Why do I get more error bars than I defined?

Hello,

I want to plot the mean curve of two measurements with 7 errorbars i calculated with the standard deviation.

However every time I try to plot the errorbars, more than I defined appear.

Could someone please tell me, what I'm doing wrong?

I also attached the figure I get.

Thank you in advance!

clear all

clc

load ('data_repod_ad.mat');

num_vector = length(data.n(2).gray_value);
error=zeros(1,num_vector);

% Mittelwert berechnen
for ii=1:num_vector
    MW(ii) = mean (data.n(1).gray_value(ii)+data.n(2).gray_value(ii))/2; 
    time(ii) = mean (data.n(1).time(ii)+data.n(2).time(ii))/2; 
    if ii==40
        error(ii)=sqrt((data.n(1).gray_value(ii)-MW(ii))^2+(data.n(2).gray_value(ii)-MW(ii))^2)*2;
    elseif ii==80
        error(ii)=sqrt((data.n(1).gray_value(ii)-MW(ii))^2+(data.n(2).gray_value(ii)-MW(ii))^2)*2;
    elseif ii==120
        error(ii)=sqrt((data.n(1).gray_value(ii)-MW(ii))^2+(data.n(2).gray_value(ii)-MW(ii))^2)*2;
    elseif ii==160
        error(ii)=sqrt((data.n(1).gray_value(ii)-MW(ii))^2+(data.n(2).gray_value(ii)-MW(ii))^2)*2;
    elseif ii==200
        error(ii)=sqrt((data.n(1).gray_value(ii)-MW(ii))^2+(data.n(2).gray_value(ii)-MW(ii))^2)*2;
    elseif ii==240
        error(ii)=sqrt((data.n(1).gray_value(ii)-MW(ii))^2+(data.n(2).gray_value(ii)-MW(ii))^2)*2;
    elseif ii==280
        error(ii)=sqrt((data.n(1).gray_value(ii)-MW(ii))^2+(data.n(2).gray_value(ii)-MW(ii))^2)*2;
    end
end
%% Plot erstellen

f = figure(3);  
f.Position = [10 10 550 400]; 
k= [[0 0 0];[0 0 1];[1 0 0];[0 1 0];[0 1 1];[1 0 1]];
errorbar(time,MW,error,'Color',k(1,:),'LineWidth',1); hold on;

grid on
xlim([160 10000]); 
xlabel('time in ms');
ylabel('mean gray value of ROI');

Answer 1

[the cyclist]

For most of your points, you are not plotting no error bar; you are plotting an error bar with zero error.

You could make the default error equal to NaN, and I think you will get what you want.

error_zero = [0 0 0.03 0.04];
error_nan = [nan nan 0.03 0.04];

y = [1 1.1 1.2 1.3];

figure
errorbar(y,error_zero)

figure
errorbar(y,error_nan)

Answer 2

[Voss]

error is a 1-by-284 vector (going by the .fig file) whose elements are all 0 except at indices 40, 80, 120, 160, 200, 240, and 280. Those are the only indices where error is calculated, by the code inside the for ii=1:num_vector loop, so when you use errorbar(...,error,...) you get 284 error bars, of which all but those 7 have magnitude zero.

Note that each of the lines of code under the if or an elseif inside that loop are identical, so the entire if/elseif/elseif/... construction could be implemented more simply as:

if ismember(ii,40:40:280)
    error(ii)=sqrt((data.n(1).gray_value(ii)-MW(ii))^2+(data.n(2).gray_value(ii)-MW(ii))^2)*2;
end

but I don't think either of the if/elseif/elseif/... or if ismember(ii,40:40:280) is what you really intend to do, since those both ignore all elements of MW and time besides those 7.

clear all
clc
% load ('data_repod_ad.mat');
data = struct('n',struct( ... % random data struct array
    'gray_value',{rand(1,284) rand(1,284)}, ...
    'time',repmat({linspace(160,10000,284)},1,2)));

num_vector = length(data.n(2).gray_value);
error=zeros(1,num_vector);

% Mittelwert berechnen
for ii=1:num_vector
    MW(ii) = mean (data.n(1).gray_value(ii)+data.n(2).gray_value(ii))/2; 
    time(ii) = mean (data.n(1).time(ii)+data.n(2).time(ii))/2; 
    if ismember(ii,40:40:280)
        error(ii)=sqrt((data.n(1).gray_value(ii)-MW(ii))^2+(data.n(2).gray_value(ii)-MW(ii))^2)*2;
    end
end
error

error = 1×284

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

find(error ~= 0)

ans = 1×7

40 80 120 160 200 240 280

Answer 3

[Theresa Kandels]

Thank you both for your answers! It worked :)