Trouble with findpeaks - 14 EEG signals and peaks on same plot - different colors

Hi,

I am trying to plot 14 EEG signals and their peaks on the same plot, but I am having trouble getting all of the findpeaks to have the triangle symbol in the same color as the EEG signal. Rather than have matching colors, and have all findpeaks have the triangle symbol, I keep getting a legend like this:

Is there a way I can get them to match colorwise and to make all of the findpeaks have the triangle symbol? I've attached a zip folder containing my matlab code and the csv file (this figure's code starts at line 36).

Also, I know it won't be easily legible with all 14 signals and peaks on the plot. I plan to add and remove certain signals as I need to, but I'd like to have all 14 signals on the same plot to start with.

Secondary Request:

If possible, I'd like to have the peaks on the right column of the legend and the EEG signals on the left column. Also, is there a way I can title a legend column?

Thank you for your help!

[dpb]

Please just attach the text of the pertinent part of the m-files instead of making us download (yet another) zip file...

findpeaks draws both the line and places the markers on the plot itself; you don't need to draw the line additionally with plot().

findpeaks, however, does NOT return the handle to the lines drawn; if use output variables it then doesn't draw the graph at all and returns the requested information about the located peaks.

So, use findpeaks as I showed before, then before moving on to the next axis, retrieve the line handles and set the properties of those as desired.

IF you do things sequentially, you can get the legend to appear as you want by ordering the line handles in alternating order and having the labels to match.

Of course, you can also use the 'DisplayName' property of the line instead of writing it explicitly in the legend.

ADDENDUM

To reorder the order of variables, just build an indexing array of column number or name in the desired order and use it in the loop to retrieve the next by reference. There's no need to create a zillion temporary variables and write every line of code over and over -- that's the power of having a programming language at your beck and call to be able to generalize instead of copy and paste like a spreadsheet.

Also, if you want a different name for the variables, fix up the 'VariableNames' property to contain them as you want them named -- there's no magic there that they have to keep the name in the file--just they have to be unique.

[Christina Diersing]

Thank you again for your help!

I got the findpeaks to work now and the plot looks exactly how I wanted it to.

I am trying to do what you suggested for the variable names, but I am having some difficulty. (I have never used table properties before yesterday and, though I read through the documentation and tried several different lines of code to try to rename the variable names, I am still having trouble.) Basically, I am trying to write a for loop like you had on my other question, but to replace the repetitive lines of code at the beginning of this excerpt:

AF3=T.EEG_AF3; F7=T.EEG_F7;
F3=T.EEG_F3; FC5=T.EEG_FC5;
T7=T.EEG_T7; P7=T.EEG_P7;
O1=T.EEG_O1; O2=T.EEG_O2;
P8=T.EEG_P8; T8=T.EEG_T8;
FC6=T.EEG_FC6; F4=T.EEG_F4;
F8=T.EEG_F8; AF4=T.EEG_AF4;
for j=2:width(T)
    NewVarNames(j)=extractAfter(T.Properties.VariableNames(j),'_');
    %T.Properties.VariableNames(j)=extractAfter(T.Properties.VariableNames(j),'_');
end
T.Properties.VariableNames=NewVarNames;

It's likely a stupid error on my part, but I have spent at least 3 hours trying to get it to work. Could you help with this please?

[dpb]

I did not see the above plaintive cry in the wilderness until now...sorry!

>> tTry.Properties.VariableNames(2:end)=extractAfter(tTry.Properties.VariableNames(2:end),'_');
>> head(tTry)
ans =
  8×15 table
      Timestamp        AF3        F7         F3         FC5        T7         P7         O1         O2         P8         T8         FC6        F4         F8         AF4  
    _____________    _______    _______    _______    _______    _______    _______    _______    _______    _______    _______    _______    _______    _______    _______
    1622835751.19    4260.38    4267.69    4316.03    4305.51    4340.00    4278.46    4278.97    4279.62    4235.26    4184.87    4456.67    4401.79    4343.59    4387.05
    1622835751.20    4264.87    4261.92    4311.79    4308.59    4338.97    4280.26    4279.74    4276.79    4232.18    4184.74    4442.69    4401.54    4334.23    4380.77
    1622835751.20    4266.28    4272.05    4316.41    4312.56    4340.51    4280.77    4286.41    4275.77    4229.62    4178.33    4438.21    4405.38    4325.26    4381.03
    1622835751.21    4262.82    4274.74    4319.74    4311.28    4342.44    4280.77    4288.21    4281.41    4242.82    4181.28    4444.87    4407.82    4330.64    4389.49
    1622835751.22    4266.15    4267.56    4318.85    4308.08    4332.56    4279.23    4282.31    4285.64    4248.59    4186.03    4448.72    4404.23    4333.72    4392.31
    1622835751.23    4266.92    4267.95    4317.69    4298.33    4322.95    4277.95    4275.77    4275.13    4235.00    4179.87    4440.77    4398.97    4326.28    4386.92
    1622835751.24    4255.26    4269.49    4309.87    4293.21    4328.08    4278.08    4270.13    4265.38    4229.62    4177.95    4425.51    4394.49    4325.26    4383.33
    1622835751.24    4254.87    4265.13    4306.54    4303.33    4336.54    4280.13    4267.82    4272.44    4242.95    4184.74    4431.79    4397.05    4328.21    4382.18
>> 

Answer 1

[dpb]

T.Timestamp=T.Timestamp-T.Timestamp(1);   % 1622835751.18846;
varlist={'AF3','F7','F3','FC5','T7','P7','O1','O2','P8','T8','FC6','F4','F8','AF4'};
%% Reg Plot - all EEG Signals and peaks on same plot
hF1=figure('Name', 'EEG Signals and Peaks - Same Plot');
newcolors={'#0072BD','#D95319','#EDB120','#7E2F8E','#77AC30','#4DBEEE','#A2142F', ...
           '#FF0000','#00FF00','#0000FF','#00FFFF','#FF00FF','#FFFF00','#000000'};
colororder(newcolors);
grid on; hold on;
for i=1:numel(varlist)
  ix=find(matches(extractAfter(tTry.Properties.VariableNames,'_'),varlist(i)));   % get the wanted variable in list order
  findpeaks(T{:,ix}, T.Timestamp,'MinPeakProminence',100);                        % draw the findpeaks plot
end
hAx=gca; hL=hAx.Children;                                                         % retrieve handles to lines
hLG=legend([hL(2:2:end);hL(1:2:end)],[varlist strcat(varlist,' Peaks')], 'NumColumns', 2);  % write legend in desired order
hLG.FontSize=7; hLG.Location='best';
title('Try 4 EEG Test');
xlabel('Timestamp (s)');
ylabel('EEG Value (uV)');
fig1.WindowState = 'maximize';

results in

where just a little frontend effort lets one reduce some 14 calls to a loop of 4 lines that can be arranged in whatever order is later desired if ever want to change simply by reordering one array of names as well as elminate the difference in colors owing to extra uneeded call to plot.

The legend is rearranged as desired simply by reordering the order of the handles passed; like everything else in MATLAB, the default order is by column and findpeaks puts the peaks line first where the desire is to have the solid line first instead. One could do it by rearranging the handle array order; I just took the expedient of passing the two sets in the desired order to legend directly.

[dpb]

"The legend is rearranged as desired simply by reordering the order of the handles passed;"

Since the channel is given in the first column, if it were me I think I'd just use 'Peaks' for the second column of the legend to reduce the clutter slightly instead of duplicating the channel ID....