And your cfg file (processing stream).
Hi Kim,
Just checked the data file you sent. It seems for some channels you have NaN values at some time points.
E.g. raw signal at ch 9 has some (see below). This causes the filtfilt function to break when called by Spline SG function.
find(isnan(d(:,9))==1)
ans =
798
799
800
801
802
803
804
805
806
807
A solution could be to replace these NaN values with the mean of the two neighboring points.
Hope this helps.
Meryem
@mayucel we could find the nan and replace it in most cases as you suggested. How do you handle it if a whole channel has nan? Is there any possibility to analyze the rest of the dataset?
Thanks for the great and fast help!
Jay and Stephen think this error was fixed on Jan 28. Download the latest Homer3 from github and try it
@dboas we tried this. With Homer3 it is not working. We wrote a script to replace the nan in the raw data as Meryem suggested. In most cases it works now.
Thanks for the feedback. I added this to our to do list to try and find NaNs automatically
Hi Meryem and David,
Our lab is running into the same issue. We have opened the .nirs file in Homer3 and written a script that replaces the NaN values as Meryem suggested. We have called this matrix 'd'. I was just wondering how to save this into a .nirs/.snirf file so our cleaned data can be analysed in Homer3.
Thanks,
Bec
English