Magnetoencephalography (MEG) is a noninvasive neuroimaging tool that offers a combination of excellent temporal and good spatial resolution, provided that the acquired signals have a high-enough signal-to-noise ratio. This requirement is often compromised as MEG signals are very weak and often masked by interfering signals from environmental noise sources present at most MEG sites. Even more challenging interference is encountered if the subject carries any magnetic material attached to the body, which is sometimes inevitable in clinical settings, e.g., due to therapeutic stimulators. Therefore, to enable reliable data analysis, it is very important to reduce the contribution of noise in MEG signals as efficiently as possible. In this chapter, we review the basic characteristics of MEG signals, give a short review on traditional approaches to suppress noise, and describe some examples of modern noise reduction methods. Specifically, we emphasize the usefulness of advanced mathematical algorithms applied on the multichannel MEG data.
|Alaotsikko||From Signals to Dynamic Cortical Networks|
|DOI - pysyväislinkit|
|Tila||Julkaistu - 17 lokakuuta 2019|
|OKM-julkaisutyyppi||A3 Kirjan osa tai toinen tutkimuskirja|