Berlin Institute of Technology, Machine Learning Laboratory, Germany.
IEEE Trans Neural Syst Rehabil Eng. 2012 Sep;20(5):653-62. doi: 10.1109/TNSRE.2012.2205707. Epub 2012 Jul 11.
The influence of pre-stimulus ongoing brain activity on post-stimulus task performance has recently been analyzed in several studies. While pre-stimulus activity in the parieto-occipital area has been exhaustively investigated with congruent results, less is known about the sensorimotor areas, for which studies reported inconsistent findings. In this work, the topic is addressed in a brain-computer interface (BCI) setting based on modulations of sensorimotor rhythms (SMR). The goal is to assess whether and how pre-stimulus SMR activity influences the successive task execution quality and consequently the classification performance. Grand average data of 23 participants performing right and left hand motor imagery were analyzed. Trials were separated into two groups depending on the SMR amplitude in the 1000 ms interval preceding the cue, and classification by common spatial patterns (CSPs) preprocessing and linear discriminant analysis (LDA) was carried out in the post-stimulus time interval, i.e., during the task execution. The correlation between trial group and classification performance was assessed by an analysis of variance. As a result of this analysis, trials with higher SMR amplitude in the 1000 ms interval preceding the cue yielded significantly better classification performance than trials with lower amplitude. A further investigation of brain activity patterns revealed that this increase in accuracy is mainly due to the persistence of a higher SMR amplitude over the ipsilateral hemisphere. Our findings support the idea that exploiting information about the ongoing SMR might be the key to boosting performance in future SMR-BCI experiments and motor related tasks in general.
近期已有多项研究分析了刺激前大脑活动对刺激后任务表现的影响。虽然与一致性结果相关的顶枕区的刺激前活动已被详尽研究，但对于感觉运动区的了解则较少，关于后者的研究报告结果不一致。在这项基于感觉运动节律（SMR）调制的脑机接口（BCI）研究中，探讨了这个主题。目标是评估刺激前 SMR 活动是否以及如何影响后续任务执行质量，进而影响分类性能。对 23 名参与者执行右手和左手运动想象的平均数据进行了分析。根据提示前 1000 毫秒 SMR 幅度，将试验分为两组，并在刺激后时间间隔（即任务执行期间）通过共空间模式（CSP）预处理和线性判别分析（LDA）进行分类。通过方差分析评估了试验组和分类性能之间的相关性。分析结果表明，提示前 1000 毫秒 SMR 幅度较高的试验比幅度较低的试验具有显著更好的分类性能。对脑活动模式的进一步研究表明，这种准确性的提高主要归因于对侧半球上较高的 SMR 幅度的持续存在。我们的研究结果支持了这样一种观点，即利用有关持续 SMR 的信息可能是提高未来 SMR-BCI 实验和一般运动相关任务性能的关键。
