calculating the variability between three different mice and within one mouse code. pleas help to confirm works as intended

Hi all, I have 6 mat files of data corresponding to 3 mice where 2 of the mat files correspond to 1 mouse yielding two mat files for each mouse.

I want to determine the variability within one mouse and then the variability between the mice, i have made a code to do this but as a novice matlab user i want to confirm and improve it. Can someone please help me to do this? Thanks!

% read in data from mat files
data = [];
for i = 1:3
    mouse_data = [];
    for j = 1:2
        filename = sprintf('feature%d.mat', (i-1)*2+j);
        temp = load(filename);
        mouse_data = [mouse_data temp.(sprintf('feature%d', (i-1)*2+j))];
    end
    data = [data mouse_data];
end

% calculate the mean and variance for each mouse
mouse_mean = mean(data, 2);
mouse_var = var(data, 0, 2);

% calculate the mean and variance for all mice together
group_mean = mean(data(:));
group_var = var(data(:));

% plot the variance within and between the mice
figure;
bar([1 2 3], mouse_var);
xlabel('Mouse');
ylabel('Variance');
title('Variance within each mouse');
ylim([0 0.3]);

figure;
bar([1 2], [group_var, mean(mouse_var)]);
xlabel('Group');
ylabel('Variance');
title('Variance between mice and within-group variance');
ylim([0 0.3]);
xticklabels({'Between mice', 'Within-group'});

[Star Strider]

If you have the Statistics and Machine Learning Toolbox, explore the various analysis-of-variance functions. It would appear that at least anova1 could be appropriate.

[Chanille]

Like what is a an example to run these features through the anova1 test?

[Chanille]

@star strider

[Star Strider]

See the documentation I linked to, first to be certain that it does what you want (since I’m not certain that I sufficiently understand the problem to be solved), and second to understand how to use the function and interpret its results.

Answer 1

[LeoAiE]

Your code reads the data from the .mat files, concatenates them, and calculates the mean and variance for each mouse and the group. However, there are a few improvements and adjustments that you could make to better represent the variability within and between mice.

1. Calculate the mean and variance for each mouse separately, without concatenating the data.
2. Calculate the mean and variance for each individual trial and then calculate the overall mean and variance for each mouse.

Here's an updated version of your code that takes these points into consideration:

% Read in data from mat files
data = cell(3,2);
for i = 1:3
    for j = 1:2
        filename = sprintf('feature%d.mat', (i-1)*2+j);
        temp = load(filename);
        data{i, j} = temp.(sprintf('feature%d', (i-1)*2+j));
    end
end

% Calculate the mean and variance for each mouse and trial
mouse_mean = zeros(3, 1);
mouse_var = zeros(3, 1);
trial_mean = cell(3, 2);
trial_var = cell(3, 2);

for i = 1:3
    trial_data = [];
    for j = 1:2
        trial_mean{i, j} = mean(data{i, j}, 2);
        trial_var{i, j} = var(data{i, j}, 0, 2);
        trial_data = [trial_data data{i, j}];
    end
    mouse_mean(i) = mean(trial_data(:));
    mouse_var(i) = var(trial_data(:));
end

% Calculate the mean and variance for all mice together
group_mean = mean(cell2mat(data(:)));
group_var = var(cell2mat(data(:)));

% Plot the variance within and between the mice
figure;
bar([1 2 3], mouse_var);
xlabel('Mouse');
ylabel('Variance');
title('Variance within each mouse');
ylim([0 0.3]);

figure;
bar([1 2], [group_var, mean(mouse_var)]);
xlabel('Group');
ylabel('Variance');
title('Variance between mice and within-group variance');
ylim([0 0.3]);
xticklabels({'Between mice', 'Within-group'});