Processing of weak magnetic multichannel signals : the signal space separation method

Thumbnail Image
Journal Title
Journal ISSN
Volume Title
Doctoral thesis (article-based)
Checking the digitized thesis and permission for publishing
Instructions for the author
Degree programme
Verkkokirja (3986 KB, 67 s.)
This work concentrates on processing of multichannel magnetoencephalographic (MEG) data. The aim of the work is to improve the quality of the measured signals in order to enable reliable data analysis. A special requirement for the developed mathematical methods is that they should be applicable to all MEG measurements regardless of the level of cooperation of the subject. This is essential, e.g., with small children and in clinical investigations. In addition to MEG, the methods presented here can be used in other magnetic multichannel measurements, too. MEG measurements are used in basic brain research and recently also in clinical examinations. The method has excellent time resolution and reasonably good spatial resolution, which makes it a very useful tool in analysis of various brain functions. During the last 20 years, the instrumentation of MEG has been developed from devices containing less than 40 channels and limited coverage to whole-head systems with more than 300 channels. Yet, many of the signal processing and analysis methods used today date back to the time of the old instrumentation with limited coverage of the magnetic field. Traditionally, MEG investigations have been performed primarily only with cooperative subjects in order to avoid the characteristic problems of MEG, including signal distortions due to head movements and artifacts caused by sources attached to the body. In clinical measurements, however, the patient may have involuntary movements and carry artifact sources such as therapeutic stimulators. This work introduces the signal space separation method (SSS), which is based on Maxwell's equations and the generous spatial sampling by modern multichannel MEG devices. The thesis describes the theoretical foundations of SSS and its temporal extension tSSS, and demonstrates the results in several applications. SSS and tSSS are shown to significantly improve the quality of MEG data under conditions previously considered too challenging for meaningful analysis. Notably, the methods have potential to expand the applicability of MEG to some new patient groups, e.g., patients with deep brain stimulators.
magnetoencephalography, signal processing, multichannel measurement, clinical applications
Other note
  • [Publication 1]: Samu Taulu and Juha Simola. 2008. Multipole-based coordinate representation of a magnetic multichannel signal and its application in source modelling. Helsinki University of Technology Publications in Engineering Physics, Report TKK-F-A855. © 2008 by authors.
  • [Publication 2]: Samu Taulu and Matti Kajola. 2005. Presentation of electromagnetic multichannel data: The signal space separation method. Journal of Applied Physics, volume 97, 124905, pages 1-10. © 2005 American Institute of Physics. By permission.
  • [Publication 3]: Samu Taulu and Juha Simola. 2006. Spatiotemporal signal space separation method for rejecting nearby interference in MEG measurements. Physics in Medicine and Biology, volume 51, pages 1759-1768. © 2006 Institute of Physics Publishing. By permission.
  • [Publication 4]: Samu Taulu, Juha Simola, and Matti Kajola. 2005. Applications of the signal space separation method. IEEE Transactions on Signal Processing, volume 53, number 9, pages 3359-3372. © 2005 IEEE. By permission.
  • [Publication 5]: Samu Taulu and Riitta Hari. 2008. Removal of magnetoencephalographic artifacts with temporal signal-space separation: Demonstration with single-trial auditory-evoked responses. Human Brain Mapping, in press.