How can I run PCA on a 3-dimensional cell arrays?

Hi,

I have a cell array (see attachment) where each cell contains an array with 21 columns (variables). Each cell represents a different participants and is of different length. I would like to run the PCA on each cell in the cell array and collect the outputs (coeff,score,latent,tsquared,explained,mu) for all participants.

The fucntion I want to use is the following with X being my cell array pre_data_1.

[coeff,score,latent,tsquared,explained,mu] = pca(X)

I am specifically interested in getting an "explained" that contains all the explained variables from each participant. In this variable each column should correspond to an explained variable from a cell/ participant.

This is what I have until now:

numCells = numel(pre_data1);
for k = 1 : numCells
    for p = 1 : n
        thisX = pre_data1{k}{p}; 
        [coeff{k}{p},score{k}{p},latent{k}{p},tsquared{k}{p},explained{k}{p},mu{k}{p}] = pca(thisX); 
    end
end

But when I run this I only get strange results like so:

explained =

1×2 cell array

{1×1 cell} {1×1 cell}

I wonder if someone could help me with that?