屏蔽电缆
工件(错误)
计算机科学
脑磁图
失真(音乐)
干扰(通信)
噪音(视频)
补偿(心理学)
脑电图
信号(编程语言)
计算机视觉
人工智能
虚假关系
声学
物理
电信
医学
放大器
机器学习
带宽(计算)
图像(数学)
精神科
心理学
精神分析
频道(广播)
程序设计语言
标识
DOI:10.1097/wnp.0000000000000699
摘要
Summary: Noise sources in magnetoencephalography (MEG) include: (1) interference from outside the shielded room, (2) other people and devices inside the shielded room, (3) physiologic or nonphysiologic sources inside the patient, (4) activity from inside the head that is unrelated to the signal of interest, (5) intrinsic sensor and recording electronics noise, and (6) artifacts from other apparatus used during recording such as evoked response stimulators. There are other factors which corrupt MEG recording and interpretation and should also be considered “artifacts”: (7) inadequate positioning of the patient, (8) changes in the head position during the recording, (9) incorrect co-registration, (10) spurious signals introduced during postprocessing, and (11) errors in fitting. The major means whereby magnetic interference can be reduced or eliminated are by recording inside a magnetically shielded room, using gradiometers that measure differential magnetic fields, real-time active compensation using reference sensors, and postprocessing with advanced spatio-temporal filters. Many of the artifacts that plague MEG are also seen in EEG, so an experienced electroencephalographer will have the advantage of being able to transfer his knowledge about artifacts to MEG. However, many of the procedures and software used during acquisition and analysis may themselves contribute artifact or distortion that must be recognized or prevented. In summary, MEG artifacts are not worse than EEG artifacts, but many are different , and—as with EEG—must be attended to.
科研通智能强力驱动
Strongly Powered by AbleSci AI