My calculation for respiratory rate shows the same bpm for my 3 waveforms.

I need to find the respiration rate as part of my project. My code is the same for each of my waveforms, but when I click run to calculate it I keep getting 90 bpm and I know they are not supposed to all equal 90. The data I have is 3 1000x1 matrix filled with correlation values ranging from 0-1. Below is the part of the code that I am trying to use to calculate respiration rate:

% %PEAK FINDING

neg_peaks=islocalmin(Corr_smooth);

figure;

plot(times, Corr_smooth, times(neg_peaks),Corr_smooth(neg_peaks), '*r');

t=(1:length(Corr_smooth))/Fs;

time_diffs=diff(t);

resp_rate=mean(1./time_diffs);

disp(['The respiration rate is ' num2str(resp_rate) ' bpm'])

My Fs=90 and I have to find the resp. rate using the negative peaks so that's why im not using findpeaks.

Any help is appreciated!

Answer 1

[Voss]

You are calculating the average of the inverse of the differences of t, but t is essentially just a vector of all time indices:

t = (1:length(Corr_smooth))/Fs;

so diff(t) is 1/Fs everywhere, and of course mean(1./diff(t)) is Fs.

I think you should be using the times where the local minima were found?

t = times(neg_peaks);

[Taya]

So I just need to change what t is equal to?

like this?

neg_peaks=islocalmin(Corr_smooth);

figure;

plot(times, Corr_smooth, times(neg_peaks),Corr_smooth(neg_peaks),'*r');

t=times(neg_peaks)******

time_diffs=diff(t);

resp_rate=mean(1./time_diffs);

disp(['The respiration rate is ' num2str(resp_rate) ' bpm'])

[Voss]

Yeah. Does that give reasonable results?

[Taya]

It made each of my waveforms different from each other, but they are all 20 bpm and below.

[Taya]

My data is from rats and I know their bpm are 60-100

[Taya]

So ig the short answer is no I do not have reasonable results

[Voss]

Is times given in minutes? If not, you'll need a unit conversion to get bpm.

[Taya]

times is something I had to calculate. Here is my code to calculate the times:

times=((1:length(Corr))/Fs);

The data I was given was 11 seconds of data per waveform

[Taya]

I don't know if that answers anything. I'm still not very familar with Matlab :/

[Voss]

Please upload Corr and Corr_smooth here. You can save those variables in a mat file and upload it using the paperclip icon.

[Image Analyst]

I was going to try to help, but then I noticed you forgot to attach your data. I'll try to remember to check back later for it.

If you have any more questions, then attach your data and code to read it in with the paperclip icon after you read this:

TUTORIAL: How to ask a question (on Answers) and get a fast answer

[Taya]

Here are the data for my 3 waveforms. Each of these is 11 seconds of data.

[Voss]

% load the data from the mat files:
files = dir('*.mat');
for ii = 1:numel(files)
    S(ii) = load(files(ii).name);
end

Fs = 90;

figure;

for ii = 1:3
    
    Corr = S(ii).Corr;
    
    % here I make up a method to calculate Corr_smooth from Corr:
    Corr_smooth = movmedian(Corr,7);
    
    % some additional parameters to islocalmin are required:
    neg_peaks = islocalmin(Corr_smooth,'MinProminence',0.05,'MinSeparation',45);
    
    times = (1:numel(Corr_smooth))/Fs;
    
    subplot(3,1,ii)
    plot(times, Corr_smooth, times(neg_peaks),Corr_smooth(neg_peaks), '*r');
    
    t = times(neg_peaks);
    resp_rate=mean(60./diff(t)); % here I assume t is in seconds
    title([num2str(resp_rate) ' bpm'])
    
end

Answer 2

[Star Strider]

Perhap;s this —

LD1 = load('USaverageImageIntensity_R1.mat');
Corr(:,1) = LD1.Corr;
LD2 = load('USaverageImageIntensity_R2.mat');
Corr(:,2) = LD2.Corr;
LD3 = load('USaverageImageIntensity_R3.mat');
Corr(:,3) = LD3.Corr;

Fs = 90;
t = linspace(0, size(Corr,1)-1, size(Corr,1)).'/Fs;

for k = 1:size(Corr,2)
    [vys,vlocs] = findpeaks(-Corr(:,k), 'MinPeakProminence',min(Corr(:,k))/10, 'MinPeakDistance',25);
    vysc{k} = [vys vlocs];
end

figure
plot(t, Corr)
grid
xlabel('Time (s)')
ylabel('Amplitude')

figure
tiledlayout(size(Corr,2),1)
for k  = 1:size(Corr,2)
    nexttile
    plot(t, Corr(:,k))
    hold on
    plot(t(vysc{k}(:,2)), -vysc{k}(:,1), 'vr')
    hold off
    grid
    xlabel('Time (s)')
    ylabel('Amplitude')
    RR = mean(1./diff(t(vysc{k}(:,2))))*60;
    title("Corr "+k)
    ylim([min(ylim) max(ylim)*1.25])
    text(3, 1.15, sprintf('Respiratory Rate = %.1f / minute', RR))
end

Make appropriate changes to get the desired result.

.