readme.txt

eeg segmentation codes...

Author:
Dr. Reda Regab Gharieb

Reference:
R. R. Gharieb and A. Cichocki, 
"Segmentation and tracking of the electo-encephalogram signal using an adaptive 
recursive bandpass filter,"
Medical & Biological Engineering & Computing, IFMBE-2001,
vol. 39, pp. 237-248.

This package contains the following files:

README
  This file
eeg_demo.m
  Matlab m-file
  This Demo program is to simulate different segments of EEG signal and to online track 
  theses segments.
sim_eeg.m
  Matlab m-file
  This is a function using linear modeling for generating an eeg wave specified by the user.
eeg_seg.m
  Matlab m-file
  This makes segmentation of eeg signal based on the center frequency of each segment.

HOW TO RUN

This EEG Demo program is to simulate different segments of EEG signal and to online track
theses segments.

For the simulation, you need Matlab and its Signal Processing Toolbox.

1. Type "eeg_demo" at Matlab command line. You can see demonstration of segmentation of 
simulated EEG signal.

2. Changing some parameters in eeg_demo.m, you can obtain other results. You can also apply 
eeg_seg.m to your own data.

<eeg_demo.m>
This m-file calls next two functions and carrys out demonstration.
This Demo program is to simulate different segments of EEG signal and to online track 
theses segments.

First, it generates an eeg signal that consists of different segments (e.g. alpha wave, 
beta wave ....) using sim_eeg function. In the upper figure, the signal is presented. 
Next, using the matlab specgram function, it represents the spectrogram of the signal 
in the middle figure. Next, it calculates our segmenting function for tracking the
eeg segments and represents the function in the lower figure.

We can track the eeg segments by the spectrogram, but our segmenting function
represents the segmentation more clearly and is easier to use because it is a scalar
function.  

<sim_eeg.m>
This is a function using linear modeling for generating an eeg wave specified by the user.
You can use both MA model and ARMA model. For usage, please use help command in Matlab.

<eeg_seg.m>
This makes segmentation of eeg signal based on the center frequency of each segment.
The function uses an adaptive bandpass filter; the adaptive filter is implemented as a 
4th order Butterworth filter; the adaptive filter needs only one coefficient to be updated. 
The center frequency of the eeg segment is proportional to this adaptive coefficient.
For usage, please use help command in Matlab.