PDF(Pubmed)
Abstract：
UNASSIGNED: To understand brain function in natural real-world settings, it is crucial to acquire brain activity data in noisy environments with diverse artifacts. Electroencephalography (EEG), while susceptible to environmental and physiological artifacts, can be cleaned using advanced signal processing techniques like Artifact Subspace Reconstruction (ASR) and Independent Component Analysis (ICA). This study aims to demonstrate that ASR and ICA can effectively extract brain activity from the substantial artifacts occurring while skateboarding on a half-pipe ramp.
UNASSIGNED: A dual-task paradigm was used, where subjects were presented with auditory stimuli during skateboarding and rest conditions. The effectiveness of ASR and ICA in cleaning artifacts was evaluated using a support vector machine to classify the presence or absence of a sound stimulus in single-trial EEG data. The study evaluated the effectiveness of ASR and ICA in artifact cleaning using five different pipelines: (1) Minimal cleaning (bandpass filtering), (2) ASR only, (3) ICA only, (4) ICA followed by ASR (ICAASR), and (5) ASR preceding ICA (ASRICA). Three skateboarders participated in the experiment.
UNASSIGNED: Results showed that all ICA-containing pipelines, especially ASRICA (69%, 68%, 63%), outperformed minimal cleaning (55%, 52%, 50%) in single-trial classification during skateboarding. The ASRICA pipeline performed significantly better than other pipelines containing ICA for two of the three subjects, with no other pipeline performing better than ASRICA. The superior performance of ASRICA likely results from ASR removing non-stationary artifacts, enhancing ICA decomposition. Evidenced by ASRICA identifying more brain components via ICLabel than ICA alone or ICAASR for all subjects. For the rest condition, with fewer artifacts, the ASRICA pipeline (71%, 82%, 75%) showed slight improvement over minimal cleaning (73%, 70%, 72%), performing significantly better for two subjects.
UNASSIGNED: This study demonstrates that ASRICA can effectively clean artifacts to extract single-trial brain activity during skateboarding. These findings affirm the feasibility of recording brain activity during physically demanding tasks involving substantial body movement, laying the groundwork for future research into the neural processes governing complex and coordinated body movements.
UNASSIGNED: A dual-task paradigm was used, where subjects were presented with auditory stimuli during skateboarding and rest conditions. The effectiveness of ASR and ICA in cleaning artifacts was evaluated using a support vector machine to classify the presence or absence of a sound stimulus in single-trial EEG data. The study evaluated the effectiveness of ASR and ICA in artifact cleaning using five different pipelines: (1) Minimal cleaning (bandpass filtering), (2) ASR only, (3) ICA only, (4) ICA followed by ASR (ICAASR), and (5) ASR preceding ICA (ASRICA). Three skateboarders participated in the experiment.
UNASSIGNED: Results showed that all ICA-containing pipelines, especially ASRICA (69%, 68%, 63%), outperformed minimal cleaning (55%, 52%, 50%) in single-trial classification during skateboarding. The ASRICA pipeline performed significantly better than other pipelines containing ICA for two of the three subjects, with no other pipeline performing better than ASRICA. The superior performance of ASRICA likely results from ASR removing non-stationary artifacts, enhancing ICA decomposition. Evidenced by ASRICA identifying more brain components via ICLabel than ICA alone or ICAASR for all subjects. For the rest condition, with fewer artifacts, the ASRICA pipeline (71%, 82%, 75%) showed slight improvement over minimal cleaning (73%, 70%, 72%), performing significantly better for two subjects.
UNASSIGNED: This study demonstrates that ASRICA can effectively clean artifacts to extract single-trial brain activity during skateboarding. These findings affirm the feasibility of recording brain activity during physically demanding tasks involving substantial body movement, laying the groundwork for future research into the neural processes governing complex and coordinated body movements.
摘要：
■要了解自然现实世界环境中的大脑功能，在具有不同伪影的嘈杂环境中获取大脑活动数据至关重要。脑电图(EEG),虽然容易受到环境和生理伪影的影响，可以使用先进的信号处理技术，如工件子空间重建（ASR）和独立分量分析（ICA）。这项研究旨在证明ASR和ICA可以有效地从在半管坡道上滑板时发生的大量伪影中提取大脑活动。
■使用了双任务范例，在滑板和休息条件下，受试者受到听觉刺激。使用支持向量机对单次试验EEG数据中是否存在声音刺激进行分类，评估了ASR和ICA在清洁伪影中的有效性。该研究使用五种不同的管道评估了ASR和ICA在工件清洁中的有效性：（1）最小清洁（带通滤波），(2)仅限ASR,（3）仅ICA，(4)ICA，其次是ASR(ICAASR)，和（5）在ICA（ASRICA）之前的ASR。三个滑板运动员参加了实验。
■结果表明，所有包含ICA的管道，尤其是ASRICA(69%，68%,63%），优于最低限度的清洁(55%，52%,50%)在滑板过程中进行单次试验分类。对于三个主题中的两个，ASRICA管道的性能明显优于包含ICA的其他管道，没有其他管道比ASRICA性能更好。ASRICA的卓越性能可能来自ASR去除非平稳伪影，增强ICA分解。ASRICA证明，通过ICLabel识别出所有受试者的大脑成分比单独的ICA或ICAASR更多。对于休息条件，更少的文物，ASRICA管道(71%，82%,75%)显示出与最小清洁相比略有改善(73%，70%,72%），两个受试者的表现明显更好。
■这项研究表明，ASRICA可以有效地清除伪影，以提取滑板过程中的单次试验大脑活动。这些发现肯定了在涉及大量身体运动的体力任务中记录大脑活动的可行性，为未来研究控制复杂和协调身体运动的神经过程奠定基础。
■使用了双任务范例，在滑板和休息条件下，受试者受到听觉刺激。使用支持向量机对单次试验EEG数据中是否存在声音刺激进行分类，评估了ASR和ICA在清洁伪影中的有效性。该研究使用五种不同的管道评估了ASR和ICA在工件清洁中的有效性：（1）最小清洁（带通滤波），(2)仅限ASR,（3）仅ICA，(4)ICA，其次是ASR(ICAASR)，和（5）在ICA（ASRICA）之前的ASR。三个滑板运动员参加了实验。
■结果表明，所有包含ICA的管道，尤其是ASRICA(69%，68%,63%），优于最低限度的清洁(55%，52%,50%)在滑板过程中进行单次试验分类。对于三个主题中的两个，ASRICA管道的性能明显优于包含ICA的其他管道，没有其他管道比ASRICA性能更好。ASRICA的卓越性能可能来自ASR去除非平稳伪影，增强ICA分解。ASRICA证明，通过ICLabel识别出所有受试者的大脑成分比单独的ICA或ICAASR更多。对于休息条件，更少的文物，ASRICA管道(71%，82%,75%)显示出与最小清洁相比略有改善(73%，70%,72%），两个受试者的表现明显更好。
■这项研究表明，ASRICA可以有效地清除伪影，以提取滑板过程中的单次试验大脑活动。这些发现肯定了在涉及大量身体运动的体力任务中记录大脑活动的可行性，为未来研究控制复杂和协调身体运动的神经过程奠定基础。
