Abstract
This chapter deals with the application of ICA to biomedical data. In biomedical recordings multiple sensors are used to record some physiological phenomena. Often these sensors are located close to each other, so that they simultaneously record signals that are highly correlated with each other. For example, in electroencephalographic (EEG) recordings the sensors are placed at the scalp within a few centimeters of each other. Therefore, the sensors not only record brain activity transmitted by volume conduction from a few dynamic neocortical processes but also artifactual signals, such as noise independent of brain processes, that overlap with neural brain activity that and may be present in all sensors. In this case, a useful tool for EEG researchers would be a method for segregating neural brain activity from artifactual noise signals, or even more interesting, a method that can segregate overlapping neural activities into independent components. Given some major assumptions about EEG signals: (1) that they sum approximately linearly, and (2) are temporally independent, ICA may be an appropriate tool to blindly separate overlapping EEG signals and artifacts into independent components.
Keywords
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.
I am not built for academic writings. Action is my domain.
Gandhi
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
Notes
MRI is a method to obtain images of various parts of the body without the use of x-rays. In contrast to x-rays and CAT scans, a MRI scanner consists of a large and very strong magnet in which the patient lies. A radio wave antenna is used to send signals to the body and then receive signals back. Given the received signals which are changing magnetic fields that are much weaker than the steady strong magnetic field of the main magnet, an image of the body can be computed at almost any particular angle.
Functional MRI refers to the use of MRI scans with a specific task such as psychomotor tasks to observe brain activity during performance. fMRI detects subtle increases in blood flow associated with activation of parts of the brain and may be useful for preoperative neurosurgical planning, epilepsy evaluation, and “mapping” of the brain.
The basis of the BOLD technique lies on the fact that MRI images can be made sensitive to local oxygen concentrations in tissue. This effect has been applied almost exclusively in the human brain to map cortical regions responsible for performing various cognitive tasks, since the oxygenation level in active cortex changes between baseline and tasking conditions.
Since the task is an alternating visual task the reference function is expected to exhibit a square wave type signal
One may ask why ICA results in correlated data. The answer is that a spatial ICA was performed, assuming independent image maps in contrast to temporal ICA giving temporally independent components.
Author information
Authors and Affiliations
Rights and permissions
Copyright information
© 1998 Springer Science+Business Media Dordrecht
About this chapter
Cite this chapter
Lee, TW. (1998). Biomedical Applications of ICA. In: Independent Component Analysis. Springer, Boston, MA. https://doi.org/10.1007/978-1-4757-2851-4_7
Download citation
DOI: https://doi.org/10.1007/978-1-4757-2851-4_7
Publisher Name: Springer, Boston, MA
Print ISBN: 978-1-4419-5056-7
Online ISBN: 978-1-4757-2851-4
eBook Packages: Springer Book Archive