Visualize Connectivity matrix using imagesc

Discussions
1

Hi,

I want to visualize the correlation matrix using the Matlab command "imagesc".
But, the matrix exported from Brainstorm to Matlab is a somewhat different matrix to visualize N x N matrix.
Here I am having
TF: [36856×1×100 double]
36856 indicates ch information
100 indicates frequency information from 1Hz to 100Hz.
The number of MEG channels is 271.
How can I visualize the 271x271 matrix at Frequency f.

conn =

struct with fields:

           TF: [36856×1×100 double]
       TFmask: []
          Std: []
     DataType: 'data'
         Time: 315
    TimeBands: []
        Freqs: [100×1 double]
  RefRowNames: {1×271 cell}
     RowNames: {1×271 cell}
      Measure: 'other'
       Method: 'henv'
     DataFile: ''
  SurfaceFile: ''
      GridLoc: []
    GridAtlas: []
        Atlas: []
HeadModelFile: ''
HeadModelType: []
         nAvg: 30
         Leff: 1
 ColormapType: []
 DisplayUnits: []
      Options: [1×1 struct]
2

Hi @sugata_hisato,

The connectivity matrix (TF field) is saved as: [Nr x Ntime x Nfreq].
Nr represents the connectivity between two files, thus Na x Nb. However, the relation between Na x Nb and Nr may not be direct as it's optimized to save storage space for some connectivity metrics.
https://neuroimage.usc.edu/brainstorm/Tutorials/Connectivity#On_the_hard_drive

E.g. in correlation, as corr(a,b) == corr(b,a) only one value is saved. In your case 271*271 = 73441 elements; then the upper triangle of the connectivity matrix (36585 elements) plus the diagonal (271 elements) are saved, giving 36856 elements.

You can retrieve the connectivity matrix in the shape: [Na x Nb x Ntime x Nfreq] by using:

R = bst_memory('GetConnectMatrix', conn);

Best,
Raymundo

3

Hi, Raymundo,

I could reshape TF: [36856×1×100 double] using bst_memory and visualize the connectivity matrix by Matlab command.

Thanks a lot!

4

Hi,
I want to get each connectivity scout label from the exported connectivity structure from Brainstorm to Matlab.
Could you let me know how I can get all the combinations of scout labels (6328 scout x scout labels) from the below information?

struct with fields:

           TF: [6328×1×100 double]
       TFmask: []
          Std: []
     DataType: 'matrix'
         Time: 15
    TimeBands: []
        Freqs: [100×1 double]
  RefRowNames: {1×112 cell}
     RowNames: {1×112 cell}
      Measure: 'other'
       Method: 'henv'
     DataFile: ''
  SurfaceFile: ''
      GridLoc: [11602×3 double]
    GridAtlas: []
        Atlas: [1×1 struct]
HeadModelFile: ''
HeadModelType: []
         nAvg: 11
         Leff: 1
 ColormapType: []
 DisplayUnits: []
      Options: [1×1 struct]
      History: {1×3 cell}
5

See the documentation for the fields of the connectivity files in the Connectivity tutorial:
https://neuroimage.usc.edu/brainstorm/Tutorials/Connectivity?highlight=(RefRowNames)#On_the_hard_drive

After you reshape the TF matrix, you get a R matrix with dimensions [112 x 112 x 1 x nFreqs].
The labels for rows and columns are RefRowNames and RowNames.