Apparently, among functions that compute adjacency matrix for channels and sources, MNE currently does not have a function that computes the spatial adjacency matrix for cortical labels (e.g., Desikan-Killiani atlas with 68 cortical labels). Similar to mne.channels.find_ch_adjacency() function, such function should have list of labels as input and return the scipy.sparse.csr_matrix of labels adjacency with shape (n_labels, n_labels). The output of such function could be used for cluster permutational statistics.
Describe your proposed implementation
I suggest it can be done relatively easy using FreeSurfer triangulations.
Describe the new feature or enhancement
Apparently, among functions that compute adjacency matrix for channels and sources, MNE currently does not have a function that computes the spatial adjacency matrix for cortical labels (e.g., Desikan-Killiani atlas with 68 cortical labels). Similar to mne.channels.find_ch_adjacency() function, such function should have list of labels as input and return the scipy.sparse.csr_matrix of labels adjacency with shape (n_labels, n_labels). The output of such function could be used for cluster permutational statistics.
Describe your proposed implementation
I suggest it can be done relatively easy using FreeSurfer triangulations.