Matrices size difference within for loop and unable to perform assignment of elements

Hello MathWords community,

I am trying to calculate maximum peak to peak amplitude of 11-16 Hz bandpass filtered brain EEG activity (shown below).

The code below is showing my attempt in doing that and I am having issues with the line " SS_Amp(:,iNepochs) = PeakPo + PeakNe;".

This line is supposed to add the values of positive peaks and negative peaks and progressively add the result-matrix (which is 21 by X number of inputs; 21 is due to the number of EEG electrodes) into a for loop. The result-matrix is supposed to be progressively added to the already-specified "SS_Amp = zeros (Nchannels,Nepochs);" matrix.

However, whenever I attempt this I get the below error message saying that the size of the left side is 21 by 1 and the size of the right side is 21 by 5. I do not understand why the size of the left side is 21 by 1 - the size of the left side should also be 21 by 5 as the number of epochs (Nepochs) should have been specified before the beginning of the outer for loop.

How would I solve this issue?

%% ===== RUN =====
function OutputFiles = Run(sProcess, sInputs) %#ok<DEFNU>
    % Initialize returned list of files
    OutputFiles = {};
    Nepochs     = length(sInputs); % Numer of inputs or epochs
    EpochN      = linspace (1,(Nepochs),(Nepochs)); % Changes the x-axis from time (s) to epoch number n. 
    Fs          = 1024; % Sampling Frequency
    DataMat     = in_bst(sInputs(1).FileName, [], 0); % Read the first file in the list, to obtain channel size. 
    epochSize   = size(DataMat.F);    
    Nchannels   = epochSize(1); % Number of ALL channels (i.e. EEG, EOG, EMG and status)

    % Generate matrices of zeros (Nchannels X Nepochs) to concatenate results
    SS_Amp = zeros (Nchannels,Nepochs); % Generates 21 by 5 zeros to be used down below
    PeakPo = zeros (Nchannels,Nepochs); % Positive peaks
    PeakNe = zeros (Nchannels,Nepochs); % Negative peaks 
    IndexP = zeros (Nchannels,Nepochs); % Indices of positive peaks
    IndexN = zeros (Nchannels,Nepochs); % Indices of negative peaks

    % ===== LOAD THE DATA =====
    for iNepochs = 1 : Nepochs
       DataMat   = in_bst(sInputs(iNepochs).FileName, [], 0); % LOAD THE DATA
       epochSize = size(DataMat.F);  
       Ntime     = epochSize(2); % Number of time samples changes depending on sleep spindel.
       Data      = DataMat.F; % Actual data containing EEG, EOG and EMG data

    % ===== PROCESS =====
    % This is where the actual process of data manipulation and calculation takes place.
       DataN        = Data*-1; % The negative version of the data
       Data(end,:)  = (randn(1,Ntime))*10^-6;
       DataN(end,:) = (randn(1,Ntime))*10^-6;
       % Last channel which is BDF/status channel is a constant number so replaced
       % this channel with random number to remove issues during findpeaks 
       Data_Chan    = zeros(1,Ntime);
       DataN_Chan   = zeros(1,Ntime);

            for i = 1 : Nchannels
                Data_Chan(1,:)     = Data(i,:);
                DataN_Chan(1,:)    = DataN(i,:);
                [PosP, LocP]       = findpeaks(Data_Chan,Fs);
                [NegP, LocN]       = findpeaks(DataN_Chan,Fs);
                [PeakPo(i,iNepochs), IndexP(i,iNepochs)] = max(PosP,[],2);
                [PeakNe(i,iNepochs), IndexN(i,iNepochs)] = max(NegP,[],2);
            end
            
       SS_Amp(:,iNepochs) = PeakPo + PeakNe; % Final calculation of max peak-peak amplitude
       
    end
end

[KSSV]

What are the dimensions of input? Tell us about sProcess and sInputs.

[MinChul Park]

hello KSSV thank you for your question and sorry for the lack of information!

sInputs = 21 by Y by 5 where

21 = 21 electrodes for the brain

Y = number of time samples which are variable and

5 = the number of files that are being used.

sProcess is irrelevant to the calculation so I've omtted the information

Answer 1

[MinChul Park]

Change in code from ...

SS_Amp(:,iNepochs) = PeakPo + PeakNe; % Final calculation of max peak-peak amplitude

To...

SS_Amp = PeakPo + PeakNe; % Final calculation of max peak-peak amplitude

Answer 2

[Paul]

Hi mcp0228,

The loop variable iNepochs will be a scalar value, hence the LHS of the assignment is a column vector.

    for iNepochs = 1 : Nepochs

       % do a bunch of stuff here
        
       SS_Amp(:,iNepochs) = PeakPo + PeakNe; % Final calculation of max peak-peak amplitude
       
    end

Based on this statement:

"The result-matrix is supposed to be progressively added to the already-specified "SS_Amp = zeros (Nchannels,Nepochs);" matrix."

maybe that line should be

SS_Amp = SS_Amp + PeakPo + PeakNe;

But only you would know for sure.

[MinChul Park]

Thank you for your input Paul however as the sole purpose of "SS_Amp = zeros (Nchannels,Nepochs);" is to concatenate the results that were generated by the nested for loop, if you attempt "SS_Amp = SS_Amp + PeakPo + PeakNe;" I get the same error message displayed above.

[Walter Roberson]

If you are still getting an error message about the size of the left side is 21 by 1 and the size of the right side is 21 by 5 even though you no longer have the line

       SS_Amp(:,iNepochs) = PeakPo + PeakNe; % Final calculation of max peak-peak amplitude

then the error would have to be against some other line now -- but which one?

We do not have your files and we do not have some of your functions, so we cannot test the code -- so we are relying on you to correctly report your error messages.

[MinChul Park]

Hello Walter thank you for your feedbacks! Sorry for the missing information.

Fortunately though, this has been resolved. I just needed to change the code from:

SS_Amp(:,iNepochs) = PeakPo + PeakNe; % Final calculation of max peak-peak amplitude

TO...

SS_Amp = PeakPo + PeakNe; % Final calculation of max peak-peak amplitude