Compress .tck files

Hi All,
We have just preprocessed a large cohort and have ended up with an extremely large amount of data - particularly .tck files at 12gb each per session. Does anyone have any recommendations on compression? Is there any reason that MRtrix does not output them in a compressed format as with other outputs?

Any input would be appreciated.

Many Thanks,

Yes, that’s unfortunately not unusual. I would suggest that once your pipeline is up and running, it’s worth considering deleting these files once you’ve extracted all the required information from them. But I realise most users will likely feel much more comfortable holding on to them just in case…

This is a good question. The main issue is that track files don’t compress well. The last time I looked into it, none of the commonly used compression algorithms managed to reduce a typical tractogram all that much – certainly not enough for me to think it would be worth implementing. The consensus here is that to get decent compression, we’ll need to use lossy approaches specifically designed for tractography – some discussion of the issue on this thread, for instance.

So it’s certainly something we’re thinking about, but it’ll be a while before anything is available.

1 Like

Appreciate the thorough response, keep up the great work.