(2 ideas) COMMIT2 Conncectome Filtering and Connectome Spatial Smoothing

[smeisler]

Hi,

I just read these papers about COMMIT2 and Connectome Spatial Smoothing.

1. Advantages of COMMIT2 include computational efficiency (quicker to run than SIFT2, and some streamlines will get removed/0-weighted, so subsequent tractogram operations will be quicker), and, at least according to the paper, more informed removal/de-weighting of false positives. I believe COMMIT frameworks were tested on iFod2 tractograms, so that much can be borrowed from the current Mrtrix pipelines. It seems they have also made more augmentations to their COMMIT framework recently (Hierarchical streamline clustering and accounting for motion artifacts/signal dropout).

2. With the exciting progress of including HSVS in reconstruction streamlines, it appears it would be possible and beneficial to use connectome spatial smoothing, which is informed by geodesic distances between atlas nodes on subject surfaces.

I hope these papers will encourage some good discussion about future features.

Thanks,
Steven