[Homer-users] Short Separation GLM without Block Avg

[Yucel, Meryem A.]

Hi Leah,

Please see my responses below:


  1.  We would like to be able to use short-separation channel regression and get the data for the entire sample rather than a block average. Is there a different function to do this or a way to modify hmrDeconvHRF_DriftSS.m to do this?
Once you run the analysis with the above function in HOMER, you can go back to matlab and load the nirs file. It has several variables automatically saved. You will find concentration time courses (ss regressed out if that was your option) under procResults.dc. This should work regardless of having a stim onset vector.

2.      How is the stimulus vector input used in hmrDeconvHRF_DriftSS.m and is it possible use the function without providing information about the stimuli?
Yes. Do you have a resting data that you want to perform connectivity type of analysis? Although you can still use this function to do ss regression, you have no control of removing the brain signal with that as well. (As opposed to looking at a functional brain response due to a stimulus where you have more control.)

3.      If we chose not to neutralize motion artifacts in hmrDeconvHRF_DriftSS.m, would it be appropriate to implement a different motion artifact removal method prior to hmrDeconvHRF_DriftSS.m? Or would the motion artifact removal go after hmrDeconvHRF_DriftSS.m in the processing pipeline?
Yes, you can opt out the motion correction option in the function and add any of the motion correction algorithms under processing stream before the GLM, and in fact, that is generally how we do it with HOMER

Hope this helps.

Meryem


From: homer-users-bounces at nmr.mgh.harvard.edu [mailto:homer-users-bounces at nmr.mgh.harvard.edu] On Behalf Of Friedman,Leah
Sent: Wednesday, May 9, 2018 3:25 PM
To: homer-users at nmr.mgh.harvard.edu
Subject: [Homer-users] Short Separation GLM without Block Avg


        External Email - Use Caution
Hello Homer users,

I had some questions regarding short-separation methods and the GLM function (hmrDeconvHRF_DriftSS.m). This function seems to be the primary method of using short-separation channel regression. However, it seems to require knowledge of experimental stimuli and has a built-in block average. We’ve watched a number of homer tutorials and training sessions and still have the following questions regarding these methods:


  1.  We would like to be able to use short-separation channel regression and get the data for the entire sample rather than a block average. Is there a different function to do this or a way to modify hmrDeconvHRF_DriftSS.m to do this?
  2.  How is the stimulus vector input used in hmrDeconvHRF_DriftSS.m and is it possible use the function without providing information about the stimuli?
  3.  If we chose not to neutralize motion artifacts in hmrDeconvHRF_DriftSS.m, would it be appropriate to implement a different motion artifact removal method prior to hmrDeconvHRF_DriftSS.m? Or would the motion artifact removal go after hmrDeconvHRF_DriftSS.m in the processing pipeline?

We have looked at using hmrDeconvHRF_DriftSS.m both within the GUI and as a standalone function outside of the GUI.

Thanks in advance for any responses (partial or complete)!

Sincerely,
Leah Friedman

—
Drexel University, 2019
B.S. Cognitive Neuroscience
lmf323 at drexel.edu<mailto:lmf323 at drexel.edu>
Drexel AIR Lab<http://www.drexelairlab.com/>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.nmr.mgh.harvard.edu/pipermail/homer-users/attachments/20180510/d7f6a70c/attachment-0001.html