Good morning Toronto!
We are finally at ISMRM 2015 after a 7 hours roadtrip to Toronto! Be sure to come see and chat up with the various members from the SCIL that are here with me. Expect us all for the results of the ISMRM tractography challenge on monday in room Constitution Hall 105 at 17h30, but feel free to also speak up with the various people wearing dipy shirts (made by Elef specifically for ISMRM and HBM this year). Also expect a ton of cool posters by us and a talk by JC about how to do group statistics on compressed fibers the proper way on friday session. My poster will be on how to get acquisition which are quality-wise the same as the Human Connectome Project (HCP) on any regular scanner in only 13 minutes, thanks to post-processing with denoising. And don't forget to make #SCIL #ismrm15 trending!
So hey, my poster session is on thursday from 1h30 to 3h30, so come say hi if you can. I will be presenting how you can acquire a high resolution diffusion MRI dataset on your regular scanner, and get awesome results and metrics from it by removing the noise. Using the denoising I presented at last year's ISMRM, it is indeed possible to obtain high quality datasets by postprocessing after the acquisition, without changing your scanner nor the sequence it uses. Be sure to come chat with about how you can also use these techniques and the improvements they bring, as I plan to provide the source code of the method on github later on.
Gab will be presenting his recent work on connectivity using tractography and new matrices for analysing the data. Etienne will be presenting how surface tracking made from T1 weighted images can enhance the initialization and seeding of diffusion MRI tractography by using anatomical priors. François work is about how to load millions of streamlines in real-time, thanks to a new fiber compression algorithm we presented last year also. Marc-Alexandre will present his work about automatic outliers removal arising from spurious fibers made by tractography which are implausible. Elef has a new bundle extraction algorithm, where you can automatically extract similar bundles from a population once you have a dissection of the bundle you want, thus removing the need to dissect them all by hand.
And with that, expect us en masse this week!