Algorithm 1:

Manifold Learning and visualization of EEG dynamics using graph dissimilarity embedding and geodesic-informed minimum spanning tree

Input: Graph set G, where $G=\left\{{ℊ}_1,{ℊ}_2,\dots {ℊ}_n\right\}$ and each ${ℊ}_i$ is a 34 by 34 EEG connectome

Output: Geodesic distance matrix GDM, where GDM $\in R^{n\times n}$

Minimum Spanning Tree MST of G

1. [Optional] Select Prototype graph set $G_{pro}\subseteq G,G_{pro}=\left\{g_{pro}^1,g_{pro}^2\dots ,g_{pro}^m\right\};m\le n$

2. Construct dissimilarity embedding using Eq. 2

3. for each $g_i,g_j$ in G

4. Construct Euclidean distance matrix $\left(\text{EDM}\right),EDM\in R^{n\times n}$ Euclidean distance between         ${\phi }_n^G\left(g_i\right)\hspace{0.17em}\text{and}\hspace{0.17em}{\phi }_n^G\left(g_j\right)$ is given in Eq. 3

5. end for

6. GDM← Apply Dijkstra algorithm and k nearest neighbors $\left(k=60\right)$ to EDM

7. return GDM

8. MST ← apply TreeV is to GDM to yield geodesic informed minimum spanning tree

9. plot MST