Forum

Removing Chunks of ...
 
Notifications
Clear all

Removing Chunks of Data Before Preprocessing

4 Posts
3 Users
0 Likes
267 Views
Posts: 5
Topic starter
(@sluch)
Active Member
Joined: 1 year ago

Hello,

The start and end of the timeseries are irrelevant to the task, and are actually very noisy, so I would like them removed before I run the preprocessing like PCA. Is there any way for me to remove chunks of my data in Homer before preprocessing? If not, is it possible to do so with the NIRS/SNIRF files? I found this thread ( https://openfnirs.org/community/homer3-forum/exporting-raw-data/#post-744) where information is provided on how to open these files, but I'm not too sure on how to proceed from there. Any recommendations?

Really appreciate any help on this!

 

3 Replies
Posts: 11
(@cpickouol)
Eminent Member
Joined: 9 months ago

Hi,

I had a similar issue, as my recording involved practice and priming blocks taken after a steady state baseline.

 

My workaround was to open the raw .txt files in excel, remove the unwanted data and then save the edits as tab delimited .txt files. These can then be converted into nirs > snirf .

Reply
Posts: 5
Topic starter
(@sluch)
Active Member
Joined: 1 year ago

Thank you for the help Christopher! Did you get .txt files from your NIRS data acquisition software? In my case (using NIRStar) I only get .nirs, .hdr, .wl1 & .wl2.

Any thoughts on how to achieve this chopping up for this data? It would also be nice to split my different tasks, so I can preprocess rest data and task data separately.

Reply
Posts: 4
(@cory1993)
Active Member
Joined: 1 year ago

There should be an option to segment the time series in the Homer GUI (I believe under Tools). I do not know what version of Homer that was first implemented in, but it is available in the more recent versions that I have been using.

Another approach that I have used is to convert the NIRStar output to .nirs format (e.g., see NIRx2nirs) and then write code to segment the data at the appropriate frame(s).

Reply
Share:
en_USEnglish