How do I create a matrix that is only the frequencies within the COI of a wavelet transform coherence?

I am measuring the wavelet transform coherence of two signals (x and y). I then am trying to take the average coherence within various frequency ranges. However, I only want to measure the coherence values within the cone of influence COI. How do I do this?

Currently I have the two signals and then conduct the wtc. The signals are attached as files

% define parameterts and signals
ts = 0.1; % time step
fs = 10; % frequency in Hz
t = (0:length(x) - 1)/10; % time vector

[wtc, f, coi] = wcoherence(x, y,fs);

Next, I assume I would create the matrix that is the coherence values within the cone of influence.

coiDatamatrix = ?? % a matrix of only the wtc values within the coi

Finaly, I would take the average coherence within the frequency range 0.125-0.3 Hz

 freqRange = [0.125 0.3];

%     % Find indices corresponding to the frequency range
    freqIndices = find(f >= freqRange(1) & f <= freqRange(2));

    % Extract coherence values in the frequency range within the COI
    coherenceInRange = coiDatamatrix(freqIndices)

    % Calculate average coherence (ignoring NaNs)
    averageCoherence = mean(coherenceInRange);

[Mathieu NOE]

double post

How do I calculate the coherence within a specific frequency range? - MATLAB Answers - MATLAB Central (mathworks.com)

[FsC]

Thank you for responding to my question, I am still having this issue and am not sure how to address it, perhaps I worded the original question incorrectly.

I fixed the code as you advised in the other post and can now extract the frequency array. But, how would I extract the coherence that is within the cone of influence. I am trying to create a mask that is the same size as wtc where all values outside the coi is 0 and all within are 1. How would I do this?

Thank you for your help!

[Mathieu NOE]

hello

I don't have the Wavelet Toolbox with the wcoherence function

I could further help you if you share the output of your wcoherence computations as a mat file

[FsC]

Yes, of course. I've attached the coherence matric (wtc), coi array, and frequency array (f). The original signal is 2 minutes long and sampled at 10 Hz. Attached is also an image of the graphical output for this data.

Thank you for your help!

[Mathieu NOE]

ok I think I got it to work now , see in the answer section below

Answer 1

[Mathieu NOE]

so now the final code

the idea is to create a mask so that wtc values that lies below the coi curve are zeroed

% specific code (to remove)
load('f.mat')
load('coi.mat')
load('wtc.mat')
n = numel(coi);
t = (0:n - 1)*ts; % time vector

% define parameterts and signals
fs = 10; % frequency in Hz
ts = 1/fs; % time step
% t = (0:length(x) - 1)*ts; % time vector
% [wtc, wcs, f, coi] = wcoherence(x, y,fs); 

% create a mask (coi) so data below coi curve are zeroed
[m,n] = size(wtc);
mask = ones(m,n);
tmp = mask;
ind = round(m*coi/max(f));
for ci =1:n
     mask(1:ind(ci),ci) = 0; 
end
mask = flipud(mask); 

% apply Coi mask 
wtc = wtc.*mask;

% plot wtc within coi bounds
imagesc(t,f,abs(wtc));
set(gca,'YDir','normal');
title('X1');
xlabel('Time (samples)')
ylabel('Frequency (Hz)')
colormap('jet')
colorbar('vert')
hold on

plot(t,coi,'--w','linewidth',3);

% Finaly, I would take the average coherence within the frequency range 0.125-0.3 Hz
 freqRange = [0.125 0.3];
%     % Find indices corresponding to the frequency range
    freqIndices = find(f >= freqRange(1) & f <= freqRange(2));
    % Extract coherence values in the frequency range within the COI
    coherenceInRange = abs(wtc(freqIndices,:));
    % Calculate average coherence (ignoring NaNs)
    averageCoherence = mean(coherenceInRange,'all','omitnan');

on your side you simply have to remove these lines

% specific code (to remove)
load('f.mat')
load('coi.mat')
load('wtc.mat')
n = numel(coi);
t = (0:n - 1)*ts; % time vector

and uncomment this one

% t = (0:length(x) - 1)*ts; % time vector
% [wtc, wcs, f, coi] = wcoherence(x, y,fs); 

hope it helps !

[FsC]

This is perfect! Thank you so much for your help and knowledge!

[Mathieu NOE]

My pleasure !

one last thing : you can remove that line

tmp = mask;

all the best