Document Type
Thesis
Date of Award
9-30-1985
Degree Name
Master of Science in Biomedical Engineering - (M.S.)
Department
Biomedical Engineering Committee
First Advisor
Stanley S. Reisman
Second Advisor
David S. Kristol
Third Advisor
Joseph J. Strano
Abstract
The analysis of the EEG background waveform is studied by two techniques, spectrum analysis and real-time analysis. The spectrum analysis is performed by using Fast Fourier Transformation (FFT) on a segment of the signal. The resulting signal is classified into different band types such as delta, theta, alpha and beta, to see the power in each band, and the magnitude of the power on the left and right side if the brain is evaluated for asymmetry. The mean power frequency shows the frequency where the maxinum power exists in each band type. This type of analysis provides overall information of a segment of signal, but does not give any information about the makeup of the segment.
Real-time analysis is developed to obtain information about smaller segments of the signal. The signal in the time domain is analyzed by monitoring the times that zero crossings occur. The time between zero crossing is converted into a frequency value, and classified into the different band types. consecutive bands are compared, and the time duration of the same type of band is calculated. Autosegmentation occurs when the consecutive band types are different, and the time duration is counted for each segment of the signal, Four second segmants of the signal are analyzed in this way, and a statistical study is done to obtain the dominant frequency in each four seconds of data.
Recommended Citation
Won, Suzanne Hesook, "Spectrum analysis and real-time analysis of electroencephalograms" (1985). Theses. 3488.
https://digitalcommons.njit.edu/theses/3488
