Measurement of EEG enabled it to get to know electrophysiological mechanism of brain. Disturbance of consciousness, epilepsia, dementia, cerebrovascular disease and cerebral infarction are included as a useful clinical example of EEG inspection. It is thought that it leads to the early detection and treatment of the above disease by analyzing EEG. However, there is a possibility that artifact may mix in EEG. When the brain wave is measured, the artifact by the heart beat, the blink, muscle noise, line noise, and the eye movement may be contained. The data of the part is rounded down when the characteristic of artifact (the peak of amplitude, frequency component, variance, and gradient) exceed a decided threshold and the recurrent part of other input signals (EOG, EMG, and ECG) is subtracted from measurements, etc. as artifact removal method. And so this study considers the artifact removal that uses Independent Component Analysis (ICA). It doesn’t need other input signals. ICA is a statistical method that is useful for removing artifacts. ICA has various algorithms. FastICA is used in this study. FastICA has the best overall performance in terms of both separation quality and computation times. This study uses and simulates ICA to the measured brain waves. Future work is how to apply the separated brain waves.
[1]
Aapo Hyvärinen,et al.
Independent component analysis in the presence of Gaussian noise by maximizing joint likelihood
,
1998,
Neurocomputing.
[2]
Aapo Hyvärinen,et al.
The Fixed-Point Algorithm and Maximum Likelihood Estimation for Independent Component Analysis
,
1999,
Neural Processing Letters.
[3]
M. Akutagawa,et al.
Systematic Identification for Inert Region Of a Brain from EEG
,
2005,
2005 IEEE Engineering in Medicine and Biology 27th Annual Conference.
[4]
Erkki Oja,et al.
Independent component analysis: algorithms and applications
,
2000,
Neural Networks.
[5]
Aapo Hyvärinen,et al.
Fast and robust fixed-point algorithms for independent component analysis
,
1999,
IEEE Trans. Neural Networks.