Connor Lane

I'd be happy to add a review of this if that would help it move forward. Lmk if that would be useful.

Hi there, I see that a PR is in progress. But I thought it would still be worth bringing up a slightly different smoothing approach. In the HCP pipeline ([paper](https://surfer.nmr.mgh.harvard.edu/ftp/articles/2013/2013_-_Glasser_et_al._-_NeuroImage.pdf)...

This is a very fair point, thanks @noahbenson! In principle, it seems like it should be possible to do it fairly efficiently as you describe. Basically running a Dijkstra shortest...

I think the issue should be fixed in [this gist](https://gist.github.com/clane9/9b58f778c39560cf49a87cbb971795bc).

Thanks for pointing that out, @larsoner! I'll see if I can sort it out.

Hi @lucascolley, so sorry for abandoning this. I can try to pick it back up if it would still be useful. To start, I rebased with main to see what...

I removed the unused variables. Regarding the signed vs unsigned int comparisons, I'm a bit confused. It seems the complaints are about lines such as L374, where an `int` is...

Hi @rgommers, thanks for the helpful comments, much appreciated. I will work on addressing your comments but wanted to respond to a few things first. > More importantly, I don't...

@shnizzedy, @gkiar I'm fine with merging this now if you want. But I planned to add more notes after going through the default config in more detail. Lmk what you...

Thanks @effigies, I think these are good strategies for interacting with the memmap data. But still I'm confused why the simple `get_fdata()` is so much slower when `mmap=True` (the default)?...