Forum

Notifications
Clear all

scaling head size of user supplied brain structures in AtlasViewer  

Page 2 / 3

Meryem Yücel
Posts: 2
Moderator
(@mayucel)
Member
Joined: 4 months ago

Regarding point 1: Do you have "load precalculated" profile option under Forward Model menu? This is another way to load the sensitivity profile. Indeed we showed that way during the course, not how you would do it by "generate MC input" option. Do you want to use generate MC input to be able to use more photons or to change other parameters? If you prefer that way, maybe you explain the steps you took on how you did it. 

Reply
Abigail Fiske
Posts: 11
(@afiske)
Active Member
Joined: 2 months ago

@dboas @mayucel - thank-you both for your helpful responses.

1) Meryem, the 'load precalculated' option was greyed out. The only option I had available was to Generate MC input, so I did this. Do you have any idea why the load precalculated option was not available to me? It is because the groupResults.mat file was generated in Homer2, as in 2)?

2) I will re-run the processing stream in Homer3 to generate a new groupResults.mat file, and then try again to use this in AtlasViewer for the image reconstruction.

3) Yes, anatomical labels for imported MRIs would be really useful. At this point, would I have to rely on an adult atlas (e.g. the Colin preloaded atlas in AtlasViewer) to project probe to this adult cortex, to find the anatomical labels? I'm not sure how else I'd be able to identify what brain regions each channel is covering. Our probe consists of 46 channels (16 sources, 16 detectors) designed to cover the bilateral PFC and intraparietal sulcus.

Thanks again!

Abi

Reply
David Boas
Posts: 21
(@dboas)
Eminent Member
Joined: 2 months ago

@afiske, re point 3, yes, at this point you can just get anatomic labels on the Colin atlas. You can proceed as you suggest. I do hope we will set up labels for any MRI. We looked at this years ago and @jdubb was concerned about how long it would take and we've always had higher priorities for our own work... but hopefully we will get to it.

 

Re point 1, a few others were having this problem in the class. Perhaps @jdubb remembers what the solution was.

Reply
Abigail Fiske
Posts: 11
(@afiske)
Active Member
Joined: 2 months ago

I apologise for messaging again, but unfortunately I am getting no further with this and am instead going around in circles.

1) When I used the imported infant MRI atlases, the 'load precalculated...' option is greyed out. I can do the Generate MC Input option, but this generates the sensitivity map with 'holes' in, as I described previously. When I use the default Colin atlas with my same data and probe, I can load the precalculated sensitivity profile. However, I cannot run the ImageRecon because it does not recognise the groupResults.mat file that I generated in Homer2. This leads onto 2) below.

2) I opened Homer3 in my data folder and said yes to converting my .nirs files (60 files in 60 participant folders) to .snirf format. I manually excluded (using StimGui toggle on/off) the blocks for each participant that I knew were invalid from behavioural data. I imported my config file used in Homer2 and checked the parameters were the same. It said something about the groupResults.mat file (generated previously in Homer2) being relabelled to groupResults_old.mat. 

When running this at the group level and the processing had finished, I received a message box error saying "error closing file...the file may be corrupt" as well as a warning in MATLAB (see attached image). There is also a box saying "Auto-saving processing results....". However, a groupResults.mat file with the current date and time is saved in the folder. This then does not allow me to open AtlasViewer, as I receive another error message in MATLAB saying "error using load. Cannot read file ....../groupResults.mat". I assume this is because the groupResults file is corrupt.

3) I'm able to project probe to cortex with the adult Colin atlas, but am unable to map our HRF data onto this due to point 2.

Really appreciative of any support you can offer, thank-you.

Abi

Reply
David Boas
Posts: 21
(@dboas)
Eminent Member
Joined: 2 months ago

@afiske let's deal with 2 and then move to 3 later. For pt 2, can you share a link to the groupResults.mat file? A dropbox of google share link or something like that? I suspect the file is huge and shouldn't be attached here or by email.

I am not confident this will enable us to figure out the problem. I fear this will require us to reproduce the problem to fix... but this means we would need your 60 data files. Ugh.

@jdubb, have you seen anything like this before?

Reply
Page 2 / 3
Share: