search

cvnlab / GLMsingle

A toolbox for accurate single-trial estimates in fMRI time-series data
BSD 3-Clause "New" or "Revised" License
101 stars 45 forks source link

Data quality and amount #150

Open cindyhfls opened 1 month ago

cindyhfls commented 1 month ago

Hi,

I was wondering what the data quality and amount are required for this toolbox to work well. In the video tutorial, it says we need high-quality data. What does this mean? Do we need 30 min? 1 hr? 7T? 3T? DOT?

Cindy

kendrickkay commented 1 month ago

This is a good question. Unfortunately it's really hard to give hard and fast rules. (By the way, what is DOT?)

For sure, 3T data is fine, and for sure, under certain circumstances 30 min of data is sufficient.

Do you have some specific concerns/issues that you are envisioning?

cindyhfls commented 2 weeks ago

DOT = Diffusion Optical Tomography, which is similar to fNIRS with infrared imaging. The resolution is slightly worse than MRI. I don't have a specific concern and I don't have a specific dataset now, but I am interested in doing decoding with neuroimaging in the future and just curious about what kind of data would suffice.

What do you mean by "under certain circumstances"?

kendrickkay commented 2 weeks ago

Well, what I mean is that it's impossible to say something like "30 min is sufficient", since there are many other factors that contribute to whether it is sufficient or not.

Regarding DOT --> Oh, that's interesting. GLMsingle was designed with fMRI data in mind. There is some chance that it might generalize well to other types of imaging modalities. But it depends on the nature of the signal and noise in a given modality. Likely it'll require some careful methodological work to see if it works well, and if not, adapt the tool such that it will work well.