Recommendation

[LPNC-LANG]

• What are you trying to do?
  I"m running gradient analysis onto phase-locking networks obtained from source-space MEG data. I have a 52-parcel resolution but am not particularly interested in identifying fine-grained functional boundaries anyway, I'm just looking for a way to embed each connectivity matrix into a one-dimensional array for later statistical analysis.
  Relatedly, I couldn't find definitive recommendations on how the choice of kernels and dimensionality reduction algorithm may affect the output? What would you recommend in my case?