hi, i am trying to plot my eeg dataset with and without filter.The plots which i got after applying butterworth filter is in an amplified form.i am attchng image (with filter) and code for plotting eeg(without filter).pls help me

%without filter

clc;

clear all;

Fs=256;

cd('E:\Thesis EEG\MIT_EEG_Database\P1');

ea=dir;

for i=3:5

ea(i).name

s=load(ea(i).name);

x=[s.val];

[r,c]=size(x);

for j=1:2

window=x(1,(j-1)*256+1:j*256);

figure;

plot(window);

end

end

Answer 1

[Kris Fedorenko]

Hi Nirmal!

I do not have a lot of domain knowledge in this area, but based on the documentation page for the "butter" function , using the [z,p,k] syntax might be more numerically stable than using [a,b] syntax.

You should be able to use this syntax as follows:

[z,p,k] = butter(n, Wn);
[b,a] = zp2tf(z,p,k);
Y = filter(b1, a1, x');

I have reproduced your code and tried it on a few EEG data files from this database .

Using [z,p,k] syntax for the "butter" function results in filtered data of similar magnitudes as the original data, while using [a,b] syntax made filter data appear amplified. So I hope this helps!

Kris

[Star Strider]

Actually, converting to second-order-section representation is preferable:

[z,p,k] = butter(n, Wn);
[sos,g] = zp2sos(z,p,k);
Y = filtfilt(sos, g, x');