How to do comparison using TTEST2 to get the p-value between two 4-D connectivity matrices.

(Hello, I'm fairly new to statistical analysis and EEG data analysis)

I have two connectivity matrices of EEG data that I want to compare. The thing that confuses me is the dimension of the matrices (4D - single), does anyone know what each of the dimensions represent? And how to find the p-value for the comparison (t-test)?

I've tried:

DTF_trial1 % connectivity matrix 1 (4-D single)
DTF_trial2 % connectivity matrix 2 (4-D single)

[H,P] = ttest2(DTF_trial1, DTF_trial2)

but it returned thousands of lines of 5 values:

P(:,:,96,151) =

    0.6100    0.0308    0.0899    0.2211    0.2078

P(:,:,97,151) =

    0.3384    0.0222    0.0878    0.2718    0.1502

P(:,:,98,151) =

    0.2317    0.0190    0.0871    0.3884    0.1364

P(:,:,99,151) =

    0.1713    0.0237    0.0869    0.5536    0.1350

Thank you.

Answer 1

[Paras Gupta]

Hi Sandy,

I understand that you want to know, what the different dimensions used in 4D connectivity analysis matrices represent, and how we can compute the p-value for comparison using a t-test.

In the context of EEG data, a connectivity matrix represents the functional connectivity between different brain regions or electrodes. Each dimension of the 4D connectivity matrix has a specific meaning:

• Dimension 1: Represents the source region or electrode from which the connectivity is measured.
• Dimension 2: Represents the target region or electrode to which the connectivity is measured.
• Dimension 3: Represents the frequency or frequency band at which the connectivity is measured.
• Dimension 4: Represents the trial or sample number

Since you want to compare the connectivity matrices across all dimensions, you can reshape the matrices into 2D arrays before performing the t-test and compare connectivity across all trials (4th dimension).

% Reshape the connectivity matrices into 2D arrays 
DTF_trial1_2D = reshape(DTF_trial1, [], size(DTF_trial1, 4)); 
DTF_trial2_2D = reshape(DTF_trial2, [], size(DTF_trial2, 4));  

% Perform the t-test 
[H, P] = ttest2(DTF_trial1_2D, DTF_trial2_2D); 

Please refer to the following documentation for more information on the functions used in the code above.

• ttest2: https://www.mathworks.com/help/stats/ttest2.html
• reshape: https://www.mathworks.com/help/matlab/ref/reshape.html

Hope this helps.