Hi, thank you for sharing this repo !
Apologies if this is already implemented, but I have tried to find an answer by browsing the source code and couldn't find it.
the original CAP paper (Liu et Duyn 2013) uses a seed-based intensity thresholding to keep a small subset of frames on which clustering is then applied. Shortly after, they perform the clustering directly on the full timeserie, in a hypothesis free, datadriven way.
The two methods seems to be compatible and able to answer different questions. Would you be open to an extension of your repo to add the possibility to perform a voxelwise clustering on a subset of frames selected by specifying a seed node in a parcellation ? Or do you feel it is too different from what your repo is doing to be included ?
Thank you for your time
Hi, thank you for sharing this repo !
Apologies if this is already implemented, but I have tried to find an answer by browsing the source code and couldn't find it.
the original CAP paper (Liu et Duyn 2013) uses a seed-based intensity thresholding to keep a small subset of frames on which clustering is then applied. Shortly after, they perform the clustering directly on the full timeserie, in a hypothesis free, datadriven way.
The two methods seems to be compatible and able to answer different questions. Would you be open to an extension of your repo to add the possibility to perform a voxelwise clustering on a subset of frames selected by specifying a seed node in a parcellation ? Or do you feel it is too different from what your repo is doing to be included ?
Thank you for your time