[Homer-users] Problem with wavelet analysis in Homer2

[antonio chiarelli]

Hello Geoff,

with respect to the kurtosis based wavelet filtering,

it definitely should not oscillate like that.

An original version of the code i provided for homer2 had a bug that i fixed,

however i'm wondering if the bug was not corrected in the available code.

Can you tell me where you downloaded the file in order to check if the bug is already there?

I'm glad to help you with this,

kinds regards,

antonio


--
Antonio Maria Chiarelli, Ph.D
Infrared Imaging Lab,
ITAB -Institute for Advanced Biomedical Technologies and Department of Neuroscience, Imaging and Clinical Sciences
University "G. D'Annunzio of Chieti-Pescara"
Via Luigi Polacchi 11, 66100, Chieti, Italy
E-mails: antonio.chiarelli at unich.it, chiarell at illinois.edu



________________________________
Da: homer-users-bounces at nmr.mgh.harvard.edu <homer-users-bounces at nmr.mgh.harvard.edu> per conto di Green, Geoff <green.1425 at buckeyemail.osu.edu>
Inviato: sabato 3 marzo 2018 01:31
A: homer-users at nmr.mgh.harvard.edu
Oggetto: [Homer-users] Problem with wavelet analysis in Homer2



Hi,

I have a question about the wavelet analysis in Homer2.

I have acquainted myself with the software in hopes to apply a motion correction wavelet to the preprocessing of data we have stored instead of bandpass filtering to increase the integrity of the data.

Currently, I have been able to convert nirx data to .nirs files (homer2 compatible) in order to apply specific preprocessing functions to the time series for specific channels. There are 2 different wavelet corrections, motion correct wavelet and motion correct kurtosis, of which I am trying to use as a substitute for a bandpass filter. I applied the kurtosis to the time series, but this affected the data in a strange way (image attached "conc_kurtwave") compared to what the data looked like before the kurtosis wavelet (image attached "conc"). Is the kurtosis suppose to have this strange oscillatory type of effect? I have read a little about it, and it seems to be an advancement on the regular wavelet analysis. Regular motion correct wavelet sets one threshold value for the wavelet coefficient rejection (Molavi et. al, 2012), while the motion correct kurtosis wavelet has a threshold that varies as a function of the frequency of artifacts in the data; it is iterative (Chiarelli et. al, 2015).

I want to be able to compare kurtosis wavelet this to the regular motion wavelet to see which renders a better hrf in the data, however, I get this error message (image attached) when i try to use the regular motion correct wavelet. I see no error in the script it is referring to. Do you know of/can suggest any possible resolutions.
If there is anything that needs clarification, just let me know.

Best,

Geoff D. Green II


-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.nmr.mgh.harvard.edu/pipermail/homer-users/attachments/20180304/60f9a45f/attachment.html