Hz) EEG spectral power density (coefficient 0.432, standardized coefficient 0.745, P-value=0.025); another possible influence was sensory AIS score (range: 0 min to 224 max, coefficient -0.177, standardized coefficient -0.512, P=0.089).Conclusion: Limited but real potential for self-assistance in chronic tetraplegics by EEG-BCI-actuated mechatronic devices was found, which was mainly related to spectral density in the beta range positively (increasing therewith) and to AIS sensory score negatively. Spinal Cord (2012) 50, 599-608; doi:10.1038/sc.2012.14; published online 13 March 2012