search

alexenge / eegSeminaR

A university seminar on analyzing EEG data in R
https://alexenge.github.io/eegSeminaR/
Other
2 stars 2 forks source link

How to locate channels on eeg data #1

Open wenboyu0803 opened 1 month ago

wenboyu0803 commented 1 month ago

Hi, Alexenge,

i am make preprocessing on BIDS data, i found that it could not locate channels after importing raw data. Thus, i try to make channel location following your course (https://alexenge.github.io/eegSeminaR/articles/v3-preprocessing.html), however, i have no idea to choose which montage. i did not find the file "electrodes.tsv".

hope to get some advice from you.

Thank a lot.

Wenbo Yu

alexenge commented 1 month ago

Hi Wenbo Yu, if you don't have an electrodes.tsv file in your BIDS data (next to the raw data files), you will need to choose one of the standard montages provided with eegUtils (see https://craddm.github.io/eegUtils/reference/electrode_locations.html). For this, you will need to know what kind of recording system / electrode cap was used during data acquisition. If you don't know this, the safest bet would be to go with the default (dat_raw <- electrode_locations(dat_raw)), which should apply a standard 10–20 system montage. Pay attention if you get any warning messages about missing channels or case mismatches (e.g., you might have a channel with name FP1 but in the montage it's called Fp1).

Note that you may not even need channel locations at all, if you're not interested in making topographic maps (as described in section: https://alexenge.github.io/eegSeminaR/articles/v8-making-plots.html#scalp-topography). All the other processing steps and statistical analyses should work fine without channel locations.

Hope this helps!

wenboyu0803 commented 1 month ago

Hi, Alexenge,

Thanks for you help. i find i have a electrodes/tsv file, but i could not read it with funtion of import_chans. After i check the reference from Matt Craddock's web, i found This function cannot recognize tsv files. i decide to put this issue aside, for my eegdata did not need to locate channels.

But i encountered another question when i remove artifact with abolute threshold, that is, i could not deleted epochs which are marked as bad epochs. The number of epoch is still the same as the one before artifact correction. 1

More interesting, when i improve the threshold, more epochs are markder. this is not logical for i suppose fewer epochs should be recognized noise. 2

Great thanks for you help.

Wenbo Yu

wenboyu0803 commented 1 month ago

Hi Alexenge, I have solve the quesition of artifact correction, by set reject = False. By this way, i could find out the epochs which are marked as bad and then, i could select other epochs.

However, i still found that the amplitude in my analysis is much larger than we assumed. Then, i found the amplitude of raw data imported by eegUtilis package is much large than that imported in analyzer. I think it is quite amazing. Importantly, i import BIDS data, the amplitude is just like your course. (left: analyzer, right : R ) analyzer

If possible, please give me some idea. My email: yuwenbo0803@163.com Thanks. Wenbo Yu