Skip to main content
. 2023 Jan 2;10(1):56. doi: 10.3390/bioengineering10010056
Algorithm1 Step-by-step rs-fMRI diagnosis algorithm
  • 1:

    ∀ rs-fMRI BOLD signal data:

  • 2:

          1.∀ preprocessing strategies:

  • 3:

               (i) with/without global signal correction.

  • 4:

               (ii) with/without band-pass filtering.

  • 5:

          2. Use ∀ atlas ∈ {AAL, TT}, ∀ preprocessing strategy:

  • 6:

               (i) AAL atlas with 116 brain regions

  • 7:

               (ii) TT atlas with 97 brain regions

  • 8:

          3. Calculate the two feature representations for each atlas/strategy:

  • 9:

                (i) Static functional connectivity matrix FC

  • 10:

              (ii) Dynamic functional connectivity dFC

  • 11:

                   (I) Use a Gaussian smoothed sliding window over each pair of brain regions to calculate pair-wise dynamic correlations.

  • 12:

                   (II) ∀ pair of brain regions, calculate the fraction of no correlation as nwk, and the fraction of strong correlation as nst.

  • 13:

                   (III) Use these aggregations for each region pair as the new dynamic functional connectivity feature to create the feature matrix dFC.

  • 14:

    Feature Selection:

  • 15:

          For each feature representation, run a univariate selector to reduce feature space

  • 16:

          For each of the four RFE-CV kernels, find the n features that provide the highest cross-validated balanced accuracy to be used for each of the kernels.

  • 17:

    Classification:

  • 18:

          ∀ classifier, for each configuration of hyper-parameters, for each reduced feature representation:

  • 19:

           (i) Split reduced Xselect, with n selected features, into k folds, along with labels y.

  • 20:

           (ii) Calculate the cross-validated score for each hyper-parameters’ configuration.

  • 21:

          (iii) Determine the best hyper-parameters configuration in terms of score for each classifier.

  • 22:

          (iv) Output the best classifier/parameters, along with its used n features.