Hi friends, I have this sample program and I want to give EEG input signal for this. Can you please tell me how to give the input and if possible provide me the EEG data also. Because i don't know how to get the data. somewhere I have seen, it is nearly 6-7 GB is there.please help me in this issue. Thank you
function [W, S] = ica(X)
% ICA Perform independent component analysis.
%
% [W, S] = ica(X);
%
% where X = AS and WA = eye(d) and [d,n] = size(X) (d dims, n samples)
% Implements FOBI algorithm.
% JF Cardoso (1989) "Source separation using higher order moments", Proc Intl Conf Acoust Speech Signal Process
[d, n] = size(X);
% Subtract off the mean of each dimension.
X = X - repmat(mean(X,2),1,n);
% Calculate the whitening filter.
[E, D] = eig(cov(X'));
% Whiten the data
X_w = sqrtm(pinv(D))*E'*X;
% Calculate the rotation that aligns with the directions which maximize fourth-order correlations.
% See reference above.
[V,s,u] = svd((repmat(sum(X_w.*X_w,1),d,1).*X_w)*X_w');
% Compute the inverse of A.
W = V * sqrtm(pinv(D)) * E';
% Recover the original sources.
S = W * X;
Depending on what you want to do, you can get the data from many public EEG databases, like Physionet, BNCI-Horizon, and many others. Try to get .MAT files so that would be easier to use them from Matlab. For instance, you can check this one: https://www.kaggle.com/rramele/p300samplingdataset
Once you those files, in general each file correspond to one person or session. You need to be able to get the RAW EEG data from each file, and in order to do that, you need to understand the metadata to check in which field the raw EEG is stored. In the one from kaggle that I pointed out, data.X contains the EEG data.
Once you load any of those .dat file, you just use your function like this
[S,W] = ica(data.X)
S are the projected sources and W is the transform matrix. If your algorithm works right, you should get this S = data.X * W.
I want to remove artifacts in EEG data sir. Please help me in this issue sir.
you cane use eeg lab to remove your signal artifacts.
But If you want to do it with mentioned code, you should know that you're working on eeg channels and if your channel number is n, then the covariance matrix of your eeg data will be n by n.
So don't worry about computional cost or complexity.
But before working with this function I recommend you study about ICA.
this one might be helpful.
https://sapienlabs.org/getting-rid-of-eye-blink-in-the-eeg/
- Badges
- Science topic
- Similar topics
- Computer Science
- Program